ML20128N743
| ML20128N743 | |
| Person / Time | |
|---|---|
| Issue date: | 05/24/1983 |
| From: | NRC COMMISSION (OCM) |
| To: | |
| Shared Package | |
| ML20127B155 | List: |
| References | |
| FOIA-85-110 NUDOCS 8507130148 | |
| Download: ML20128N743 (562) | |
Text
{{#Wiki_filter:N 19, b. O h> UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION In the matter of: Docket No. / PEER REVIEW MEETING -. f - BMI-2104 REPORT (DRAFT) RADIONUCLIDE RELEASE UNDER LWR SPECIFIC ACCIDENT CONDITIONS VOLUME II: A BWR ANALYSIS i [y' 1=e A67 Location: Washington, D. C. Pages: i Date: Tuesday, 24 May 1983 I 9507130148 850425 PDR FOIA ALVAREZB5-110 PDR TAYLOE ASSOCIATES
~
Court Reporters 1625 i Street, N.W. Suite I004 Washington D.C. 20006 g (202) 293-3950 D
- - - .. p;, - ..;
,7.-.-,. ,; - - . .x, . , = . .
- 9,
. .~ . ,
1 l l 1
- =
UNITED STATES'OF AfiERICA NUCLEAR REGULATORY COMtiISSION f% j. 2. BMI-2104 REPORT ~(DRAFT) 3 RADIONUCLIDE' RELEASE UNDER-LWR SPECIFIC ACCIDENT CONDITIONS 4-VOLU!!E II: A BWR ANALYSIS 5 ' 6 Room 1046 1717 H Street, N.W. 7 Washington, D. C. 8 _ Tuesday, 24 May 1983 9 The Peer Review. Meeting commenced at 10 8:30 a'.m., pursuant to notice, Mr. Helvin Silberberg,
. - 11 Peer Review Chairman, presiding.
12
' COMMITTEE' MEMBERS PRESENT:
13 M. Silberberg, Chairman-14 ~ - M. Jankowski
^
g- 15 R. Bernero v
- g 16 D. Cooper vi u 17 -
a R. Vogel l' 18 D. Roe
't 1
- 18
[ W. Castleman n- ~
- 20. W. Kastenberg l.
21 { A. Reynolds 5 l 22 R. Ritzman 23 .L. Zumwalt
- 24. C. Johnson I
~ _q, , 4p.,,,;,
- " ' ' ' ' * * * ' * ~ ' * " " ~ ~ * ' ' ' ^ '
.=,=. g a::~.ag :- - - ,
w . . :.a . . .
-,.-s
*'7, 2'
'I
, LCOMMITTEE MEMBERSPRESENT: Continued' L .. - ;..
, - s D 2- - D. Walker 3! c l4'
.-- 5 ~
-6 g - .
, 7 .. .
8-9
'. - 11 12
~ .13 .
.14 '
.4e
[ '16 w
-o
- 17 -
4
.I
-e . gg -
'4
-r
- 19 .
.g"
'.r-k 21'-
-3 g n 24
,r-P. Ig
- M ,
%,' gQ* %Y#% M *Ph JN'b 8F *N "" M* 9' N '
~
=" ' '
?: ,
},,
TA-195-p. 3 MM:jli1:1
!- ~
1 _P _R _O _C 'E _ _E ' _D _I _N . _G _S Y.- 2 MR. BERNERO: Ladies and gentlemen, I would like to 3 get started please.- 4 Good morning. My-name isLBob Bernero. 5 About two weeks after the last' meeting of this 6 august group,-I was appointed to-direct the Accident-Source 7 . Term Program Program Office, so that's~why I'm here. 8 I hope to be seeing more of you now. As you 8 undoubtedly know already, the NRC and its contractors is 10 embarked on a. perilous path, trying to look at the physical-11 . chemistry of coremelt and fission product transport in order 12 to predict in a much more' realistic way what actually can get
- 13 j out.in a nuclear reactor coremelt accident. In order to have 14 both a technically sound and a usable estimate.of such source 15 '
terms, we do need -- in fact, absolutely have to have -- peer 2 l 16 . review, sound peer review to test the scientific basis on O l 17 which these predictions-would be made. I j 18 You demonstrated in the first review of the.PWR, l' 18 [ when all the long knives came'out, that this is a functioning 8- . peer review. I read those comments from the last cycle. And { 21 y . said! you. know when a peer review is working, when the 22 [ people who read the comments turn red in'the ears. And I am 23 pleased'to say that the process is~ working. I hope that our 24 ~ recycle through the. contractor of your comments is reasonable, l (.2 26 proper, and satisfactory. And I look forward to the further
, , 7 ,
- MM: j l'.1: 2' 4 1 comments you provide.
f'); ti 2 Now-I'11' turn the meeting over to-Dr. Silberberg,
- 3. :who-will chair =it for the balance of the two days.
. MR..SILBERBERG: Thank you, Bob.
5 As.you can tell, a' lot has' happened since we-last
'6 met.- I know there was-a period of silence, where you didn't 7- hear from us, and'that.was because everyone was going back to 8 their~ respective cubbyholes and drawing boards, trying to ,
8 bring together and bring to bear'the comments -- our response
' 10 . to your comments from the 'first' meeting.
11- By way'of introduction, before bringing you up to-12 .date, .where we have been in the last four months, what I would 13
. like to do is just note a few regrets from people who were 14 not able to be here today.
15 First, from.the United Kingdom, Mr. Abby and 3
-] 16 ~
Mr. Potter were not able to be here today.
.17 ' Mr. Thorguson, from Canada.
18 Bob Hilyer, from the United States. 18 -And Saul Levey sends his regrets -- personal h
# problem, he was not able to be here.
a 21 . But nevertheless, as ever, he was productive and he { 22 sent in'his comments ahead of time. And he asks that they be 23 read.into the minutes, and we will do that sometime -- probably 24 tomorrow.
)
25
- i. Carl Johnson was nice enough to tell me that he was
,. ..,.,,+a... ~ . - .. a . . .- -.- - ,
^ * ' '
. .. ? . ~ __
- Dis] l 1 3.
'5 I going to be an. hour late, so-we can expect Carl here pretty ! -2 >
soon. 3 In the meantime, we have'with.us our colleagues from 4- overseas. 5 We have Mr. Petrangeli, who was here at the first meeting, from ENEA, Italy. 7- I don't see-Mr. DeMunk.yet, from the Netherlands 8 I assume he will arrive shortly. Licensing Authority.
' And Dr. Soda, from GERI, from Japan.
10 Now, there will-be, throughout the course of today,
~ ~ II j a variety of backup materials that will be provided that I*
Battelle's staff.has been working on diligently that we will I* be passing out in the due course of the meeting today, things
" like a separate appendix on a-description ~of the TRAP melt 15 code.
I I' I must say, in addition to Battelle, Oak Ridge Labs 3 17 e ha~ v e put together and brought with them, I hope, their report 3 18 j on the tellurium chemistry and the basis for how tellurium is I' being treated in this report.
! 20 W And we will also provide a summary of peer review comments that-I will say a little bit more about before ! Jim Giesecki makes his presentation.
23 (Slide.)
- 24 As Bob Bernero noted, we have taken your comments
}- ,,
quite seriously and digested them. And within the first month
MMtjl 124 6 1 after our -- after the meeting, there were a number of things (~% ' ( 2- that became quite evident to us as a result of your feedback, 3 as well as our own impression from the first review. 4 I've listed here four items -- there were many more 5 -- and I want to assure you that they have been correlated 6 and summarized. But there were four comments in particular 7 that actually had an impact on how ASTPO did its planning and 8 how we revised our approach to the program that has unfolded 9 since the last meeting. 10 For example, many of the commenters -- there was 11 a consensus that said the data base for the supporting -- the i 12 data base which supports the codes and that might beiused for 13 code validation is not visible in the work that we presented 14 and that this needs to have its own emphasis.
. \
15 5 Second comment had to do with the fact thatIthis l 16 work has large uncertainties to it and in it, obviously, at O 17 this stage.
$ And some parameters are more sensitive than 1 # 18 others.
I 19 k And so a need for an uncertainty and sensitivity i 20 l analysis was presented, Along with some of the other support-5 21 ing activities, we have now added these as supporting ' 5 22 activities to the Staff work, as well as to the contractor l 23 work. 24 Also, as you recall, there was much debate on the f M question of just what the containment loads are and the
MM;jl 1:5 7 1 challenge to the containment in terms of containment response, p t 2 particularly with respect to the early failure mode question. 3 And you will see later that we have addressed this 4 in a supporting activity. 5 Finally, the question of new technology in dealing 6 with reactor cooling system upper plenum, in terms of thermal hydraulics modeling, was' clearly an area that was noted by 8 consensus of the reviewers, in terms of working with -- start-ing with the MERGE code, which Battelle and everyone else 10 admits is a first start to the problem. And we have had to 11 address that in our supporting activities. I* Now, putting this altogether, basically we now have I3 (.) four elements to what we call the reassessment of the technical I* bases for source terms. (Slide.) 2
$ Element One is a preparation of a summary report on 0 17
? the data base for validating codes to predict releases for the 3
18 E codes that we used in the Battelle reports. This will bc g 19
, e put together by various NRC contractors making contributions I 20 8 to it, and the Oak Ridge National Laboratory has the lead and E 21 i
g will publish the report as an Oak Ridge document, with due i 22
- credits to all of the contributors.
23 Element Two is basically the same element. It's the
/ Battelle analyses for what has now been enlarged to five
\.J g5 plants. And I'll discuss that when I go into the scope in a L ~ 4
MM:jl 1:6' 8 1 little more detail. 2 We have now identified Element Three, as Bob Bernero 3 noted, as a very separate and distinct activity that is 4 deserving of a status all by itself, namely the thorough peer 5 review of the scientific basis for the work -- first, what we 6 call the technical experts' peer review, which is what is 7 going on now and what we had in January and will continue,. 8 and then a broad-based review by an independent scientific 9 organization. 10 And we have been working on this for some time, and 11 we expect, shortly, to receive a proposal from the American 12 Physical Society, to hopefully participate and conduct this j ' 13 review. 14 Finally, Element Four, which always existed, but now 15 o is a lot clearer in our minds, is basically the element which 2 l 16 picks up all of the supporting staff activities that you will 0 l 17 see in a moment that evolve from the first peer review, as i 18 well as a way of pulling together all of the contractor f i 19 products, together with the Staff-supporting activities lead-ir 20 ing to an appraisal of the significance of the reassessed 21 source terms, when we are all done. { 2 22 I will say a little something about schedule before l 73 I close, so you can see how they all fit together. Cnd t.1 24
.' 3
'XZ' L'/
' fee
' 9 1
MR. SILBERBERG: The scope for Element 2 has been
' C'N 2
\. enlarged to include now the Zion large high-pressure 3
containment PWR right here today. We now refer to the
-Battelle reports as the BNI-2104 series, and they are.
LS Battelle Reports Volumes 1Lthrough 5, 1 through 4. And 6 they, if you will, stand alone. And for that reason, we 7 refer to the contractor designation and not to a NUREG 8 designation. (Slide) la Element 4 has expanded as a result of the peer 11
~
[ comments. Starting with an appraisal of the products from 12 L Elements 1 and 2, we are now adding in parallel studies
~13 that will largely done by the staff with some support from
( l' NRC' contractors on the state of the art.of the loads which J4
~' 15 one -- if you will, best-estimate loads -- wnich one might 16 expect to be needed in order to determine the response of 17' containment for, if you will, steam spikes first, hydrogen la burns, and other items that are challenged.
l' Containment response in terms of early 23 containment failure. . If there is anythding we learned from 21
- tne Surry study -- and really, it shouldn't nave been a 22 -
surprise -- was tnat early containment failure is really 2J the thing that really challenges risk based on the current 2" technology available now, whereas in late containment 25 failure we have a lot of other retention mechanisms in the TAYLOC ASSO CIA TES 1625 t Street, N.W. Suite 1004 w eshington, D.C. 20006 (202) 293 3950,
10 I containment that are acting to reduce source term. 2- 'i The uncertainty.and sensitivity analysis, we 3 have.'been reviewing with several. laboratories, possibilities for now this might be conducted. And early in 5 June we w'il make a decision as 'to how we will organize
'- Lthis and what'the scope'will be.
7 If you will, this is a rather challenging task, 8 and since we want to get the first study done in 6 months, it will, of. course, ce limited in many respects, but lu hopefully will form the basis for an ongoing follow-on on 11 uncertainty and sensitivity analysis that will be more 12 l comprenensive, that we can ultimately use at the ens of the 13 program to pull these tnings together. So you will hear t I*
-g more about this in'due course.
15 Inis element also picks up what we call the 16 review and appraisal of work going on outside of this 17 program, if you will, in the industry; namely, tne IDCOR i la study and the ANS source term study.
~
This will be reviewed l' by NRC contractors as well as the staff and wi11 allow us 20 to maxe a comparison'of wnat -- now others have calculated 21 similar events and similar sequences. And we will, of 22 course, factor those results into our appraisal. 23 Finally, we have decided to ask the Los Alamos 24 people to use TRAC initially to explore and improve the 25 analysis of'the upper plenum in the reactor coolant system
, TAYLOC A SSO CI A T ES 162 5 I Street, N.w. - Suite 1004 w ashington, D.C. 20006 (202) 293 395J
v- -~ . ., 11 1 in the vessel to, if.you will, support or extend the MERGE 2 analysis. And this is now underway. 3 In Element 4, we will, of course, appraise the
- comments from the peer review, first this group and then 5
finally the indepennent group that will do the appraisal. Now, when we are all done in this time frame, we 7 will nave what we call NUMEG-0956, the real NUMEG-0956, 8 wnich will be the sequel to 0772, and will, in effect, be a reassessment of where we stand at this point in time, if to you will, at the end of this year ano into the spring. 11 we will not release the draft of 0956 for 12 comment until after all of the peer reviews are done. And 13 you will see this on my next slide, l' (Slide)
~ 15 We should have the code validation report l'
complete by August and reviewed. The five plant reports, 17 still draft but issued as draft, complete in September of te ed3. The peer review, last peer review, which will be the 19 one that I referred to, the independent peer review. The 20 broad-based peer review, will ne we hope around May of '84 21 and would then lead us to 0956, publication of craft for 22 comment and then -- I mean broad public comment. Tnen 23 December '84 would be a final report. 28 Now, wnat is the schedule for the peer reviews 25 of this group. t
~, fAYLOE ASSO CIA T ES 162 5 I Street, N.W . - Suite 2004 W ashington, D.C. 20006 (202) 293-3950
~ ~
_ :. EJE . 12 1 (Slide). p
,( 2
~Let me give you some of tnese dates and what 3
our-current plans are so that hopefully you can make your
-*- plans for tne summer accordingly.
5 We are here today. .we hope-that'in early July, 1maybe secono ween in July, what-we would do is complete the , 7 Peach' Bottom analysis and review and Grand Gulf and bring
- - e on sequoyah, Sequoyah review.
tio w, we recognize at this meeting today -- ano 10 we ceg your indulgence -- that we'did not have all the l 11' materials available to-you on both Peach. Bottom and Grand l 12 Gulf. l l 13 But we feel that with the materials that you l'
' nave, plus the oral presentations and the handouts, that 15 there are enough new issues -- some old ones, too -- but l'
enere are enough new issues in dealing with the BhR designa 17 that the next 2 days-will be well spent and will be an la excellent, if you will, precursor to the July meeting, l' wnich would then allow us to bring into final focus the l 2 'J Peacn dottom and Grand Gulf plan and Sequoyah. If.possible, 21 we would add an extra day to the meeting if we neea to. i 22 Hopefully, around tne middle of August, second 23 l ~ week of August, we would review the Surry plant, revised to 24 bring into account additional analyses that Battelle is l 25 currently planning and currently working on that Jim i 1
/
(,- fAYLDE ASSO CI A TC$ 16251 Street, N.w. . Suite 1004 w eshington, D.C. 200J6 (202) 293 3950
13
-1 Gieseke will report on as the next speaker.
( 2 We will have_the Zion plant reviewed in August, 3 and it's not clear at this time how many of the sequences
- we will go into in Zion. It could very well be that we may 5 '
just pick out one or two that would allow one to get a reasonable comparison with a surry-type plant. And that is 7 still to be evaluated. It's really a function of time and 8 resources. , Tnen finally, we would ask this group to review 10 the coce validation report from Element 1. Hopefully, you 11 woulo have the code validation report to review someti.ne in 12 early July or mid-July. 13 Now, because we have brought on two additional ( .,) l' areas of work which in themselves require we think separate
15 consideration, we have two possible additional meetings 16 that we will schedule. One, of course, would be dealing
\
17 with containment loada and response, and we think this is would be'the appropriate meeting for specialists that deal l' in these particular proolems. So I suspect we may narrow 2'J down the review here possibly. 21 Then on the uncertainty analysis, since that is 22 so important and has auch impact on judging tne results of 23 the entire program, I believe we will nave an extra meeting 26 on that, D'u t it's not clear what tne composition of that 25 review will be. (, fAYLOC ASSO CI A T CS 162 51 Street, N.W. - Suite 1006 W athington. 0.C. 20006 (202) 29).3950
p _ -
-- - - n _
. , . . .. c _ ..
14 i 1 That's really all:I had to say. If tnere are 2 some cogent comments or what have you, I would be happy to 3 answer them now for clarification. Doug Cooper. 5 MR. COOPER: Do we still have under consideration E-the simulation of the TMI incident and the comparison 7 between what was predicteo and wnat was found? 8 MR. SILBERBERG: Yes. That's sort of an ongoing 9 activity in the office, and Walt Pasadack of our staff is 10 following tnat as well as Mike Jankowski. 11 We have some ongoing work at several of the l 12 laboratories that have been looking at this problem. At i 13 some point, I ara not sure when, but at some point it will 16 find its way into the program, maybe as part of Element 1 15 or maybe as part of Element 4. l' But we think that the TMI-2 accident and its 17 value here and its relationship.and its role nere will have 18 to be placed in perspective. So you can expect that.that l' will'oe done, ana we will certainly invite your comments on , 23 te, 21 But we also, in dealing with TMI-2, part of the 22 problem.is it is not quite the same sequence as the 23 sequences we're talking about here, so it needs to-be 24 looked at very carefully and dealt with in its own special 25 . way. But neverthelesa, that is certainly a useful thing to ( TAYL0E 4550 CI A TES 1625 i Street, N.W. . Suite 1004 W annington, D.C. 20006 (202) 293-3950
~ , . . - w 15 L 1 do.
2 MR. KASTENBERG Could you tell us a-little bit 3 about wny you. chose Zion as the fiftn plant instead of 4- .something like'the MARK-II containment? 5 MR. SILBER 8 ERG Certainly. We could have also additionally chosen MARK-II. I
'MR. CASTELMAN:- hhat made you choose Zion-2?
e MR. SILS8R8 ERG Zion represents kind of another class of hign'-pressura containments, really different than, 10 1,ge s say, the Surry class or the Oconee class. And also, 11-it allows us another comparison with a companion study 12 en,ge s going on in IDCOR, which has included Zion-2. So.it 13 kind of gives us -- it spreads the high-pressure I* containments, and it-also allows us to communicate with the 15 IDCOR'results. Those, I think, are the main points. 16 As far as the MARK-II, what we will~try to do in 17 Element' 4 of.tne staff appraisal of risk and regulatory 18 signiticance, we will try to place the MARX-II in l' perspective relative to the results we have on MARK-I and 23 MARK-III, and to the extent that we find perhaps that it's 21 worth an extra calcuation or two, we may indeed have to do 22 taat. 23 MR. BERNERO: I' think it's worth adding, the 28 licensing s'taff La deeply involved in the review of the 25 Limerick PRA, which is a MARK-II containment. And we have Q~, , T A Y L. 0 C ASSO CI A TES
$625 I Street, N.W. - Suite 1004 w asnington, D.C. 20006 (201) 29) 3950
l 1 1 communication with them, and some of tnis program's C. 2
.. calculations may be generated to assist them, out the 3
MARK-Il focus you will find there in that review rather
" than here in these five plants, just due to the limited 5
resources. 6 7 8 9 10 11 1 12
! 13 ! 14
{ ,, '.
%- 15 16 17 14 17 l
20 21 . 22 23 24 25 1 1
'q TAYLOC A SSO CIA T ES 3
162 5 I street % W. - Suite 1004 d ashington, D.C. 20006
.. (2J2) 293 3950 lN.O
*b,
_ _ . . , _ _ , . _ _ _ -..,.. , _ _ . , _ , _ . . . . . . - . . _ _ _ , _ . - _ _ _ . _ _ . _ - _ . _ - _ _ _ _ - . _ _ . _ . _ _ _ , - _ . , . _ . , = . . . _ , - _ , . , _ _ _ _ . _ . ,
.T U 17 7 rgl 1 MR. SILBERBERG: I would now like to introduce
.. 2 Dr. Jim Gieseke from Battelle, who will tell you about what 3 Battelle -- how Battelle . is addressing and has addressed 4 the January Peer Review Comments; and in concert with Jim's 5 presentation, later on we will be passing around a summary 6 of the comments that we received from various reviewers and 7 observers from the January meeting.
8 A rather concise summary has been put together 9 by Chris Ryder over here in back of me from our staff, and to sometime throughout the two days if you have any comments 11 on how we have interpreted your comment or how we summarized 12 it, or abridged it or what have you or you think some clarifi-
\ ',..D / 13 - cation is needed, please make a note of it on your copy, 14 and sometime during the meeting if you can see Chris Ryder, $ 15 why he will be happy to take your notes ar.d make appropriate l 16 corrections. Chris is over here. Jim Gioseke.
0 l 17 MR. GIESEKEs. Thank you, Mel. I'm happy to be 1 8 18 back in spite of what you might think. t i i 19 (Laughter.) i j 20 We rather enjoyed all the discussions the last 21 time; we certainly thought we'd enjoy doing it again. { 2 M (Laughter.) l (Slide.) 23 There are really three topics today. This is the 24 first of three that have to do with this.. BMI 2104, Volume 1, 7 25 is the topic of our last discussion, the Surry plant and
.- : :a = . . . . .. . . - . . .
MM3- . 18 "rg2.
~
1: its' calculations. !We want talk first about what.we are doing
, i ' ' 1 -- in_ response to the review comments, but before we get into 3 that too deeply, just to restate the situation and make it 4 clear'where we are going with all of this--
8 (Slide.) e --I would like to read for.you the objectives of 7 the~ study, which-are to develop updated from release plant. 8' fission product source terms for - five types of nuclear 8- power plants and for accident sequences giving a range of to conditions. 11- I think that's importants to realize that we're 12 looking for ranges of conditions to see the effects of, as you la will notardown here, the effects of different conditions and 14 l different assumptions. The estimated' source: terms that we 5 to derive are to ly. based on-' analyses of fission. product released I [ 14 from the fuel, transport and deposition af those fission pro-0 l l 17 ducts usina improved ~ computational tools in a' consist'ent step-I 18 by-step manners g 19 Our second objective is to determine the ef fects I, 20 on fission product release associated with the' differences f . 21 in input which are associated with differences in plant design 3 l 22 and accident sequences primarily. We've also done some sen- . 28 sitivity' calculations'as we've gone along, just to see some l l . 24 major impacts of different consumptions. l' (" 25 And the'last objective, then, which is a result,is
> r =+ ,
au 19 rg 3 1 to provide in-plant time and location-dependent distributions (~,
, 2 of fission product baths.
3 (Slide.) 4 Now to give you an overview, I'm not sure at this 5 time whether this Zion will be a separate report from the 6 Surry Revisited. But to go through Volume I--is the Surry 7 plant- not " slurry plant." You don't know how many slides I 8 corrected that on, I must have missed one-- 9 (Laughter.) to -- which we talked about before, which was done 11 using the MARCH 1.1 code. 12 Volume II is a Mark I, BWR, Peachbottom plant. 13 One of the topics of discussion today--sequence is AE, TC and (TT.) 14 TW.
! 15 Volume III of this is Mark III design, BWR, I
[ 16 Grand Gulf is the selected plant. We're looking at sequences i f 17 TC, TQUV, TPI. This is also to be discussed later, calcu-1 2 18 lations that we have to date. I h 19 The next volume issequoyah ice condenser contain-l l 20 ment. We're not entirely fixed on these, but we think we'll { 21 probably--it looks like we'll be using the S211, S2D, TMLB, s 22- TML sequences. l 23 Then we come to Surry Revisited, which is revisite(! 24 in the sense that we'll be using MARCil 2 code rather than
\-
25 MARCil 1.1 code. nd rg i fis
~
MM:31 3:1 (fl' rg) 20
~
1 You'll hear about the major differences that have (, 2 to do with, of course, slumping, the same sequences as before 8 for Surry, and finish up with Zion's. 4 We now mention the sequences -- and I totally defin-5 ed on that -- to clear up a little bit the differences -- well , 6 before I go on and talk now about we have done with regard to the Surry plant, I do want to acknowledge input that we have 8 been receiving. (Slide.) to Westinghouse has provided us with information on II the upper plenum design; Stone & Webster on the co'ntainment 12 geometry and the compartmentalization.
- r. 13 EPRI -- Dick Vogel has chased down information on
{} I* the concrete composition for us. I And Sandia, of course, as you are aware, provide'd us l 16 with release from the core concrete interaction. 3 17 g And Oak Ridge has provided input in terms of release 18 j from the fuel. MR. JANKOWSKI What about B&N7 1 2o I MR. GIESEKE: On the Battelle, Northwest has pro-I vided this information, the spark code for the boilers. But 22 l w,,11 acknowledge them when we get to the boilers. 23 (Slide.) 24 We had a number of peer review comments. As Mel k 25 mentioned, these are being handed out. This is a summary that t
v n =4.p , ; . - - - - ----- - - a m ;--- :------ n-- -- ;- 4 ;;a ;z.; : 3---- ;--;;;;g [ $301:jii Liil:
^
{ .c
~ '
f t 'i. 21
, , j ,
f.; : . r ;. M .
' 4
( p . wn ( i
' 11
@.. Mell alluded' to by: Chris Ryders th'at - will ' beCh,re,vided to you. b g^; V,i t, [ U 8 But;the generalitopics in there, I'd.like, for your-informa *
^
;f ,
-fg ' ?
.t' ion,it'o quickly;ryn.through_have to-d6 with -- the cenments kl; 8 2 44' w
. ere categorized under,the , $o,pi'es of acriident sequence and
. , ,. (
- 7 8
~
system bthavior, fission' product release from fuelt. thermal #. ' s: jf i * .s 8 thydraulics;
- a. chemistry, . primary" system s s
transgedt((deposition q q and
.7=
reentraiArrent,'tr nsportfdeposition'an[ dehntrainrrffit - it: ( , ' sp 8 containment,.codervalidity sensitivity,.and also rep 6rt Y }'* structureia[organizationreportingeneral. < q '
, s. '
-l
#~ These topics and theico^mments we have read and t
) V w, <
11 ' .
. considered.jaAnd the pasult'of these consideration % have ledI. -
d, w 3.3 . .y3 12 tosomemodificatirN3or,additionalcalculationsthat6we're d .f s j 13 doing.from the'SurrysanaIyses. / 14 (Slide.) .~ g Y 15 I I have tried to usel t!his slide to show you where f.
, e (
18 l the material'is golwi tc ahy Ir, . o. u g7 volume 1, yhl'h
/ m.y
- 2 V 2, is the first Surry analysis, using . [dl .
.e I
e gg j ' the MARCH 1.1, is probably,4Joing to be. published in a matter , F " of a- few weeks,' prNtty 'niuch as is, with some modifications l l' I " r noted on here or some considerat.lons.- And these will be
~
g * {. 21 discussedtodaylkihese:additionalcalculationsormodifica-O 3 e s 3 ;c
) - 22 tion ~s.
[f V 23 ' _ Tnere are'other* issues that d are addressing, th t$ st. to cf rify what- we ar[' *
~8' and we wi 1 run throu 3
Q, y- , L L doingtothe{Ehrryanalyshs/'asaresultofour:lastpeed, -
- t. :
1- ; > c ,; 3 , g,* ,N Q , l' *
, '- )-_r i ;, . N_"
L . , L !: 2 1: __.:_ l_l_l::_!:T_2i_ _:-_l' :.'
- _0
^'
- - .5 - l._ T-
~
1
e - ma :. , . g- . . . .. . .. .. , , ~ . .
' :MMtjl'3:3' 22 I' ~ review meeting.
'. (c-2 First of-all,-weiwill be-using MARCH 2 code for all 8 ~
. the sequences.- :That will not appear in Volume 1, but it is a 4~ ~
basisIfor'a volume.which I identified on the previous slide
.5- '
,,being' Volume'5.
6 AsLfar as the upper plenum geometry in the MERGE calculations, we.obtained'information, as I' mentioned, from 8
~WestinghouseL-- as.I mentioned -- with some details. We'll befgoing-through and using a better definition of the: upper 10 plenum' geometry.
II' Improved release rates from fuel -- this.is basical-
;<1 12 ly the work done by Oak Ridge in considering two types of 18 release, one being the control rod -- release of control rod I'
materials, and the second item being tellurium release. And {. <3 15 3 y. I think you may hear some of that today or tomorrow, of what j 16
-they have done.
0 17 g The fourth item is an issue that came up as a 3 18
~1 i question: What is the effect of the inventorp distribution o a p 19 f ,f : q in the' core?
{ g.- t . We.had assumed that it was a flat profile. We
-thought'it might be good to do a-little bit of' sensitivity
.h 1 22 2 analysis. We had-done a little bit on this to see what the 23 difference might be on release rates.
24 a We brought up the issue of decay heatining of i ' 9.' 26 l deposits as possibly having an impact'on subsequent deposition
$l l .;
' ^
b *i >_
j MM:jl 3:4 23 1 of the fission products in the primary system. We have some C 2 calculations, preliminary sorts of calculations that we can 3 talk about today. And that issue will be addressed also in 4 Volume 5. 5 There were comments dealing with the nodalization 6 And when we redo it, we will look at of the primary system. 7 the nodalization, expecting that we use a finer nodalization. 8 There was a question that we had in our minds about 8 This all has to do the Surry geometry of the containment. 10 I think we may have used the with the core concrete release. i 11 We don't know what the impact wrong diameter a little bit. 12 of that is -- wrong diameter of the cavity. But we correct 13 that, and we get melt temperatures for MARCH now, and we would ( f. 14 look at the possible impact of water in the cavity on core M 3 15 concrete release.
! I8 Another item is the effect of water condensation on-o II h to the wall and what that does to deposition in the con -
3 I8 f tainment -- or as referred to, as diffusiophoresis. We've i 18
,{ done some calculations we'll report today to show the impact E
j 20 of that -- plus , that will be -- it's been added to the NAUA 21 { code, so that will be automatically be included in the next 22 l volume. U , (Slide.) 24 We keep going down our list. I think there was 25 some question about flow rates that we used before -- spray
r- - 24 MMsji 3:5 1 removal efficiency in our calculations for the containment, 2 removal from the containment. We've gone back, checked those. 3 We have some calculations to show today, and we'll be using 4 the same sort of procedure for the next issue. 5 We looked, just very briefly, at. sensitivity to 6 That will be mentioned today in a talk. drop size. 7 We mentioned, when'we ha'd our last review meeting, 8 that we had not included the deposition from the flow back 8 through the primary system, after melt-through -- the reactor 10 We had not taken into consideration any deposition, vessel. 11 and that will be included when we redo it. 12 The questions of' compartmentalized containment is 13 - to be evaluated, and we expect to use a better compartmentali-a 14 zation in the calculations. 15 Now that we have obtained information -- better 3 [ 16 information, definition _of the containment geometry -- and
=17~ we had not. prepared this appendix before, and we are handing 18 It will go in f<
out an appendix today for you to look at. 18 It just l5 Volume 1. There was a place for it in Volume 1.
# So, that will be included.in
!- wasn't' completed before.
- 21 Volume 1 and referred to by appearing in Volume 1.
22 Then, there was some analyses that were done on
.]
23 pressure spike sensitivity, or hot-drop model sensitivity that 24 -we will discuss. briefly ~ today, relating to the containment 25 fail'ure analysis.- This is not a part of either of these, but 1
=
-'D -
g w e- ^ $ A
~
,. : MM:jl 3:6 25 1 is input to NRC's containment failure analyses task that was
- p 2 We thought it'would warrant a brief go-through on mentioned.
3 that, because it was a major question that came up in the 4 peer review meeting in January. 5 (Slide.) 6 Okay. So, I have now a list of the topics we will 7 be covering in this part, the additional results with the-8 additional calculations that go into volume 1 8 (Slide.) 10 TRAP melt 2 description, as I mentioned, will be 11 available and handed out.
~
12 S2 BU Pressure spike sensitivity, relating to containment I3 failure, is a topic. [% s} I Transport and deposition in the containment, with 15 diffusiophoresis. And there's questions regarding spray. 3 l 16 One more item I! forgot ~to put on here, which is the
.o 17 effect of' decay heating, and we won't forget it just because a
I8 it's not on the list, but we will talk about that. f t 18
.,! So, these are the items, and we'll take'them in ir 20
- order, starting with.the[ TRAP melt description, which is just
'21
{ a handout. 3
;j 22 . go, y 11 go through -- Pete Czybulskis will talk 23
.aboutLeont'ainment, questions relating to pressure spikes.
24 Ken Lee will talk about transport'and deposition in
\.. / 26 the containment, as revised. And Rich Denning will. talk about g a. . . . . - . .c _
---- _=
126 Majl, 37
~1 decay heat on the primary surfacesfand the:effect on deposi-4 aa
[%. 2 tion. 3' "So,-with'that,fwe'11. move on to those talks on some 4I of'those' topics, starting with Peter Czybulskis. 5" 1MR. VOGEL: Mel,~I've been assuming we'11 get copies
.si of the slides?
7: MR. SILBERBERG: .Yes,;we'll. cycle them through after 8 each speaker.. O MR. GIESEKE: Do'you want. toL have . the viewgraphs -
~
8 ~~ in one' big pile or, individuals.for each of the. issues? II'.. MR..-SILBERBERG: -One --
\
12 MR.LMALKER:: Are you going to do any more on the 13
)- secondarystructufes?'
1 14
. -MR. SILB,ERBERS: For the a 111ary. building?.
.. q-j -15
'MR. GIESEKE:. Yo'u're: talking about the auxiliary 5 j
-[. 16 building? 1 1-0 i.
-17 l :. MR. )fALKER * - . Yes..
I 18 MR.: GIESEKE: We're getting all the geometry, infor-2 18 '
-8 mation an:1 checking f that:out as Ee' go back through.
"# 'MR. CZYBbLSKIS:... Good! morning.
jL -
-As a result-ofethe'discussionsiat the last' peer-3
}. 22 review?meetingc there aresa number'of' question's raised about:
[ the pressure spik'es.Lthat.-let to the ' potential early containment (., 24: failure;injsurry. K.) y Just let'me make,a pointLof clarification.. I t a t Jr L '"
+ 4 c. nwh s-- y. - t
. ep ' -
ee p or'&'w & , k ** ~ *H- ^
,A
m.m z__. . _ leth{l 38LL , ,
'I:
believe,in'the. draft of BMI 2104,.welmade no statement about b-m 2- ~ the likelihood.'of containment, early containment failure. I-m
' I think all the release : calculations. were predicated 'on the f act 4- that the containment's failure ldoes take place. 'It said noth-ing.about'the probabi11ty, and I t) ink-there is some misunder-
~
5 ' 6 standing as to wh'at 'the intent of. it was, Whether the probabil-
~
c '7-itycis-low'or highfis still.'an open question and one that will 8' be addressed further.:
- What'I will~ describe'to you today is.a-series of 18 -
sensitivity studies that'we hav'e performed since the last peer II 2 reviek meeting to'try to_get some insight.on.how-the peak 12 - pressure associated with an'early. pressure spike might vary as
~: .
13-
) ,
a functiion of modeling- and input l assumptions. .
- I4 I might also note that'these calculations were done 15 with MARCH 2;.as opposed to MARCH.1, which was used in the 16-other study. There were rather substantial revisions in-the 17 input in the physical.~ description of the planttin these studies 3
j . 18 4 from the one we used earlier.
,E.
I In particular, we had the benefit.of some. detailed-
; h, #- '
discussions with Westinghouse-on what the internals of the
. 21 Therefare.still'some outstanding questions, reactor look like.
22 and i I'm expecting 'somen.further input from Westhinghouse as to, 23 in particular, the lower support structures. .But these.calcu-24 lationstwere done!with the:best information we had available. n, So,Llet me-talk about.these. J r +-#+ > = 5. - m E.r> , * -- ~ h m e r: . .e 6" + p , _ -,./. ,'J su a s A A
m -._%. >_ . . , ~ . . . . . - - . _ . . - . - . - - MMajl 3 9
- 28 I' ~(Slide.)
p .=
,2 The objective?of these studies was to investigate 3 the_ containment pressure loading as a function of modeling and 4: inp'ut assumptions, very simply.
5 (slide.)
'6L The way we went about this exercise is illustrated 7 here.
8 As I said, we~ calculated the series of MARCH-2 9 cases, varying the heat transfer' assumptions and the heat to transfer models that we are utilizing included what I call 11 1: articulate heat transfer, which.-is basically a very simple 12 quencher . model - that assumes, for a givenparticle size,as long as
,([)lv 13 the water is in the cavity,'the. particles will' find the 14 water,until you either run out-of water or.until the particles g 15 are quenched.
_I
] 16 That's a point of~ departure, basically very.similar f 17 to what we have been doing for MARCH 1.1 for some period of I
l 18 time.
..I -
;g 19 We did this for a series of particle sizes.
k' 20 The next set of calculations'was particulate heat
~21 transfer, up to the point that the core debris' form a debris 5
j 22 bed. And the time for forming a-debris bed was determined on 2 thef basis of levitation calculation. Basically, the more rapid l 24- the heat transfer, the more' steam you generate, the harder.it. 25 is.for.-the bed to form. l g -- p.., y y -y+- 9
. - . . . , . - . , ,+ . .- .. _ . ;
'MM:jl. .3 10~ 29
.1 .
Again, these were done as a function of-several _ f i-- L.1 " 2 particle' sizes. 3 The third. set of calculations were,-again, initiated But instead of using the
~
4' with - the : particulate : heat transfer.
- 5t . levitation c'alculation, we just' assumed that the-debris would
's' form as soon as the particles' solidified.
- 7. . Then, there was a series of calculations done in 8 which I varied a number of' assumptions about the in'-vessel 9 ' behavior, about--how the' vessel fails, to see what difference cnd 3 10 .- it makes on the predicted contrainment response.
-11
. 12 13 14 -
5 15 4 y { 16 0 17 0 3
^ 18 5
19
.,h .
2 20-t
; 21 22 24~
(: ,
' Y._Y ' .
. ~
~
\
4 I & p$r* - pp 6*
.-y.#r *ui p ,.,9 y , , _g s, ,-
" ~
.~ - - ~ .-
(T ?1'y- ! , c 1 L. ', ;. . - 30c i.MMaily 11,
> , 1 : (slidej)
- n_
Just'.to. provide: some' perspective, .let me point' out
~ ,-yL; ; ,:2 -
the ibase l case here. .Th'is' was Surry TIJiB, prime, using the-
^
- 8 -
- 4 ' assumptions' on = the : code . as it is. Th'is is kind.of the l5 L reference time. scale,.with' steam: generator dryout-68I minutes,.
.j- 8 core uncovering-:at.I99', start of; melt'at 120, start of' core
- 7. - i s' lump fl49.c 81 ,
; r might: point .out. here I; use what I call a - gradual
.r P
'8-Lcoreislump.model,?.as opposed..to letting?the! core slump all at
,10 l
~
'on'ce. =Though it' turns.outLthatifor these calculations once.
- 11f the. core slumping.. starts,.it proceeds very rapidly to com-
- 12 ' -pletion. Of course,'as I said,-~the core started'the slump
-hi
.J.
18- at'149~ minutes and collapsed'at 151 minutes. - Bottom he'ad
- 14 Lfailed about a minute.later,- so most of the-~ calculations that 15
-I .will be talking about!' take place following~ head f ailure .and j' 'le ztreat the debris failure in tihe cavity.
-O p - 17
- (Slide.)
- g. ,
~ 18 Let me throw up 'a reviewi pressure and. curves.
;;f :
19. Just a~few examples,.;and.then I'will summarize what
=E.
8' ' fis1seen.-
.,r.
- 2.21- ThisL i s a typical pressure-timeLhistory. This
{ 22 : Lhappens"to;be for5 debris particle. size of:one inch.
~
'And this "I i s IforhtheLparticulate points model.iThe' core uncovers in-
~#
Jn - , ithisharea- :then melts., The head falls at this point,-and. _ _&t l . .#. :lthis is1where you"get ;the interaction :-in the bottom' head, w T -
-,Nw 4
g i141, w 80 V N'"
MM:ji~4:21 31 I In'this particular case, the' peak pressure is about 80 psi. M" V
- 2 That's for one-inch particle size.
L 3- MR. REYNOLDS: I-have a question. Is this for
'4 w'ater coming down on top'of the debris from the accumulators?
5 MR. CZYBULSKIS: This is the water coming down on 6-top of-.the debris due to the accumulated discharge at the time 7-of head failure. 8
'(S lide . )
1 8 Same sequence, same modeling assumptions, just a 10 smaller particleisize. Youiget essentially the same behavior, II but instead of 78'or 79 psi, I think this one goes up to 84. 12 (Slide.) 13' And'if'we go to the other extreme and take a very 14 small particle size -- and there's nothing scientific about 15 the particle size choice -- I start out with a 2/10ths-inch a l 16 diameter particle, went up a' factor of 5, to 1 inch, and down 8 17 o a factor of 5 the other way. 18 [ We go to a very small particle size. You get a s I8
-- l higher pressure spread. This'is for the quench model. !
g.
- 20 r MR. REYNOLDS: Wh'at are'the othe'r contributions, .
21 a besides steam and hydrogen? The-total'seems to be' higher than s the sum of the-two. MR. CZYBULSKIS: That's air in the containment.
.f - ,
c . MR.' CASTLEMAN: . What governs the amount of hydrogen-25-there? Is it governed by-the. surface area? - _ , 1, .~ - . .
; MMa jl- 4 : 3. 32 I MR. CZYBULSKIS: .Let me come back to the hydrogen
*, 2
- for' a moment since you asked the question.
3-MR. SILBERBERG: Excuse me. I would like the 4 speakers to. identify'themselves -- name and affiliation, 5 please'. 6 MR.'CZYBULSKIS: This is -- (Slide.) 8 -- the hydrogen, quantity of hydrogen in the
' containment, in one of the particulate cases.
10
' Basically, what we see is a very rapid increase in II a hydrogen, essentially at the time of head failure. Essential-12 -
yy ,17 of this hydrogen comes from the metal water reaction 13 during the in-vessel phase, and there's relatively little I4 hydrogen produced during this quencing phase,.because the 15 debris is at temperature'for so short a time it doesn't have-2 [ 16 time to react. So, you get a rapid release of-hydrogen when 8 17 e the vessel ~ fails,-very little additional ~ hydrogen until you 3 18 ' f get into the concrete. And then you see the hydrogen come up. 1: 18 - l- I'm trying to' keep this short. Let me depart from I 20'
'e- the figures and just give you a summary of a number of the
~
21 cases that we ran. 5 22 (Slide.) 23 - I apologize -- this;is a very busy slide. Let me
- just try to point out some'of the~ things on here.
> \.L - g_ The first three cases that I.showed you, the P 7,- - - ,,,, v = a s # - s =
g=
,y v- _ .- 4 e , m .: .
^
} ;j >,. . . , , : ~ n ~
71MM jl(4 4D:c -
- 9. .r1 ; ,
_ J -33
}
?~ .
'l' particulate quench.for a one-inch particle, we had a peak-J~
('% - < ' 2J ' pressure of 76..- ForJa 2/10ths-inch particle, we had'84. And m f3l for a:.0'4-inch: particle,;:we hadLa pressure of 94 psi at;this o M 2. 4 - - particular1 assumption.; This,is for.the simple. quenching model. 4 15: . If we;now go-to theDlevitation-model ---which are
. s 8 the'next.tw'oLcases,-
and$in.this case, instead of the parti - ~ 4 L
~ ~
- 'cle: quenching. completely down..as far as.it can go, tit forms 8-a debris' bed.when-the. steam (generation rate drops,'so you
~
f ' don'tijget Lquite as!much quenching: as you do in the first _ case. . J 8\ IAnd instead of a 76 psi,~wethad^74.- Instead-of 84.psii we
~
- ll '
had an-82. psi.- It doesn't' make-an awful lot of difference I 12 whether you use a quenching.or the-levitation'model. 13
) Now, if we go to a' series of calculations-where we
~
14 -
. ignore the levitationiassumptions and.just switch to the _
_15 debris bed on solidification,-you get some interesting results. 1 am .
.g. 16 iThat's not necessarilylmore true than anything else. If.you ,
C
. r.
'take the-small particle ~.and. switch.to debris be'd on particle ,
j 18' size, you-get.76'ps"i;rbut a very-interesting thing happens-
- s
.y ,. .
- 8' '
- l. :
here, that debris bed is not.very coolable and you get'- .for 8 - thatEparticulate . size, you ._get .tihis . kind of pressure-time' r
# - 21 l history.
~
[{ 22 '
.l( Slide.)[
f (23: Instead of'the:real. rapid spike,-you canJsee some:
~
24 : slope -on this . curve. - The~ reason:you see the slope on that-
,,y curve l'[this.-
s p_.
. , .__ __ J- -.
,_ u - : 4..,yu_adm :i v
_ 3 ..
. i, ' , , ,: . w , ,~ . . . . : , , , , 7. , ; , , , - , . aa .,;,a,.--=, , , . ..-;----..
-.--n., ;-., .:L...-.-. v.- :.+i.w al
IG1:jl 4:5 -
-34 1
(Slide.) 2 You get a tremendous amount of hydrogen generation 3 if you-have an uncoolable debris bed, recollect the base
'4 cases down, and here youIh ave twice the hydrogen.
5 If you.try to h'old-the debris bed and it doesn't
'6 quench, basically =you get rapid hydrogen generation.
'7 ~
(Slide .~ ) 8 . MR. GINSBERG:- Ted-Ginsberg, from Brookhaven. 8 ils it true, in all these cases, the interaction is 10 water-limited? In other words, you're dealing only with 11 accumulator water,-and you basically evaporate all the water? 12 : MR. CZYBULSKIS: Right. :In this particular case, 13 there's a finite amount of water available to interact with
'I4 the debris; that's correct.
15 Going back to some of'these cases, let me just-3
- j. -16 point out.one case, the lowest pressure case that we predicted
- 17 was bahically one where we tried to' form a debris bed on 18 I_ sclidification. The debris bed didn't quench. So, I said if E
.! 18 it quenches, lee's not do-the' debris bed calculation. Let's 20 go straight _to the concrete attack, and this is the one that 21 '
leads'to.the' lowest pressure, far' and away, because basically j 22 it didn' t ' evaporate .very much water, . quenched the debris a 23 little bit, formed the bed, and then went on to the concrete 24 attack. I
~ (y : ,
r 4 Then, I went through a series of calcul'ations where 4 . s, h , 3 m ; +,. ..e. ~+ 4*-- . - _ ~ _ ,
MM:jl' 4:6 35
-l' where-I tried to look at what difference to.the in-bed .M 2
assumptions make and what happens later on. 3 I look at the tensile strength of the bottom head 4
'as.it might affect the. timing'of the failure, saw very little effect on the containment pressure. I looked at the amount 6
of structure that came down with the debris bed and-basically 7 multiplied the core support structure by a factor of two, 8-saw minimal effect. 8 I changed the cors-slumping assumptions as, again, 10 there wasn't much of an effect; changed the effect of core-II melting t'emperature, instead of using 4130 nominal inches , I 12 used the UO2 m,elting point -- again, minimal effect; changed 13 the core meltdown model, and it seemed to decrease the pressure a little bit. I And the last case-shown on this slide is -- I tried a j 16 to' combine a number of things.that I thought might contribute 0 17 e to increasing the amount of energy that's contained in the 3 18 { core debris and therefore might increase the containment peak i j 18 pressure. And lo an'd behold, for the same particle size, of I 20 ~ e .2, I went up to 85 psi from the 82 I had with the reference 21 case. 22 gg,. basically, what we are seeing.in' this set of 23 calculations is a relative insensitivity of the peak contain-24 ~ ment pressure to the modeling assumptions. And what it comes L 25 down to is as long as you have a finite- amount of core debris
~
, w m .. . d. ..~m.e-+v . e r w ~~~% * ' "
. f." .
- m
~ , ._ _ m ~. .
,.y... ,, . -
-l
.e ]MM jlf447;~
~
' 'Y -
36. yRf J '
}
- a
'If f iriteracting w'ith ~ a . finitef amount- of water -- which is: the point
-fgr
~ '
t t .
, . 2 .;
jthat Ted made:a'momentlagoi---and;you mix them together,.the-- . 53 , precise (as'sumptions of how:you-mix..them'together are not all
~ --
i4 Lthat important.-.You tend to. transfer.the' energy, generate.the }
.i
- :. 5 ' steamiand get more.'or less the same' pressure response.
- et
.Now, what this:means,.in terms of containment 7
~ failure, I.think that's something that~the Task' Force will .;
a . address. m It's an. ongoing. activity.. '[ 8? i In_ terms.of tihe things'though, the 82 psi-is not a I
' 10 -
.very' strong challenge for the contaiment.
But just to put~
- U ithingslin'some kind of perspective, I would also like to
~
12 present some results -forL.aidifferent containment design at
. 13 -
Surry'.. - I 14 (Slide.) 15 fI Surry is a calculation - . it's actually on a B&W e j '. 2 j
.l- .. 16 :
containment, oc B&W reactor.in a containment-that is almost.
~
.0
- 17 '
1 the same-volume as Surry.is. ~ And'there's some key. differences b 1 1 If'I do the. March 1.1 calculation ik this' design, I get
= - 18 ?here; F
! l
}{.
= ;;
- 18 110 psi; pressure.- - IfLI do the'more or1less.-identical calcula-
. ,r 8 ' tion with March 2, I?get-sbout the same' answer. I : r f [ .,
' 21
- Then,.$if'Iimake'some assumptions'about'the debris f
I f 221 ~ beds, if I don't 1et thejstsel' react',.I get a lower pressure; f A ~ if Iflet.the steel? react-in one.of:these forced-degris bed: 24 ; t 2: 7. . ;..- isituations, I-getla' higher pressure . - 'And it's largely due'to .
- . j '
~8 additionai hydrogen generation. :
7 L
- . a Y
e n < w 5
& .d .i.j q F y. e- + 9WM-^ 4 84. a av. -.
++-ei k. W,- + 4' * * *" *'D " ;
, , , +. + . ~ . :,vi, , n~;~ . , ~ , " '4- ,n ,, J- r , - n-+v.--,s .r, -
- ~-~ ~ r~ ~
~*r n'~' "*"n
LMM jl.4 8~ 37
~
-1 -
.If I don't'let the:corium react, of course I'get.the g,w 2 Llowest pressure around.
3 :So,II will'stop at that point and entertain any
~
4' questions there are. 5 MR. COOPER: Ted Cooper, Harvard. 6- Wh'at was the primary' difference between the two 7' containments that you studied there? 8 MR. CZYBULSKIS: .I guess I gave you1the comparison 8 and didn't really give'you the' contrast. I'm glad you asked 10 .the question. - 11 The key reasons for the difference in the peak 12 contiaininent pressure - there are two or three. One, you 13 6 nil will recall, is'th'at Surry is subatmospheric containment, so V 14 it starts out at 10 psi, absolute, as opposed to 15. So, 15 there's 5 psi.. And if you take that difference and raise the 3 l 16 temperature, it's bigger than that.
' 17 The second difference that's fairly important to 18 the peak pressure that you reach is the water inventory in i~
18 j' the primary system. 5 1 It turns out that Surry has a relatively 'small water 21 inventory. The water inventory affects the containment 5 22 pressure at the-time of the head-failure.
$. Obviously, the 23 higher the' pressure at the time of head failure, the higher .. (,
_ 24 the maximur.t. So, there'is that difference <between the Surry. A
%) 25
~
calculation-and the B&W reactor.
, , .y .s -+ -
, - - - - ---n- , --
MMi31I 49 38
,m 1 The - third dif ference -- and that was the point that - Ac. 2' Ted Ginsberg made during my presentation -- in the case of 3 thelB&W designi you are not limited to the accumulator water.
4 There's' additional water available from overflow in the sump. 5 EMR. COOPER: Would that suggest, then, we might'be 6- able to mitigate certain' kinds of sequences by getting rid 7- . of some of the primary; water, rath'er than trying to put as 8- muchlin there as possible? 9 MR. CZYBULSKIS: I wouldn' t j ump - to that conclusion. 10 MR. KASTENBERG: Bill'Kastenberg, from UCLA. The
- 11. hydrogen acts as an uncondensable in all.of these calcula-
~
12 tions; right? 13 MR. CZYBULSKIS: That is_ correct. 14 ' liR. KASTENBERG: So, explain why, in a case where 15 you had all that hydrogen generation, it tended to spread the a
!- 16
-peak pressure, rather than just raising the peak. Why did it C
17 spread it out over time?
]
I 18 t MR. CZYBULSKIS: It's a question of.how fast you i j 18 are reacting. In the case of the simple quenching model, I 20 _r everything that happens happens very rapidly. In the case of { 2 21
- the debris-bed model, which is the one-that led to the rapid fg - 22 hydrogen production,lhow fast you generate.the steam is limit-23 - ed'by.the debris-bed heat flux,-and you require the steam-to 24
~
, have the reaction with the. water. And that's the debris bed.
25
-- lower heat flux tends to spread the process out in time.
_ _ . . _ _ . _ _ - . _ _ _ _ . - . _ - _ _ _ . _ _ _ _ _ -._.___.___.___.m.__m.-_m___m____ ..-._-__.m _-- _ _ . _ _ _ - _ _ _ . - . _ _ _ _ _ _ _ _ - -_ ._ . _ _ _ _ _ _ . . . - -__-_____--_-_.--__.-_._____-.__-____-_-______--____--%- - _ - - - _ - _ - _ _ _ _ _ _ - - _ - _ _ . _ _ . . _ .
39' MM:jl 4210:
=1 .
So.youlsee some of that' slope that you do not in a simple n.
' i .;.
a 2 - quenching. 3 Incidentally, the point.you'made about hydrog2n 4 acting-strictly as a noncondensable,'in all of these~ cases 5- . your st'eam inerted, so I did'not: talk about possible effects. 6 'Any'other questions?
- 7. (No' response.)
-8 MR.-CXYBULSKIS: Thank you.
MR. SILBERBERG: 5'Thanks, Pete. 8 Gnd t.4 10 [ l 11 ; l 1= c3 14 !
\ ~
, i 5 15 b
ll 16 - I I O 17 ,
- a. ,
2 l 18 r k .l y 19 !
$ I 20 I
-l E. 21 24 25
= se-,#-w - - - , ,. , g g e . ..s
p w y. 40 h. 1 MR. LES: 'My name is Ken Lee from-Battelle l
.D*
, 2 -
Columbus. This morning-we are going' to talk about 3 modification or improvement of the model for removing also
" particle by spraying, especially in the. case of S2D, aus we
'5 -did the lastitime.
t 6. I am going to talt about the possible 7-modification o'f this model. And one of the ques'tions which j' a came up last time was the effect of water droplet size on l 9 ~ removal weignt. So we are' going to address that. 10 (slide) ! 11' As you can see, we have been' using this model' 12- -for modeling the aerosol behavior in the containment. ! The 13-reason we do not -- we do not have anyfwater spraying l l' meenanism in the core. So what we did was to' incorporate ( )
.' 15 l this simple first order aerosol removal rate. Where you l
16 l have the aerosol concentration.with decay in this fashion i 17 where N is the number of spray drop concent' ration and this 1a l is the velocity of the water-drop.. Small g is the settling 19 l velocity of aerosol particles. 20 Because of the hydrodynamic interaction between 21 l the aersol particle ana. water drop,.there is a. collision ! 22 called fission. And at that time we have the two , 1 23 mechanisms. The first is inertial impaction effect, and the 24 secondEbeing the intersection mechanism. And'of course, 25., Stokes number /in the inertial impaction._It is written like (' , TAYLDE ASSO CI A TES 16 2 5 I Stre e t, N. W . - S uit e 10 0 4
. w sshington, D.C. 20006 (202) 293 3950
41 1-this (indicating). 2- one of the prob'lems we found at that time was 3 really'this measured impaction effect is not all tnat great, simply'because particle size we are dealing with is 5' on the. order of l' micron. Water droplet size ranges anywhere_between 500 -- 400 to 1,000. microns. 7-It turns out that-this comes from the model 8 wnich incorporates the Stokes flow, but the water droplet 9 size in our case is large. The assumption of the Stokes 10 flow probabfy is not accurate, because of this increasing 11 Reynolds numberieffect. 12 Bonded layer around.the water drop might be a 13 little bit thinner than what we thought it would be, wnich 14 means particles could penetrate a little bit farther toward-l/ 15 the water drop. , 16 (Slide)^ 17 This is terminal settling velocity as a la
. function of particle size., This is the predicted terminal 19 velocity using the Stokes flow model. And you can see if 20 you go up to higher --~ the quality of this slide is not all 21- that great. This is 1,000 micron right here, and you have 22 100 micron. Previously, we were tal.<ing aoout the size l 23 range of the water drop, this range. And obviously,-if you 24 draw a straight line, obviously there is a deviation 25 between the predicted terminal velocity and the actual -l - TAYLCE ASSO CIA TES 162 5 ]I Street, N. W . - Suit'e 1004 w ashington, 0.C. 20006 (202) 293 3950 g .. . ~ . _ _ . . _ . . . -
t.
42
-!1 terminal settling velocity, which means the Stokes flow y ~. _
2 model applies. 3 So I think wnat-I am going to do is to employ an
- inertial impaction mechanism, which is based on laminar or 5
potential terminal flow regime-or mayce we can go up to a 6-potential flow model so'that we can properly incorporate 7
'this inertial impaction effect in this, e
(Slide) 9
-What I have here is collision efficiency as a 10 function of proper size. In this calculation, what I'have 11 is only intersection effect. And tne first thing you notice-12 is that collision efficiency is-real small simply because 13
.tne size ratio by errors of particle to the water drop is 18 small, but nevertheless in this collision efficiency
.O 15 increases rather rapidly as the droplet size decreases. So 16 in this case, it's.in inverse proportion to the scale of 17 water droplet diameter.
18 Now that chart is just the collision efficiency. l' But with a fixea amount of water available, if you have a 20 small droplet size, then you are creating more surface 21 area,.which means that your removal rate will probably 22 increase. 23 So at this point we realize that the effect of 24 aroplet'. size is ratner dramatic. The last time, I think we 25 used a' droplet diameter of 1,000 micron, and some of you [ , TAYLDE ASSO CIA TES 16 2 5 i Stre et, N. W . - Suite 1004 - W ashington, D.C. 20006
~
(202) 293-3950
,. . , , . a. n... . -.-
7--, 43 1-
. _ 'didn't agree with that 1,000 micron. So I think the next S. . 2 time around, we are going'to -- well, first of.all, we are 3-going to improve the model, and then we are going to incorporate this couple of droplet size, water droplet 5 size, and see what kind of effect that will nave on the overall' concentration decay in the case of S2D.
7 (Slide). 8 Anotner question which came up at that time was whether the model was capable of accommodating the 13 existence of a two water flow rate', one being injection
' 11 ~
spray rate and then about 5 minutes later there will be 12 recirculating pump operating..And then both of them will go 13 on until containment fails in the case of diameter. l' otnerwise, that injection pump will turn-on at the time of 15 20 minutes, will continue operating until water runs out. 16 In our new model we have to incorporate these 17 two separate flow rates, so now we can handle this'in la containment injection pump flow rate of about 3,200 gallons l' a minute. And subsequent recirculation pump flow rate of 20 3,500 gallons per minute. 21 (Slide) 22 so we are reexamining this collision mechanisms 23 we have in the spraying model,'and we are~ going to use this 24 correct terminal settling velocity especially for large 25 droplets. We are censidering.to incorporate another i (j ~
-fAYLOC ASSO CIA TES 162 5 I Street, N.W. - Suite 1004 w ashington, 0.C. 20006 (202) 293-3950
44 1 collection mechanism for small particles. I don't think the E 2 effect-of this will be dramatic, but for small particles l-3 Brownian diffusion might contribute to the collection
' mech'anisms of these aerosol particles by spraying droplets.
5- And.of course, this is, as I said, two separate 6 i spraying flow rate. l ~7' Again, we do not have any removal mechanism due 8
- to diffusion freezes. And that question came up.
(Slice) 10 So we went back and made some calcuations with 11 diffusiophoresis mechanism and another one without 12 diffusiopnaresis as a-removal meenanism. This case happens 13 to be an AB delta. I don't know whether you remember all I l' [ < ,s, tais or not, but this'is D2 metal release,'and after that l \ I ll l
"- 15 th'is is the aerosol particle release during the t
l 16 vaporization release. 17 It. turned out that diffusiopharesis is not all la that important. There are a couple of reasons why they turn l 19 out not to be important. First of all, I think the water 23 surface area is not too large compared to the containment-21 volume, and then thermal hydraulic conditions are such that 22 the wall temperature might be cooler than gas temperatures. ! 23
'But as time goes on, the wall particularly picks up the gas
.2" temperatures, so that you really don't have a substantial 25 temperature difference between the containment atmosphere l
~
TAYLDE ASSO CIA TES 1625 I Street, N.w. - Suite 1004-W ashington, D.C. 20006 (202) 293-3950 l L-
r-. . . . = - ... . . - - - . . . . . .- - - ----- .-. - 45
,, 7 1
-and-the wall'.
f: t- 2 Really, that's all I have to say today.
~
3 MR. SILBERBERG: Ken, I have a question. When you say the wall surface area is limited,~ do you include the 5-surface area prov'ided by other internal structures within
'tne containment?
7-MR. LEE: Last time what we did.was we used a 8 single volume, but I don't'thinx we included all of the l available surface area. 10 MR. SIL8ERBERG: If you did, do you think it 11 would still make a dif ference? 12
!st. LEE: According to this result, I still don't 13 think it will make a lot.of difference. But I have to 14 mention that this mechanism is now-in this, so next time 15 -
arouna it will be --'all the calculations will include this 16 mechanism. 17' MR. CASTLEMAN: Will Castleman, Penn State. I 18 have a question about ~the collision deficiency. Are you 17 saying you think the expression is correct.that the main 20 problem nas to do with the Stokes settling velocity, or are 21 you questioning tne overall relationsnip for the collision 22 . efficiency? 23 MR. LEE: The expression we had-is the one we had
-24 the last time. But I think we are. going to adopt a new one 25
.wnich will be more realistic, because the water droplet
' TAYLDE ASSO CI A T ES 1625 I Street, N.W. - Suite 1004 W ashing ton, 0.C . 20006 (232) 293 3950 .s
a- . - ,- . .- . .. .a . . . . . .
.46.
I size is so big that this expression, which is based on the
-Q 's. 2 Stokes flow model, won't be any good. So we're going to 3' have a new expression with that, and I think that'will increase-the collision efficiency. .
5 MR. CASTLEMAN: The new expression will be the 6 diffusion of the extreme temperatures around the droplet? 7 MR. LEE: I think right now we're looking for a 8 simple expression, which can get into -- we're not going to, start from scratch. I think if we can come up with an 10 ~ analytic solution similar to this -- I know there is one 11 expression wnich is good for the potential. flow regime. But 12 if we can find an expression which can cover between this 13 one and the potential flow regime, I think that's what l' we're going to have.
.O 15 MR. COOPER: Doug Cooper, Harvard.
16 From tne scrubber literature, the~Walton-Wilcock 17 data that has been correlated within impaction parameters 18 wnere over an impaction parameter plus a constant squared, 19 that's probably good toward the potential flow of the 20 high-velocity regime. 21 One of the tnings that's going to. increase your 22 velocity and should be.taken into account is the pressure 23 of-the sprays. If you have even 1 psig on the spray, I 24 think you are going to get a few meters per second initial 25' velocity of the drops. And because they have a substantial r .
-("j . T A Y L O E. ASSO CIA TES 1625 I Street, N.W. - Suite 1J04 d ashington, 0.C. 20006 (202) 293-3950
p . r .% _ . - .- s . - - - _ - . - - - Y $1 '47 I' mass, Ihey have a substantial characteristic-time or
~
L 1 2' stopping distance. .So that: initial velocity could carry 3 through a substantial portion of the containment.
~"
MR. LEE:: Right. I am aware of that possibility 5 also.. I think at one time we.just' assumed the droplet on , the terminal setting lies there. Agait.. I guess it Just 7 depends'on what kind of droplet size we are going to use. 8 Another thing I might add, like you mentioned, . 1
~
i .tnis expression is such that you have got-to have a Stokes t- !~ 10 number larger than 1.214 essentially, and that's kind of 11 hard to have. That might be true for the case of, again, 12 ' Stokes flow. -But I think'if you-go to this potential flow 13 regime, you really don't have to have a Stokes number -- l' MR. COOPER: That's right. There's no critical O- 15 Stokes number really in the other one. l ! 16 The other thing I am concerned about is the 17 diffusiopharesis, which I mentioned in the memo to'Mel. But is as you pointed out, it's.not so much a collection by the 19 droplets, because often that'S' offset by the 2'3 enermopnaresis. But this business of flow to the walls I 21
- think needs yet some more study, l
22
-In the scrubber literature it's been found,.for 23 l example, if.you condense roughly a quarter of the vapor, if 2" the gas is roughly a quarter water vapor and you condense 25 pretty much all of that on the walls, you take out about a l
( TAYL0E A SSO CI A TES E 1625 i Street, N.W. - Suite 1004 W ashington, D.C. 20006 (202) 293-3950
x, :. - 48 i ? 1 quarter of the aerosol concentration. 2 MR. LEE: Exactly. 3-MR. COOPER:- I just believe in the physical 4 situation that there's - going to be an awful lot of water l- 5' vapor produced.
- 6 The calculation I ran th' rough for the Surry 7
. situation suggested 5,000 pounds would give you almost a e
containment volume's worth of water vapor, and that should,
'- be condensing on~the walls, running down the side of.the 13
( walls, being heateu up again. _I think that's going to give 11 a lot of collection. Ana I was surprised by the results of 12 i the first calculation. It may-be we're not getting enough k 13' l heat conduction out of the containment. L 14 MR. LEE: Right. I think.what happens is cefore l 15 the coremelting starts; you get a lot of condensation in l 16 the containment. But we don't have an efficient product at 17 that point. Actually, that was one thing we found.- - l 18 The second thing we thougth we had -- we 19 consider this condensation of water steam into particles, j 23 ano you have to compete with -- I mean this 21 diffusiopharesis has got to compete with that condensation l 22 of the water vapors and particulates. But!again, this 23 model does not calculate 1the. water vapor concentration in 24 the vicinity ofLthe wall. We~just take the condensation l 25 rate.which is calculated by the thermal nydraulle cause. l J\ '~~ , TAYLOC ASSO CIA TES 162 5 I Stre et, N. W . . Suite 1004 d ashington, D.C. 20006 (202) 29J-3950 E - -- __n_.___ _ _ _ _ _
a ** m 7, .. .
- .. . a M, %
;4 i t t ,-
~
s ,s
-49 s s i s' i'
'I r
1-p . So that's what we have found so far.
- t. 2 ~
But'I-agree-with you, s f: 3
.MR.'COOPERr, The fina[ point I wanted to make is f.
v; y ,, i
- v
.that at th'e'beginning or.ca; lean we should.get' adiabatic 5
-expansion in t.his chamber, and that should lead to real
,s ~1 n
- supersaturation and sernaps enhanced particle growth and, '
tnerefore, enhanced particle deposition. j u 8 MR. LES: . Right.. 9 MR. WILLIAMS: David Williams, Sandia{
~
k ,g . 10-The comments I sent in on the S2D at:alysis are f I 11 g T i* 4 on the whole tWing. ; IlMoted ! I had a great deal di ' . , . ' 12 difficulty rconciling 'the expression being used tf.dee for 3 13 spray efficiency with the results that were coded for S2D ' y ' s . , N 14
. delta or S2D sequence both'for gamma and delta. *s i 4 15
\
MR. LEE: S2D-gamma and delta. Okay, s + 16
.MR . W'ILLIAMS: In particular, the? suspended 17 r '
-aerosol condentrations in the S2D delta sequence that were i
Id quoted in one of the figures indicated a particle residence l' kind of no more than a few seconds, whereas as you noted in 20 that expression for tM impaction, fficiency, that term is
< 3 21 extremely small fors particles of the order of a few microns 22 i -I
-- no, the interceptio. .1 term is extremely small; the a
23 impaction' term-doesn't come in at.all for Stokes-diameters 24 v.. less than 1.'214. t O 25 And yet r/pe actual re'sults quoted seem to-imply
, r < ,
v ~l ' Q
,a
)(
3 a1 (, ~ .. T AJL OE , ASSO CIA.f ES ' 'N
~
- 162 5 I Street, N. W . - Suite 100 4 '
'N W ashington, D.C. 20006
^
(202) 293 3950
; q, y 3.
-l . . ,- -+
", k'
^
Ms- u __ _ . _ - J. - .
- p su
& il -a;1arge collection efficiency have you looked into that at
& ^'& v 2
. _ _ all?-
3- -MR. LEE: Right. Can you identify yourself once jd; *. 4 again? b . 5' 'MR. WILLIAMS: David Williams, Sandia. s-MR. LEE: Rignt. I think I read your comment, and 7 you are' precisely right. I.think this expression is, like I
- y. -
8 said, is'not correct. 'One thing, because of the flow model 9 .I just talked about, and~the second one, as you pointed to out, this began.if Stokes number becomes less -- well, I 11-guess you're-tal<ing about two . takes, and it-really should h . > 12 have been suppressed-for the regime where the Stokes number 13
- - is less than 1.214. So that was a mistake.
14 I think I read your comment, and you were l 15 precisely rignt. s e 16 MR. WILLIAMS: Thank you. That clarifics my
.\
17 concera.
.p d la ~
MR. RITZMAN: -Bob Ritzman from SAI. N[; 19 Did I understand you, Ken, to say when you redo 20 tnis again you'are going to use a spray droplet size L 21 distribution? i '22 MR. LBE: -No. [h 23' MR. RITZMAN: You're not going to-use a 24 1,000-micron spray drop size?
.25 MR. FLEE: I think'we're probably. going to use 400
- ; TAYLOC ASSOCI A TES
- ('%' ' ' 162 5 I Stre et, N. W . - Suite 100 4
.v i W ashington, D.C. 20006 -
(202) 293 3950
- y Y;
j;~ g - - . . . . . . ~ , - . . - - - - . . - + , ~
r- . _ _ . . . _ , . . . _- ~ . . .. .. __, -. _ z . , 51 l or 500.
- e. But I think I will have to ask your opinion or the f-t- ,
2 opinion from -- I~think'the last time we used 1,000 micron, 3 there was a-lot of objection to the size of the water L
- droplet. But we are not going to model a size dispersion 5
~
of.a. water drop. I aon't think it's rea11y worth it. It's just a matter of.which number you pick. Probably surface is 7
.a million diameter water droplet. .
8
.MR. COOPER: Typically, the sortlof mean diameter 9
surface volume ratio. 10 MR. ROE: Donald Roe, Roe & Associates. l-11 Could you-describe how you are handling the 12
' condensation and heat transfer at.the wall?
13 MR. LEE: No, I didn't' describe that. l' MR. ROE: Would you describe briefly how it's 0 15
-done, what level of modeling is included?'
l' MR. LEE: We just take.the condensation rate 17 provided by the thermal hydraulic code and use that rate to 4 I la calculate the removal rate. l' MR. ROE: How is that condensation rate n 20 determined? What's in the thermal.nydraulic modeling? 21 Rich, maybe you could answer that. 22 MR. DENNING: We.use the MARCH code, and that nas 23 empirical-neat transfer correlations of the type, the 24 Togomi. type of correlations. 2 5 --- MR. ROE: Were you considering wall heat (L ~ T A Y L O .E ASSO CIA TES 162 5 ' I Street, N. W . - Suite 1004 w ashington, 0.C. 20006 ' (202) 293-3950 A
.a .- .-~ :-
(
- .~ . ,
=
...+;.... .
_ . . . . . . . - ~ . . . . . . . . .m.-._ gc . < . . .
- s. -
V *
-a a p ~-
. 52.
-- 1 ci .capac ity?,
L [f ' i 2 .
~
MR. DENNING: .Yes.We'are'modeling the thickness 31
'of.'the wallsand heat. transfer into th<a walls.
~* s MR.SILBERBERG - Thanks, Ken.
5 6 7
~8 9-10~
- 11'
- 12 13
- 14
'a'g -
15-
- 16
- g7 .,
i d ';
' 19.
20' 21
- 22
-23 24
'25
-\r;; -
-{'
-T A Y L D E ASSO CI A T ES l16 2 5 I Stre et, . %. W'.- - Suite 10 0 4 -
~
W ashington, D.C. 20006 . X ;(202) 293-3950.
-D 2
^ gee '
1
~*f;, - >
.._.<.-.,,.-.,.._uma.-- ;,+ -
__ z-j; , , ,+, r --- .x
__. _- - _=_ . _ - -___ . - _ . _ _ _ _ _ __ _ _ _
. h )1 a3 fd '
l' MR. ~ DENNING: I am Rich Denning, and I am going T
~ ' D'. 2 to'be ciscussing a small study that we did to look at the
'3 effects of decay heat'that is deposited on the surfaces of
- tne reactor coolant system and its effect.on.temperaturea.
5 (Slide) As you will recall, the MERGE analyses that look 7 at:the reactor coolant system temperatures as a function of 8 time are cone:quite independently frcm the TRAP analyses, and the flows anc temperatures that are predictea by MERGE 10 are then input into TRAP. 11 TRAP-predicts [ deposition of fission products on~ 12 the reactor' coolant surfaces associated with those fission 13 products as decay heat. But the effect of that decay neat 14-is not then cycled back'into the MERGE analysis to 15 oetermine: what tne effect is on the temperatures and the 16 lows. 17 he are definitely--not intere ted:in coupling the la MARCH, MERGE, and TRAP analyses,'but we recognize that in 19 order to do this problem right, that maybe eventually what
.20 is required. So what we wanted to do is to look and see 21
,.what-is the effect by going through~a single iteration of 22 MERGE and TRAP analyses.
23 And I think'that there are basically two issues 24 that are of' concern here. The first one is how important is 25 the-feedback of the decay heat from the fission products TAYLOE ASSO CIA TES ( 162 5 I Stre et, N. W . - Suite 1004 W ashington, 0.C. 20006 (202) 293 3950
. -w% _ .# . .y _ . .m- i ._ 2 . - - ~ , - - . + - ----a--------- - 'c - - *
. [ ,
v , - N. ' ' W ,
~
4 il ~during the time period offour TRAP analysis on changing.the 2L thermal nydraulic conditions within the reactor coolant-
- ) ; system. .. ,
Then theremis a very' closely!related question
~
- - 4: 5 _ -but a1very.important one-that says, what happens in the' '
- . .'? long termLif we put fission-products.on. surfaces during the
- t 7
Lperiod wnen welhave flow in the reactor coolant system? 1 8
- 'What.we do.is basically-we just forget about them-then.- -
'9-aut there really;isLa. question,7what happens'to l
' 10 - lthemLin:the long' term? After we move.on to'the-other phases.
l11 :og '. the = accident, is it possible.tnat these. fission products
~ 12' will.cause~ surfaces to-continue to heat?: Is it possible 13- fission.productsiwill be reevolved?':Is it possible we could 1
! set up." strong-convection patterns'within.the-vessel-'that
' ~
. 14' 15
{- Lbring air.into th'e vessel, change theLchemical' form of-the 16 -fission ;procuct? ! 17 So basically,-these are..the two issues that we r la are trying-to address ~. Thefapproach is"to adapt MERGE-so..we
' 19 . can ' include . fission. product decay heat source term in ;
calculating the structure. temperatures,..and then-to perform ' 211 -one' set of iterations, MERGE run, TRAP run, take the
- - 22 ' fission products from the TRAP;run', putithoselback into s
23- fMERGE and run again and'see if we'.see a'significant
, 24 ..~ dif ference between' the- temperatures of the two MERGE runs,
{.
'do!we'see-~a;sihnificant-. difference in the' deposition in-the 2
.a u
~ '
?) s i A' Y L' O E ASSO CIA T ES -
1625 I Street, N.W. ' Suite 71004 -
, W ashington, D.C. 20006 b '
. , (202) 293-3950.
m
.W P'm W i k 'e a fC' *M _M h d -
# I '
,79 w - g y *
'N-'
4-.w a .,y , , .- g. -- .~.--:-.-a -n x
,. c ,g ,e W )
+
_ .;. s:7 _ ,
- 37
' 5 5' l ll- :two TRAP' runs?r
. Ef7 .2; :(Sl'ide) '
' 5. ^
~ 3' The accidentfsequence for which we did this was 4 the.S2Djepsilonicase,' orkthe S2D case. 1 It doesn't matter
.5' what:th}e eontainment failure mode is;here..I just-wanted to ;
E , review.for you what.the flow path is in this sequence. J 7
-ive have-assumed a small break in the cold leg, a
~
aE flow rath that<go'esLfrom the core into the upper plenum.
; region down'a hot leg through?the steam _ generator. That's-10 impor ta n t".- The steam generator not only.is massive, but-in 11- this-particular. case could'~very well-nave active cooling.
12
. go.the fission product decayJheatIthat;goes into here'could ;
-13' be. carried away quite easily-and:then out to the-14 ~ containment. I i
15- (Slide)
- 16 In order.to~do this, we did a series of UtIGEN 17~
calculations for-three burnups corresponding,to three l 18 different' core. regions: 11,000, 22,000, and 33,000 19- . megawatt-days per metric-ton. 20-TnenLwe-looked ~at-different. times after 21.: shutdown. We actually added-the J three core regions back
- 22 ~ together again. We didn't look locally to see what the 23 effects were. i 2'
And we" grouped by major groups the noble gases:
. 2 5. . :
an iodine group, cesium group, tellurium group, and then 7
~il TAYLGE ASSO CIA TES ~
16 2 5 ' I S tre e t, N. W . - S uit e ~ 10 0 4 W ashington, D.C. 20006
;(202) 293-3950
. _x .p m,- .y . .- .. . m-< -
. := ~ A-- ~~ =~~.;-G---- " - &- < J-1 56 t
.. .f ili s everything~else: thrown into a balance.-This group'is really
;f ' .
1 _ . 2- represented by aerosol, 1 3: 'And'we looked at ~ the fraction of the decay. heat thatsis associated with the full-core: inventory of each of e 5 these at.these differentishutdownLtimes. .And then this 6' table ~is'put into'the. MERGE code. ;And then we looked. We 7
~
-did:a MERGE analysis,.a TRAP analysis. And we' looked to see u -how many fission' products did-we.have deposited in the 9' ' upper plenum.
10'
~
. ( S lide )
11 '- .I am going to'focu's here.on the upper plenum. J12 TheLthree: major? regions ford'eposition here-could be~the ' 13- . upper' plenum, the hot leg, and steam generator.- But I am 14 .really-going to. focus.here'in looking at the-upper plenum.
. -9I 15 This shows noble gases are actually resident in
- 16 'the region and not onfthe;' structures. 'This is the TRAP 17 analysis, fraction of; core inventory. You will notice the 18 - ' iodine which was transported in-the form of' cesium iodide 19 initially started-to deposit on surfaces,. cut then as the 23.- surfaces heated up, was driven back off.
21 The cesium wnich was primarily deposited as 22 cesium hydroxide, some of.it. deposited as aerosol, the 23 - tellurium, and then the balance. 24 Now, if:you7took'thisLfigure and the'last. figure 25 and multiplied.the fractions of-core inventory times the
- ( j~ TAYLDE ASSO CIA T ES 1625 I Street, N.W. - Suite 1004 W ashington, 0.C. 20006
-(202) 293-3950.
_ .a, w,.:.: _
. . . . :. _ . .g ,
~~y 57 1
-~.
heats,fwhat'you would see is at the end of this time period t 2 most of the heat-is; coming from the tellurium.and thei
~
3 aerosol, a little bit more from the tellurium than from the
~4 aerosols.
5 The tellurium, incidentally, you will see that
~
6 tnis behavior for the BWRs'that we will get into later
'7 'today-may change substantially. We probably won't see as 8
much' tellurium deposited in the primary system. . 19- :-But in this case, most of the heat in the IJ structures at this time at the end of the period is. coming 11 from-the tellurtum,-almost as much as is coming from the 12 bulk of the aerosols. 13 (Slide) la
.The representation for the upper plenum, if you t';
15 recall, for the Surry PWR in its first volume draft, we did 16
- no nave' good representation of data for the upper plenum
~
~17 structures. We subsequently have that. But this study was 18 done before we had that.
19 The case that we have used is what we have i 20 called the hot upper plenum case, which minimized, was at 21 ene-lower bound of wnat we thought the surface area and 22 structure mass in the upper plenum was. 23 So the effect that we see here may be 24' overemphasized over reality, although'I don't think 25 greatly. The maximum it could'be would be a factor of 4. j- TAYLOE ASSO CIA TES 16 2 5 I Street, N.W. - Suite 1004 W ashington, D.C. 20006 (202) 293-3950
.__m._ ____._____.2_ __ $_____. _ .- . _ _ _ _ _ _ _ _ _ _ . _ . _ _ _ _ _ -____.__.__.E #
Q: L. +
._ =L 4 x.= --.: - _ . - -- . - . . . - . . _ . . - - :
58 1 But there might be some overemphasis of this effect. [' ' 'h 2 This~is the temperatures that we see in the
.3 ' upper plenum without the fission product decay heat and 4
with: the fission product ' decay neat.- And at the end here,
'5
-where we are slumping.into the lower plenum, we have a 6 difference of about 400 degrees Fahrenheit, which is not 7- insignificant. And'the; rate of heating of the structures is 8 fairly significant here due to fission products.
9 (Slide) 10 The effecr of that on the fission products - is 11 shown in the final slide. ' Recall that this particular -- 12 other than-the tellurium,.there was'not a great deal of 13 retention in this particular case. There is virtually no l'
'effect on the aerosol deposition, whien is a major j
15 contributor to the heat source.- 16 There was-enhanced deposition of the 17 higher-volatility materials on the. aerosol particles. That la is,'instead of depositing on the surfaces, they deposited s 19 on the particles. 23 Ana we saw some change in the' location of the 21 ceposition, and we also saw a change in the deposition 22 mechanisms. Much more of the' deposition was happening 23 associated wit". the aerosol deposition rather than the 2" direct vapor deposition. 25 As far as the total primary system behavior, we (> TAYLOE ASSO CIA TES 1625 I street, N.W. - Suite 1004 W ashington, D.C. 20006 l (202) 293-3950 ' l l
.mn} r [ 3;
~.
gw 9
,. ~ q.+
-t
-=-=- -
- s-
-~, ~
~ = - - -
i y
~
59-l (;g 21
- ' sawJsome :significant .-- well, quite signifi' cant on the
~M'
~
2] ' cesium iodide ef tect .on the retention. LWe saw only about
~
~
& ?3' halffas:much retention ~L. J s
' 4.- fBut in that particular case, we'had very little 25 ; retention'anyway. 'So;I am.not'sure how meaningful'that .
particular1 amount is...A small!.effect.on the tellurium. 7 LS mallteffect on the i cesium hydroxide.- 8'
-Our conclusions from this are that this could.be
~
~
a significant effect.. 'Weido n'ot: plan to adapt our,models t
-10 forithe analyses.thatiwe.are performingIfor the last three
- cases.-~That would:be'a: major-undertaking.
12 - We do, however,;think that.it'does raise, 13
~
~
continue to-raise, significant questions-:about-the:
- - - - - ~_ . . 14 L long-term behavior of-~ fission products.that are deposited
-y3 E ~
- 15. ~ -in the reactor cooling ~ system. I.think this would beia l'
, major consideration in~the sensitivity studies'.that are 17 going to be initiated thisisummer.. -
18 1-will.be glad to. answer an'y questions. 19 -MR. WALKER: Dee halXer'from hestinghouse. 20' On those numbers,-you had about 60 percent of 21 tne tellurium on the' surfaces and 2 percent of the 22 aerosols. What if you got 3 or 4 percent of core' decay heat 23 on the surfaces?
'24' 'MR. DENNING: .Right. 2 percent of each of'those.
l25 MR. WALKER: ~ Not very much, actually. sy. ,
$~.):~ T A Y t. 0 E ' ASSO CIA TES
~ 1625 I Street, N.W . - Suite 1004' W ashington, D.C. 20006
' (202) 29 3-3953 ~
, , .--,:-.,- i . . - - ..- ~ ~ .- ,
p _ - ---- , .- ' 6u
- 1 ~-MR. DENNING: Not very much, actually, but enough p
s . 2 to-have a fairly significant effect on the heating of_the 3 structures. And also,. recognizing the uncertainties in our
- depositions are significant, if we were to get into'a 5
regime -- and tor example, this afternoon'you will see a 61 ' case'where we get some' fairly significant. deposition of 7- iocine on the upper plenum structures.-- we could get into u regimes that involve quite'a bit more heating than that. 9 MR. COOPER: Your description made me think of 10 something. When you get the vapors onto the particles, then 11 - you deposit particle plus vapor on the wall. 12 Does the code keep track of the fact that it 13 still has material that is potentially in vapor form if it 14- gets heated again, or does it.take it'now and connect it to O k 15
~
the particles in a way that won't come up? 16 MR. DENNING: It'will allow it, if it condensed 17 on the part'icles, it will allow it'to vaporize off the id particles. 19 gg, coopgg Good, 20 MR. VOGEL - Vogel, EPRI. 21 Tne relative importance of the fission product 22
. heating depends.upon what you have assumed with regard to
'23- convective heat transfer-from the core to the; upper plenum.
24' I assume, since at the moment ~that is sort of an open 25 question, that you probably didn't assume very much. Is
, TAYLDE ASSO CIA TES 162 5 I Street, N. W. - Suite 1004 j W ashington, D.C. 20006 (202) 293-3950 ,
I i
y e-m ,. +
? m;. -
- - e 7v v 7 m-~m ~ -- ~-~-m---' s="~- ~ " " -~ -
, j '61
< ~
1 y ,
-I that incorrect or'what? .
I ,1 1 4_ .%; f 2 MR.' DENNING:/ Let me answer the' question. I-am - i not surefthAt:I completely un'derstood it. 3 ? ' ' s - I think there-is t
- O 4 c almajor-iss'ue'that.regards-how much natural convection-exists l'n the upper ~ plenum, particularly'in the period
~
--J 5 J'? Lafter we,meltithrough the lower head of the vessel. And-in
~7, theilong. term, the tnermal: hydraulic condition:in tne 8
ll : vessel.-is going to be.an" extremely difficult problemito , , _ resolve:LWhate.is really happening.in the vessel in the'long 10
~ term, howimuch convective cooling is there of surfaces?
i.
-11' '
In the'~ shorter term,:what we assume -- whattwe ;
w , t 12- .did'Was we nad-flows driven from volume to volume-by the.-
'13 Igeneration~ rate of, material. . Within:the upper plenum-the +
t .-- l'
-heat:tran.sfer structures due.to' natural convection was '
l 15 based upon natural convection correlations based upon L. - 16 GrasnofLnumber types of' considerations.-
~
17~ MR. VRLKER: One more. In the TMLB'-sequence, tne L . Y .la small-break. sequences, you have' a- fairly short time between i o < .19 1the introduction of' material on the lower head and the l ~ 23 vesse1 melt-through. 21 I guess my question is: .Is that a result of the' 122.- code, is-it something-you really~believe, or in.Surry is it 23 a result of.not having much water..down because'of the t
;24 atypicality of the lower-end region because of'the steam 25, sweep section subsequent to core slump?
l. r l -- , t 7 , TA LOE ASSO CIA TES
- T 162 5 ! Street, N.'W - Suite 1004 l
W ash' i ngton, D.C.: 20006 L(202) 293 3950 ..;
.[ p g *
-,,,-Mq
@, f .e - De m &4.. 1.;4 * * + .'-'*8- y b $ % F 41DrF U
- k . - - - f rr u -* * -e a - r "v -
-wr v ru --N 'm e vw- rww,- 6
..__, _ , _ = _
w
, 3-
.a. : 3 p =-
na -p a:. - - . ..a . . L w - ... . -
.e-- - .. ,
m- , . 62 -
~
. 2. . 11 'MR. : DENNING::- With' regard to thel pressurized B
I((i ~ Ecases,:-like-TMLB',=I---think there is-little uncertainty that
.2 r -
- 3~
, shortly ~after the core falls into the lower plenum, you 4- -
wil'l-Lfail the lower, head because of1 the internal; pressure l '+ - 5 ~ and the-strength. YouJreally don't h' ave the melt through.the lower
- h'ead .1 ' Alliyou have;'tojdo is really heat.it up. So I:think
~
, 7,
~
- S.
~
there;is very-little. uncertainty in..those particular cases.
^9~ '
^
[ 'There is;more uncertainty in the-unpressurized-t L 10 cases as to just whatithe duration.will'oe,.'and you will t x 11- see this~ afternoon or this morning when we talk about BWRs
.. t 112 quite a different predictive behavior-than for those 13 particular sequences and quite an effact on the types of 7 14' tne periods of release, the processes that are going;on
- 15 'while the core is in the lower plenum region _.
~
' 16 ' MR. WALKER:
You'know, itjjust seems.to me it's.
-j
- 17. difficult to convince yourse1f you will.have.a. coherent -
'18 enough slump that you will[i drop a whole bunch of-material i-19 down there and drop through the bottom head in a hurry. It
. 20 seems to me you will-drip' material cown there and have a l 21 fairly:significant steam sweep _before you get through that 122.- head ifSthere is any water at all-down there. Maybe that's-F 23 just an-impression.. ' i- 24 MR. DENNING: Yo'r u questions are certainly good P 25 ones about do you.really. dump the material down in the I , y%. Ff TAYLOC ASSO CIA TES- , 1625 I Street, N.W. - Suite 1004 3, W ashington, D.C. 20006 (202) 293f3950 7m ,_ -- ~ = . . . .
4
~63 +
- l. ilower plenum quickly or does it more gradually-slump into
-(. 2 tne lower. plenum. ,
3
.In the reanalysis we are going to do, we are going to be looking much more at gradual slumping type of 5
-models and the effect of enat.
' In audition,'you will see that in the MERGE 7
' analyses and TRAP analys,es we have'done for the BWR, we 8- nave considered this period of lower plenum behavior'much 9 better'than we did in the-surry analyses. We go back to the 10 surry analyses; we will oe looking more at what happens 11- when we are producing this rapid generation rate of steam 12 at the end of this period in vessel.
13 .Pete, aid you nave something you wanted to add
'l* to that?
C:,') 15 MR. . CZYBULS KIS : Peter Czybulskis, Battelle. 16 I Just wanted to amplify your' comments that, in-17' fact, in most of the things that we are doing today, we are 18 shying away from the coherent slump down in.tne lower 19 plenum all at once. 20 In particular, for example, the.Surry 21 calculations that I discussed, even though they were 22 addressing what happens in the after-vessel head failure, . 23 we use.wnet I call the gradual slump model. We have taken
.2" into account the holdup of the d'bris on-the support 25 - . structures and then-the stuff-tua.ing into the head
( TAYLOE ASS O CIA T ES 162 5 I Street, N.W . - Suite 1004 W ashington, D.C. 2JJ06 (202) 293-3950
e ._a _ _ .....a_ _. ._.
~
~ .. 1 : J.
~~
~'~ i
{ 64 1 . subsequently.-
-2 lit just turns out there-isn't an awful lot of
-3 delay in terms of the' attachment of the head, as Rick 4
pointed out correctly, in.the case like TMLB' where you 5 operate'at 2200 or 2300 psi. It Just doesn't take'very long 6' to heat the head to a point where there's very little
-7' structure.
8 'MR. REYNOLDS:- Reynolds, GBA. 9 Why are you considering more and more the i 10 gradual slump as opposed to all of it coming out at once? 11 Is there any experimental data or any evidence to move in l 12 that direction? l 13 MR. DENNING: That's a matter of interpretation. 14 There are people that read into the behavior of out-of-pile 15 experiments at KFK that you do have that kind of slumping, 16 more of tnat kind of slumping behavior rather than more 17 . coherent drop. I 18 I think that intuitively we tend to believe that 19 l the gradual slumping behavior is more realistic than the 2J sudden dropping. But you raise an excellent question about i
- 21 the state of validation and understanding of these.
22 And a point that I would like to make as often 23 as I can today is the sensitivity of the results we get to 2' modeling assumptions in the core region and the primary 25 j system thermal hydraulics. Although you will see' reported 1" :T A YL OE ASSO CIA TES 162 5 I Street, N. W. - Suite 1004 w ashington, D.C. 20006 (202) 293 3950
n- . 65 1 todayfsingle values for fission product releases, you
,.-~.
2 should recognize tnose. values could be changed dramatically
'3 by just such things as-the slumping behavior.
- We will show you an example this afternoon, not 5 intentional sensitivity study but basically the same 6
sequence that we have done with Peach Bottom and Grand 7-Gulf, where we get very large differences in primary system 8 deposition just due to a fairly minor change in the 9- treatment of the modeling of the core slumping. 10 MR. WALKER: Rich, in this particular case, it is 11 important too because it's a question of whether the 12 fission products are in the vessel available for release or 13 wnether.they've been transported down.the system and vented 14 out somewhere as a result of transport in the TMLB question 15 case. 16 MR. DENNING: Yes, that's right. 17 MR. RITZMAN:- Ritzman, SAI. Is I want to try to get this picture straight of 19 fission product hearing and reevolution and transferring to 20 particles. In a location if a surface heats up from the 21 deposit fission particle, reevolves and somehow it 22 partitions more favorably to aerosol particles, is that 23 downstream where the aerosol is cooled, or does it occur in 26 the same cell? 25 MR. DENNING: I thinx it's downstream, but I i" 'l IAYLDE ASSO CIA TES 1625 ! Street, N.W. - Suite 1004 W ashington, D.C. 20306
.(202) 293 3950
m . 4 4
~, -. - , n,x . ; .c. - ~.a - .g~.-.-,..
5 Y' , _ 66 0 1 .wou ld:'lixe.tollet Mike.see.if'ne could answer that because
' .(<
.2 ; he'has. looked at th'e TRAP results more than I have.
13 'Do you-want'-to attempt'to answer that? 4- 'Also,-we have to-be careful not to 5- . overgeneralize this particular-result to all cases. 6
- l' r .g 9
10 f gg-12 ~- ,
, 13 14
- Y 15 16 17-Id 19 4
' 23a 21 22
- 4. . 4 24 l
1 . 3 %
.- ! T E S'
'TAYLOC . A SS O C ' A
\~' 162 5 I Street, N.W . - Suite 1004 W ashington, D.C. 20006 (202) 293 3950-7 K f. ,
l . . < _ ..
~ ^ ~ .7-
[ 67
'l
,~ MR. KULHMAN:- Mike Kuhlman from Battelle.
l :2 what we see in this analysis, Bob,'is something
'3 that is similar to wnat we have seen previously, which is a
,- * . vapor deposit on the first structures that it sees when ' 5 these betin to heat.up previously due only to the hot steam 6 coming from-the core region. Now we have this additional
'7
~
neat which is coming from the decay heat of the fission 9 products themselves. , You.wina up preinvolving them into the gas 10 change, wnich at some' point in these analyses it's actually , 11 being cooler than the surfaces than are -the particles being 12 treateo. Thro ughout,~ the analysis is being at the 13 temperature of the gas stream. We are. incurring some 14 problems there because there is no fission product heating 15 on the particles themselves, allowing them to reach the gas 16 temperature. 17 what you are seeing is mainly transport la downstream. I think-the figures Rich gave in terms of 19 reduced deposition or retention efticiencies were only for 20 tne upper plenum. 21 MR. GIESEKE: That was for the entire plenum? 22 MR. KUHLMAN: For the entire plenum system. That 23 would imply then there are more of the vapors being ' 24 transported with the particles wnich are less subject to 25 deposition'than the steam. generator. Had the material then
' T A YL DE ASSO CIA TES 162 5 I Street, N.W'. - Suite 1004 W ashington, D.C. 20006 (202) 293-3950
~~ ' ~
c - - . xa x., - 69~ m 1 _ transported downstream with the vapor, it certainly would J 2 have been. plated out very efficiently in the steam
'3-
- generators in tne system, out.it wasn't able to-because of the. slumping _ particle.
5 MR. RITZMAN: The decay split there, the gamma portion of aecay-heating is more penetrating than the beta ] 7 energy. -how do you partition the beta and gamma 8 contributions in this analysis, or did you not? MR. DENNING: We.put the' decay. neat in the la structure. We are not modeling the depth of the structure 11 -- actually, we don't nave to model, we don't have to do 12 analysis. The structures here are thin enough so that it 13 doesn't matter. he put the decay heat, whether it is beta
- ' l' or gamma, into tne structures.
15 MR. COOPER: Cooper, Harvard. 16 Just a quick comment. The heat transfer to and 17 from the particles in the. gas stream is so rapid, I don't 18 think you will find the decay heating raises them aoove the 19 gas temperature. 20 MR. SILBERBERG Thank you, Rich. 21 We will declare a 15-minute break. 22 (Brief. recess.) 23 MR. SILBERBERG: Thank you. 24 I would just like to add to the comment that Jim 25 Gieseke mace,: and would liXe to acxnowledge of ficially for TAYLOE ASSO CIA TES 162 5 I Street, N.w. - Suite 1004 w ashington, D.C. 20006 (202) 293-3950
r - 69 1 the record that in the ensuing 3 or 4 months,-the 2 cooperation that we requested from the industry at the last 3 meeting regarding input of information to support our work
'was very well receivea by a number of groups.
5 And we would like to express our appreciation h ' for t'he support we got from Westinghouse, General Electric, 7 particularly for the BnR reports, obviously, and from Stone 8
'& hebster.
9 And with that, back to Jim Gieseke, who will now 10 discuss the Peach Bottom analysis, 11 (Slide) 12 MR. GIESEKE: To move on to the next topic, which 13 is, as I referred to it earlier, Volume 2 of BMI-2104, la wnich deals with the MARK-I design, Peach Bottom reactor in ad 15 particular. 14 Ano I reiterate the acknowledgments, since there 17 has been additional input to us on this topic, particularly 18 the help we'Ve gotten from General Electric in providing us 19 with all sorts.of information that we need, and also PNL, 20 wno have done the development of the SPARK code which we 21 are using for aerosol removal in suppression pools. 22 (Slide) 23 Just to quickly go through -- I don't know if 2" you can see this very well or not -- the whole procedure 25 that we are following in this is very similar to the r TAYLOE ASSO CIA TES 162 5 I Street, N.W . - Suite 1004 W ashington, D.C. 20006 (202) 29).3950
70 I procedure that..we'followed previously for'the Surry ' (' _2 reactor, where we begin with the selection of the plants l 3 and then the accident sequences, specify the geometry and
- " 'the accident sequence phenomenon, and:then go through the 5 . thermal hydraulics analysis with the MARCH code for both' the f ue l - neatup, whi ch comes over -- ' information is used 7
over this way, and in the primary system and containment 8 thermal hydraulic information, which reads in the other two directions, followed by MERGE-calculations for the reactor 10 coolant system in this case, which then is followed by the-11: - transport and the. reactor coolant system _of the TRAPMELT 12L code. 13 Information from MARCH used for release from tne l' fuel in the CORSOR code also gives us the initial
15 - temperature of the melt for the core-concrete interaction l' model, the release from the core-concrete' interaction as 17 does the information from tne CORSOR code which tells us 18 that is still there when it arrives at the concrete.
19 This then is combined in'a series here, and this 20 will. nave to be described indivioually for the sequences. 21 It's not a straight-through shot in all the cases. There is 22 a lot of comp 11 cations involved in nandling the codes. 23 sometimes you run an hour code for a while, then you stop 24 it and start it, a SPARK code, then you pick up the 25 calculations later.. That will be explainea. [ , ' j_ TAYLDE ASSO CIA T ES 142 5 I Street, N.w. - Suite 1004 w ashington, 0.C. 20006 (202) 293 3950 5
i 71 11: That is the final calculation'which all this in 2 leads towards, which is then released to the environment. 3 (Slide) ,
Just to briefly give you a list of the topics we 5
will be. discussing. In case you are wondering what I am doing, this is an introduction. l 7 (Laughter) e Following that, we will be hearing-from-Rich Denning on the sequence descriptions and thermal la nydraulics.- Mi4e Kuhlman will tell us about the reactor 11 coolant system transfer as well as release from the pool, 12 then transport in containment and attenuation in 13 suppression pools, Ken Lee. l And I will try to summarize it l' i a little. bit following those. O. 15 Witn that, we will go on to Ricn Denning.
'A' (Slide) 17 MR. DENNING: I am' Rich Denning, and I will be la describing the accident sequences, what happens in these 19 accident sequences, anci the thermal nydraulics, 20 particularly the primary system thermal hydraulics.
21
, .The design that we will be talxing about is a 22 MARK-I design. And for this particular design we have 23 selected Peach Bottom 2 as a representative reactor. And 24 the reason is obvious. That's the reactor that was analyzed 2$ in WA3H-1400. So, for historical reasons, we wanted to
, iAYLOC ASSO CIA TES 3
162 5 I Street, N.W. - Suite 1004 w euhington, D.C. 20006 (202) 293 39$0 ,
72 1 analyze that reactor. In addition, we have quite a bit of j, t 2 information on that particular reactor. 3 I guess I should also say I would like to thank Steve Hodge for all the help from Oak Ridge that he 5 provided to us in describing the reactor from his experience with Browns Ferry. 7 The sequences that we selected to analyze here 8 were basically tne risk-dominant sequences, he used the same philosophy that we described before for the Surry 10 plant. That is, we were looking at severe accident 11 sequences, particularly one that nad the greatest effect on 12 the risx. 13 he wanted to cover a spectrum of behavior, and 14 that's why we have selected the AE sequence, a large LOCA O 15 sequence of failure for the CC system. In WASH-1400 the TC 16 and TW sequences were overwhelmingly tne dominant accident 17 sequences, not only from their consequences but also their 18 predicted comparabilities in WASH-1400. 1Y NRC has a program called ACEP, accident sequence 23 evaluation program, which is reevaluating the probabilities 21 of accident sequences from the 1,300 PRAS that are out in 22 the world today. 23 The results of that are quite consistent. The TC 2' and TW are still very important accident sequences. TQUV 25 also an important accident sequence. You will see we TAYLOE ASSO CI A TES 16251 Street, N.W. . Suite 1004 w ashington, D.C. 20006 (202) 293-3950 L
-- . - =. .-- .. . . . .
- , .. ~ , . .
~
73 e ; : 1
?
I- ' analyzed.that'for-the Grand dulf. ! [, . L t ,
- 2
- The-'AE isn't really'an important accident ;
3~ f sequence from a PRA. viewpoint, but our other objective was
*^
'really to:look atia spectrum.of conditions inlthe primary (r <
~S' system and. the contiainment. So tha,t's why we picked this
-' _p articularisequence. ?And I'will be. describing what those l
7 are if'those initials' don't mean anything to you, f 8-
'(slide). f
.The code tnat we're usingchere is the MARCH 2 r
'10 ' code. .' MARCH 2 is'still-under development. The version'that f t
11
'we have u' sed here, essentially this' version,-nas been r
7 12 ,. 1 available to the national laboratories sincefapproximately } 13' January,.and they are undergoing peer review comments on l' it. 15 We really had a number of participants'in coming I'
.up with the MARCH 2 code.. It really is broader than the I
17
. laboratories that are indicated here. I 18 Tnere were a number of objectives for developing ;
l' l MARCH 2'rather than the.' MARCH 1.1 code. Those objectives 20 l primarily related to since MARCH 1.1 had'been released, the " 21 number;of' laboratories such as Brookhaven, Sandia, Oak 22 l> Ridge whien had been developing their own models for 23 studies that they had been performing with the MARCH code. 4 2' ' So there were a number of new models, improved l-
'25
[ modelsLthat were available in various. forms and various !
/
t i g", fAYLDE A $50 CI A T CS ! 1625 t Street, N.W. - Suite 1004 I w ashington, D.C. 20006 l l' (202) 29).3950 i i t- l
- . . ~ _ ,. .m - t .-_m_ . _ . _ _ _ _ _ . _ . _ _
74 r: 1 versions of MARCH spread out around these laboratories. 2 So one of tne-reasons was to upgrade the models 3
.in the code, to put'in these models that. people had been working on for a few years.
5 There are problems in the transportability of HARCH 1.1,-partoly related to the code structure, partly 7 related to the language. It's very CDC-oriented, some of it s
~ evenL the Battelle computer oriented. Some of the changes in i
MARCH 2 do relate to code structure language. It's written 10 in-FORTRAN 77, although the version we've been using here L 11 is a FORTRAN 4 version.'And there are some corrections also-i 12 in MARCH 1.1 that have been-included in the MARCH 2 code. 13 I am going to quickly identify some of the 14 improved models that are in MARCH 2 over MARCH 1.1. This is 15 by no means comprehensive. Under each of these there is l' quite a story to tell. 17 One of the most important changes is in the . La representation of the decay heat. In the MARCH 1.1, the 19 component of decay heat coming from the heavy elements is to i not included. Tnat is included in MARCH 2. So typically, 21 you see higher accay heat, somewhat shorter times for 22 neating up tne core,-this type of thing in MARCH 2. 23 Water and steam properties, quite frequently in 2' MARCH 1.1 there were single values that were used for water 25 and steam properties independent of the particular regime TAYLOC ASSO CIA TES 1625 I Street N.W. - Suite 1004 < W eshington, 0.C. 20006 (202) 293-3950 m 4 m
75 t -
-l 'you are in. There have been a number of enanges for p
- 2. improvements there.
-3
-Heat transfer correlations in the core, they nave' been upgraded significantly. Some of these make fairly 5
l significant results in the results; some make'very little results. .But most of these changes do improve the 7
-credibility of the code.
8 LIn the area of debris coolant interactions, , 9
-there are a number of cnanges, and I will only identify a 10 few of these, including ' a debris beds to represent . the 11 heating of the debris bed -- I am sorry -- the. cooling or l
12 l heat transfers from the debris bed have been incorporated. ! 13 These are actually Brookhaven models that have been l' incorporated in the code. , i ' ~(2) '15 These things can make so,me significant 16 differences in the' behavior. And ,ou saw some sensitivity 17 stuuies where Pete shoped hee effect of some of these t. ld options. ! l' The zircalloy' steam reaction modeling has been 20 upgraded. There is steel-steam reaction that is now allowed 21 L in some regions, such as in the reactor cavity, and this 22 can have quite a contrioution to the amount of hydrogen l 23 that's produced. There nave been significant changes in the 24 hydrogen burning models. Many of the HECTOR models have now 25 been incorporated into the FMRCH code.
' 'j TAYLOC AS$0 CIA TES 1625 ! Street, N.w. Suite 2004 w ashington, 0.C. 20006 (202) 293-3950 ,
y ,- .. ~ , 76
- l. There are some improvements in the way that we
;f' 2- have used MARCH that aren't necessarily new models or new-
'3 -
options in MARCH 2.but which are things which have evolved and are ' easier with the MARCH 2 code, such as the more 5 gradual' slumping that Pete Czybulsmis talked about before. 1In these PWR' analyses, we have used what we call
'7 l a gradual slumping model. That is, when the fuel gets to a 8~
fairly. low amount of melting, we allow it to start to slump 9 and provide its heat to the water in the. lower plenum area 10 and then later eventually will get to a period where we 11 will then slump the rest of it into theflower plenum. 12 And you will see that's going to have a ! 13 significant effect when I show you the results of the l' temperatures in the reactor coolant system. O 15 I would like to point. out one thing to make l' clear that's not included. One of the models that we are 17 currentiy putting into MARCH d is a SHROUD model. If you l Id are familiar with the BWR design, there is a can around l' each assembly, a zircalloy can around each assembly. 23 Cak Ridge has' developed a SHROUD model for MARCH 21 which treats that can separately from the cladding. We are 22 currently incorporating that into the MARCH 2.0 code, but 23 it was not incorporated in the version we used to run these 26 -analyses. So the zircalloy that's associated with the can 25 is assumed to be at the same temperature as the zircalloy TAYLOC ASSO CI A TES
*' 162 5 i Street, N.W. . Suite 1004 W eshington, 0.C. 23006 (202) 29).3950
._ c t'
~
, 4
-) iti;77 .
- o ,' ' ' N m -
g ^g
~ '
- y. h 1! .that's-associa,ted with the cladding.
~ _ _
A J ,3 7 :- .S
~2 '
'TbhtTis probablyinot a terrible approximation,. .[
-3I ibut l't.certainly~i,s one:that we willLbe looking at later in
~
~
-g'
*~ sensitivity studies'toiseeJwhat' thatieffect is. 'V
. 0ak< Ridge 1 A 5 .has actually-done'some' sensitivity studies.'with our;model .
'l
'x '
tosee-whatthati,effectlis,-toy.- ', ;i 7 ",- .. 4 y, - yh
' tj .- MD-
'%,, 1 s
,sj
-10.4
- 9 s i s>
.gg 12 10
( . .__
,13 < -s=
.T 9
gr 's .>
'('
Y - 3 .
~,-
.U g3-
"~
3 .
- . y);
. 3 16 * .y
, n i 17 - 'j.
q
' E %, #
id y., is 19 1 q 4 ' 23 21 22 - f.l i 23 ,e
'I cf t
i 24 af
"\
,. s 25- ?
( ' *II t
'4
? L_ .
,f fAYt.OE A$50 CI A TE5 '! ,
16 2 5 ! Stre e t,' N;. w . . Suite 10 0 4 - 1 l W eshington, .D.C. 20006 - T' i '~ 3 (202)'293-3950 ' ~ 3, y t. .
. $$9
^
% =f .>
_b,~
]
~
' ~
' "~
u.
;rg4: -
4
%, 1 The BARK.I design-isEsignificantly smaller than M? 2L ithe PWR'containmentithat'We talked about at the last meeting.
3- This. region'that I'm tracing out here represents the contain-4 ment--not this! outer building here, but this is the. primary 4: N- 5' containment, _-is a steel. shell. This is a steel containment.It's 8 [ a free-standing steel ~shell. This is the drywell region . 7=
.here.- This is-the wetwell region here, where the suppression 8
- pool-isLthere is a vapor spaceIhere. The volume here in the 8-
-drywell region is 159,000 cubic feet. The vapor space in the 10 wetwell is 119,000 cubic feet. ~That-compares with approxi-
-11 Emately two million' cubic feet for the large size PWR--so a 12
- much sm' aller containment design. And this,-o'f course, will ,
13
' affect' residence times'and things like that. -
14 The purpose, of course,-for this kind of configur-5 15 ation is.if one gets a pipe break accident, then the steam 4 l 16 is. relieved through vents into the suppression pool, and as a 17 . . you will see in.the transient type of sequences we'll be 3 18
-looking at, the relief is through spargers into the suppression 1
L 18 ' pool. Vr}! 20 Now let me talk-a little' bit about the failure
.4 21 ~ og-this containment. In WASH-1400, there were analyses done j
s
~ 22 '
, that indicated that the. failure pressure, the ultimate failure 23 pressure for this containment, might be in the range of 24 250' psia, but there were some approximations in those
'%C' 25 analyses.- In the treatment of this' containment in WASH-1400, M ~
f[.
*: ,- .. ,--y~.,,_. . . . . , , , . . . - ,- . .
- h. 1_.__.._.
_ n MMB. 79 rg5 4 - l' a failure pressura of 175 psia plus or minus 25 psia was (~% ' 2 assumed. That's 175 psia. It was also assumed as a result 3 of the analyses that the f ailure location would be right <here 4 in the wetwell. Now, that had a lot of implications for the 5 accident sequences. 6 First of .all, recognizing a failure locatior. in 7 this region, it was felt that we .could not assume that the 8 failure' occurred there, that the water would stay.in the 9 torus, so the dominant accident sequences in WASH-1400 or 10 the major consequence accident sequences are ones in which 11 containment failure occurs before core meltdown begins. 12 As I explain the TC and TW sequences, you'll see 7
} 13 why that happens, and in . which--that the containment has 14 failed before the core is melted down, and where the water
$ 15 here is basically assumed to have left this location, and l 16 that the' blow-down from the primary system inta the suppressio n L 'o lI 17 pool area doesn't reallyLgo for water, so the relief is basi- '
i 18 cally in this room here. r i i 19 I'll talk'a-little bit more about'the failure I 2 modes for this building, the reactor building, later, but r f 21 let'me;say in WASH-1400, it was felt'that there were two 3
' 22 kin'ds of-failure.
l One is--there is a-quadrant of this region
'M down here that'has a ceiling that goes directly to a. floor
;. , M' .that'goes i to~the-outside,.and it was assumed that one-fourth
- *\.
j N- of.the time corresponding to that quadrant, the failure in d
)
~_ _ _ _ _ ..
m ..
r - ou MMg rg 1 this building would occur in that quadrant, and you would
- 7. .
! 2 have fission products released into this vapor region, and 3 then directly to the outside, and these are called the 4 " gamma prime f ailure mode sequences. " They are the large 5 consequence accident sequences in WASH-1400. Let me repeat 6 again: it is assumed in these sequences containment failure 7 had preceeded core meltdown, the release from the primary.
8 system goes directly into this region bere, and then directly 9 to the outside'without the potential for deposition. 10 The other failure mode that was assumed in WASH-11 1400 was -- this is for this reactor building out here now-- 12 that you had a blow-out occurring in blow-out panels in this region of the reactor building, and there was transport that O 13 14 occurred from this room either up through connected volumes h 15 and then out or up an annular region, that is, between this t l 16 free-standing steel shell and this concrete wall that is { 17 behind there, and the analyses that were actually done are i g 18 really for this pathway here--that is, up this annulism and I j 19 out the building, and those are called the " gamma failure 20 modes," and have somewhat less consequences than the gamma f i 21 prime, but still quite large. i 22 The amount of retention within the annulus is 23 significant, but less than an order of magnitude effect. 24 Okay. Ncw, subsequent to WASH-1400, the NRC had
'N '
25 a study done in Ames, Iowa, to look at the failure levels u
di MM8 > rg6-1_ for steel containments. One of the steel containments they
~ .,s*
2- looked at was a MARK I design. The predicted f ailure pressure 3 'was 123 psia, significantly less than 175 psia in WASH-1400, 4 and the location of failure was felt to be in this place 5 right here-- this arrow goes a little too far.-- right at that 6 point right there, in a drywell region. That, of course, is 7 going to have significant. implications. It.means that the
'8 water will not.be displaced down.here. We' assume the water 8 is not displaced as a result of containment failure.
10 Any releases that occur from the reactor coolant 11 system into the drywell are going to be able to go out of 12 that failure without going into thel suppression pool in some 13 of these sequences. This will be clear as we go through some 14 of the sequences. 15 This reactor building here is really divided into e j 16 two regions, the lower region, below a refuelling floor, and 17 an upper region. The lower region has a three psi guage de-1 18 sign pressure to be'able to withstand a tornado. The upper i 18 [ : region is.just a steel shell, a metal-siding type building E
! that has very little strength. Tnere are also blow-out 21
{ . panels in this region. There are. blow-out panels from this s 22 region'into that region, and'there are blow-out panels from l 23
~ here into the environment. You will:see as we get into the-24 analysis of the.se sequences that the. integrity of this 25 building can be.very important, and the things that-happen g
~ -
MM8. 62 rg7 1 in this building can be important with.regards to mitigating
,. s
- n. 2 the consequences of the' accident sequences.
3 I will d'iscuss one sequence in which we have looked 4- 'at-the effect of retention within this building. In the 5 other sequences, we have assumed that the failure is such 6 that'we would either blow down from here into the--this
~7, cavity room, and out into the environment. We would have 8 passed up into'here and failed this part of the system, or 9 that the failure of the primary. containment is sufficiently 10 violent that it leads to failure of the secondary containment
- 11 reactor building.
. 12 There is one sequence, however, TC, in which we 13 look at the effect of this reactor building'on additional Oz.f
\C/
14 -_ retention. h ~ 15 , MR. KASTENBERG: Rich, does the_ change of your 3 -
- g. 16 ' idea of where.the containment fails change the relative f 17 risk dominance of the sequences?
1 lr 18 MR. DENNING: No, it doesn't change. These h '19 sequences are so dominant from a probability standpoint as e 20' well as'a consequence standpoint that_it doesn't really f j 21 affect that, but you will see it will have a significant-3 : n, 'effect on the release patterns, what periods of the accident 2 Lare important from a release standpoint. 24 - We'have taken this pathway here.for-the study. i .:
' A"/ 25 We don't really feel the question has been really resolved
,y__-9 -_.y_
4 ,.e- p ' " ' ' ~ ~*
c- .- -a,_- MM8 83 rg8-1 adequately as to whether~the failure is here, there, or
.p. . 2- possibly some'other place, and it might even be a probablistic 3 consideration,.could very well be something that differs 4 between different designs.
5 But we have for the purposes of this study con-6 sidered the drywell' failure. 7 MR. KASTENBERG: .That was the next question. r l 8 In the material we were sent, I guess it confirms that is the 9 only one you've looked at. 10 MR. DENNING: That's the only one we looked at. l
'11 MR. KASTENBERG: My next question was going to 1
i 12 be, why didn't you look at both that failure and the wetwell 13 failure. 14 MR. DENNING: It was only a matter of time. That!s h 15 the type of thing that certainly could be a recommendation l Y j j 16 ~ that could come out of this' meeting, but basically we want o l 17 to see the results of that. I i 18 MR. KASTENBERG: All right. I recommend it. l I 6
.[ 18 (Slide.).
20 MR. DENNING: All right.
-[
- 21 I've.already-pointed out the things on this next l- :N view graph.
That:is, for the study.thatLwe have.here we are
~
i- 23 going-to_look at 132 psia, and f ailure _ in the drywell.
. 24' (Slide.)
i
'~ 25 Now I would like to describe the AE sequence, a.
.a._ . - . . . _ . ,
=- . . . ;._ . _ . . , . _ .. MM8 84'
-rg9 1 flarge LOCA with failure of the ECC system. This is a se-2 -quence that happens very rapidly. The beginning of core 3 nelt is.at 12 minutes. Now this is very important. In the AE 4 sequence'as.we have analyzed it with MARCH II, we predict a 5 containment failure that is 132 psia is reached before we .
6 have melted through the bottom ~of the reactor vessel. 7 So the failure.of containment occurs during the
.a core meltdown period. I'll show you the implications of 9- this as far7as flows'and that type of thing, then reactor 10l vessel failure occurringiquite a , bit later at approximately l
11- two hours, and the only failure mode of the containment we 12 really. considered here is the gamma prime -- that is, direct 1
- ; 13 - release to the environment.- ;
14 We haven't considered dhe gamma, although the ^
\
t l- 15 gamma failure modefis certainly a' possibility for this se- { k= 16 quence, just as it is for'the TC sequence for which we've
'8 17 analyzed.
e I 18 (Slide.) r 19 MR. SIGAL: Can you elaborate why the contain-I ' { m' ment. fails at 34 minutes?. 5 21 MR. DENNING: The reason the containment fails y . n at that time is a result o'f hydrogen generation, the amount 23 - of hydrogen we predict has been produced due to zirconium 24 water reaction is. sufficient at that time to lead to con-
~
25 tainment failure. -
<MM8' 85 (i = rg10; 11' MR. SIGAL:. Subsequent question -- is it because
'~'
Li 2 you include the shroud zirc. alloy with the clad zirc alloy? 3 - NR . DENNING: .That certainly is a contributor 4 to that, Raj.. If there were:no shroud reaction, then cer-5 .tainly you wouldn't have failure at that time. However, we 6 do. expect' to -have significant shroud reaction. Pete, do 7- .you want to comment on that?.
'8 MR. CZYBULSKIS: ' Rich,.I'm not sure that you're 9 . totally' correct that the shrouds dominate that effect. There 10 'i s'a tremendous amount of zirconium in that core, whether-11 you look at the cladding alone or whether:you combine them 12 -with the shrouds..
13 ' At the' time of containment failure, I don't have ( 14 the exact numbers at my. fingertips, but I believe you react h 15 on an order of the third of t.he available clad. I'm not sure
.3
[ 16 that including the shroudsLis particularly important, but it 0 l 17 is~a point of concern. 1
'!' 18 MR. WORMAN: St'one and Webster. Relative to the
.I.- . .
g 19 . hydrogen, I wanted to clarify a point.- My reading of the E l t 2 document was.this: .that a hydrogen burn or explosion, but j - 21
~
~
rather an overpressurization due to the' hydrogen' existence [ 22 in the containment--these are inerted cont'ainments with hy-23 -drogen'in'the-first place. Now, you inert them in addition
. t, . .
lM with steam,,- You_ are not; treating the. hydrogen explosion) M'
~"
~ you're treating an overpressure.of hydrogen.
l
MM8 86
- rgil 11 MR. DENNING: Exactly. It's not.a1 hydrogen burn in this! MARK I sequence.
2 It's a matter- of the non-condensed 3 full gases being pushed into the wetwell vapor space 4 and really overpressurizing_due to the noncondensables. 5 MR. WORMAN: I might add :it would be good if the 6 analyses to support of that assumption include in the report, 7 ~ just state in the text that's 'how this happened. People .from-8 Missouri would like to see that.- 8 MR. DENNING: This_is the flow path within the core. 10 Basically what'we have is a core region, a dome steam separa- -
'll tors. I've.only shown-two of the steam separators here. And 12 in-this large LOCA sequence, we've had a failure in a re-13
. circulation loop, and so. on its way out into' the drywell, 14 the pathway is through the steam separators down an annular 15 region and out into.the drywell.
4 [ 16 . The representation that is used in MARCH and in o 17 [ TRAP-looks like this: 1-j- 18 (slide.) l 18 i The regions-that are modelled are small top guide
# region shroud head,. pipes and separators, the-lower-outer 21 ' annulus, and'then exiting out into the containment.. 'The z
22 _really important region.here_is the separators. l _ 23 (Slide.) 24 The kinds ^of' temperatures.that1are_ predicted with
~
r ,.
'~ 26 ' the= MERGE ' code subsequent to the MARdH analyses --' these show . ,. .u . u _ ,_ a . . _ m. c . - . . .
MM8-87 rgl2' 1 gas temperatures, the structure temperatures are fairly g 2 similar, and that's really what I'd_like to show you. 3' (Slide.) 4 The important things I would like to point out -- 5 this is'the separator temperature. It's going up into the 6 region of approximately'2000 degrees Fahrenheit, and staying 7 infthe region Sf greater than 1500 degrees Fahrenheit. And 8 this region is very important from cesium iodide deposition 9 as you'll see. It?s because'these temperatures stay pretty 10 high-in this region here for the steam separator.in this 11 case -- Dr. Ritzman?' 12 MR. RITZMAN: What is time zero on this plot? 13 Is_that the beginning of core melt? 14 MR. DENNING: Beginning of core melting. h 15 MR. RITZMAN: I'd.like to make a comment now for
- f. 16 the record -- that that'be put onfall view graphs or all 8 17 figures because it's very confusing what zero time means.
e i
- MR. SILBERBERG: Thank you.
, 18 E
'h 19 MR. DENNING: Okay.
t 20 Now1the things I would like to point outLare the
.{
21- turnover here. This is partly as a result of the slumping { 3
' 22' and partly asia result.of just i steam coming out of the lower 23 head. 'But what we see is a turnover in these plots with the 24 ' temperatures -- where the higheristeam flow rate as the core
'%- 35 is falling into the lower plenum and producing more steam --
~
u., ,____w._ _ . ca. ...y,...-- ,! _., 3 .. . .. . .,s , g ,x-. iQ . m
MM8 . - , i 88 rgl3
, ~ .
1- Lit's~ turning over the temperatures in the~ rest'of the system. t.
\ 2 You'll see some cases particularly this afternoon 3 where the - peak temperatures are really clipped of f, leading 4- to more retention than we have in this particular case.
5 All the cases that you will see this morning, the steam
'6 separator-temperatures are going to stay up in this region 7 of 1500 to 2000 degrees F. kulthat's going to have a very 8 important impact on'the retention'of cesium iodide.
9 .(Slide.) 10 During the. period prior to primary. containment 11 . failure, the flow path is from the drywell into the suppres-12 = sion' pool. Anything that doesn' t get scrubbed out in the 1
-t 13 suppression pool is rea.11y getting stored up in this region 7
14 right here. h 15 (slide ) 3 [ 16 After containment failure, the material that was o l 17 stored'up in here, the fir'st thing that happens, remember , I 18 ' containment failure rate, is going to happen right here. g 19 It happens while weere still melting down in the s 20 ' core region. After containment. failure, we blow down this {' 21- regioniback into the drywell, and then out, and then-through j tt some pathway to.the environment. In this particular case we ZI
~
~have-only considered --
24 LMR.-VOGEL: Where is grade on this? It looks like N 25~ the break on the lef t hand. side might be below grade.. 1
- . . . _ _ _ . _ _ . . _ . _ . _ _ _ . . - _ 2-..._. ~- - .
.y
MM8 89
-rgl4
-1 MR. DENNING: This, for example, is, I think,
, 2- unrealistic. That I'm pretty sure is below grade. Yes.
-3 Actually..the pathway is. kind of like those.for that parti-4 1cular flow pattern.
5 MR. COOPER: Rich,.when-you blow down back into 6 the drywell, are you now bringing'all or much of the water 7 from the wetwell into the dry containment, or at least some s 'of it? When you look at it,'it looks like a siphon or some-9 thing that would bring it back into the drywell. 10 MR. DENNING: I don' t believe that ef fect is 11 credible. It certainly-isn't one that we have taken into 12 account. 13 -- See, there's a vacuum breaker here that opens. 14 MR. COOPER: So you have a check valve. h 15 MR. DENNING: I see, I'm'sorry. I see exactly l 16 what you're saying. There's a vacuum breaker here that 17 allows the flow to go from here back to here, and that { 1 18 is where the break-- r
- t. . . .
p 19 MR. COOPER: So you won't get the liquid from 8' ~ the wet suppression well into the drywell when you get the 20
.[
5 21 pressure difference? A EndL tape.22 MR. DENNING: No '. 23
' 24
'i -
~\ /.
25 pre
- 4. h- m-- in, p=* -
se p -- = = gow.
TA 195 90-JMM jl1.9:1:
'l
' MR. WORKMAN: - Could you1 address the large horizontal-qC 2
- e. arrow going through the large 8-foot concrete section of the drywell?
3 'MR. DENNING: This is to represent -- there are a
-4 number of ' paths through liere into.different regions. And 15 - there is'a potential for flow paths into here and up through 6 here.
7- In the case, for example, where we considered the 8 containment -- I'm'sorryf the reactor building remaining in-9 tact,Lthe; gamma sequences.it would have been a pathway some-10 - thing like this. There are a number of penetrations that go 11 ~ through here, with openings from this area into rooms there. 12 But basically, what I was really trying to show is 13 .' there are a' number of possibilities here and quite a bit of 14 uncertainty as to just exactly what flow path there is here. 15 But we should recognize, for the gamma prime [ 16 sequence, we're assuming that we have-something that's a 0 1," 17 somewhat direct pathway like that. 3. 18 (slide,)
.[ Now, I'm going to move on to the TC sequence. This
.{'
g 20 - is a transient, with' failure to' scram.
}1 21 In this sequence, basically you have a transient
-accident, the rods don't go in, nor does liquid' control go
~
~
l
' 23 into' shut down the reactor. You are dumping a large fraction 24 The heat-removal of the' decay heat into the suppression pool.
Q-l 25- capability for the suppression pool can't keep up with it.
.%_s- - .. - . . . . ~ . . . _ . . - -. ~ _ _ . ... _- .- ..
91 MMaji .9:2: I' You overheat. that water, raise the pressure in the containment y-K
'E 2 - up to the point where it ' fail's, and you fail the containment
.3 .before you've started the meltdown.
~
14 At this point, then, you cavitate pumps that are 5 providing water back int'o keepsthe core cool. You then start 6
~
to heat up,.begin coremelt at about 1-1/2 hours. And in our
~
-7 analyses, we predict pressure vessel failure occurring about
-8 -'217 minutes. We've~ considered a gamma prime and a gamma
~
8 containment = failure mode for this particular sequence. 10 MR. REYNOLDS: What might start the transients?
'll What'are the things ~-- are you sensitive --
i 12
. \{. MR. DENNING: It has to be a transient that would, I
. '13 believe, cause isolation of the steam line. There are a 14 variety o'f transients that can lead to-various conditions.
. i
- j. 15 7,m not sure exactly what the points are essential to get into 3 1
[ 16 a TT sequence that gives you trouble, but it's that type of o 17 ~ thing. There are a large.-type of transients that can start i j 18 the sequence. 19 - MR. REYNOLDS: What you did you assumed for this { ! A ! 20
.l -particular calculation. Does it matter?
21 MR. DENNING: Its doesn' t matter.
.3 ;
22 f MR. KASTENBERG: Just to go back for a second, you 23 -
~
only looked at the gamma failure mode for the sequence, ifI 24 - recall, from the material we were'given?'
.i.
# MR.' DENNING: .You will see results of the gamma l
. . ~ - , . ~ . - . , . . . . - . - - - - . - - . - . - , - - -
. .. .. ~
y A +-4>--u- -.. y , , - - . , ,
MM:jl 9:3- 92
'l prime and.a~ gamma sequence.
~(*
-2 Now, in-reality, there are a number of gamma 3' ' sequences.-- that is, behavior ~in the reactor building. And
.4 as I talk about-some-of these uncertainties, I'll identify 5 some of those.
6 The' flow path here is from the core to the steam 7- separators,,through the steam dryers, and out the steam line.
.8 There's a. bypass pathway around the steam dryers once the 9 water level has dropped below a certain level. And that's 10 - represented by about 15 percent.
11 so, there is some flow that is bypassing -- that 12 the important deposition areas within the system are the 13 steam. separators and the steam dryers. 14 (slide,) 15 This shows the type of representation.for the
-0
[ 16 MERGE 'and TRAP analyses -- the top guide area, shroud areas, o l 17 pipes and separators. 1 - j 18 ~ This is.the bypass area to the steam line, steam j; 19 dryers, and upper outer annulus, then connecting through a
! 20
-). relief line'and down to the suppression pool. The suppression j 21 pool'-of
, course,'is not modeled in~ TRAP, it's modeled by the 2
22 SPARC code. f 23
. ~(Slide. ) '
24 - The tiypes of temperatures that we obtained in the -t 26 TC sequence are' fairly similar to the'AE. We see.the turnover
- ~ _ . _ . . _ . . . . . . _ . . . - - -.
'T
- MMajl 9:4-- 93'
. behavior. We see'the separators and the steam dryers being
~
1 ( 2 at temperatures in this range, 1500 to 2000 F. And that
-- 3
. turned out to be very important as'far as the cesium iodide 4- deposition is concerned.- .
5 (Slide.) 6
.The' flow pathway from the vessel during the period
'7 before you have melted through-the lower head of the vessel 8
.is through-a' steam-line and relief line,:into the suppression 8
pool. 10 ' Remember, the containment has. failed in this case. 11' And back'then, from this vapor space to the. dry well, to the 12 failure -- and in the~ gamma prime sequence, to the environment 13 __.in the gamma sequence, into this volume here. And then,
' I4 this particular area here has a standby gas treatment system 15 of significant capacity that can take flow from their reactor g 16 building and this upper area' of the reactor building, draws 0
- 17 from both of those equally, takes them through trains and a
18 filters.and out through to the environment.
-I.
a j 3 18 In the TC gamma sequence for TC, we found the 7, 20 -
- l. capacity.of the standby gas trestment': system to be-less than 21 the rate of steaming from here, to that we have a significant 22-
- , ;f. amount of material escaping out through the walls of the 23 building, even'though.the standby gas treatment system was 24 '
operating' during4 the time period of release. 4 (_, 25 Additionally -- and.this is getting a little bit 1
- ,. a .n e + ..n lr. .4p e e ax -' s - - ~ e=e - em , -* ' ~ .
e T 4 T - *g -"M-u y
~
. 1
- 'MM jl1 1:5z 94
- 1: '
' ahead.of the story -- there's a fairly limited capacity of 2 the HEPA' filters associated'with that train to retain the 3
' aerosols generated here and the material that went through the 4- standby gas system. There was.some. filtering, but that did f
5
~
not haveta major effect onto the environment..
.6 MR. RAJ SEHGAL: I guess the steaming rate is propor-7
=tionate to the power that you assume'during this transient. I f'
8 think you assume about 30 percent power? 30 or 40 percent? 8 Which is-very high -- TC. 10 - MR. DENNING: Then, I think Let me make a comment. 11 - Pete should p'robably_'also address this. 12 Although we have a power. level of 30 percent at the 13 beg' inning of core uncovery, as the core uncovers, that's 14 reduced in modeling that's done within the MARCH code, so 15 that it does not stay at.30 percent power during this period 3 l 16 of core uncovery. 17 Pete, do you remember the details of that enough 18 to give.some idea as to what that looks like?
~
I 18 [ MR. CZYBULSKIS: We use the 30 percent equilibrium E 20
.l power. level while the core is fully covered. I believe, from
. 21 everything I've read and discussed with my friends in GE, 22 ll
-that11s an appropriate power'1evel to use for this particular 23 -
transient. 24 As the core begins'to uncover, we reduce the power 25
-to the. point of when the core is nearly dry.- It is down to
.-.;.,_._.. ,.. _ _ , _ .a _ . .-;. .. -- .
% M' , . L ': :: - ~
, JMMajl:936f -
95- i ,,.. , 'w
- , :1' Ldecay heat.; And from:then on in', we're down2-- back to. decay
' (7: "ph 2- h'e at 'as ; a function 'of--l time.
[3 ' MR. RAJ SEHGAL: I believe the best estimate values 4- - are lower: than 30:-percent,: even when the core _ height is full. 5 MR. CZYBULSKIS: I wonder if our General Electric
. '6 . ; friends'would. care'to: comment --
7- MR.' DENNING: Excuse me j~ust'.a second, Pete. Let s 8 . me comment,.for the' recorder here, Raj Sehgal, again, was
'8 ' commenting' and saying he _ felti that the best estimates for this
- 10 equilibrium power level were less~than 30 percent. I believe 11 :that the'30' percent is a good. number.for this particular
- 12 sequence..
In-the-Peach =Bo'ttom reactor, you'11 notice, for
' ~
13' 14 Grand Gulf, wefdo use a'different value than the 30 percent. g- 15' 1The principal effect .here, I think, Raj,'would be'with regard _ -5
) 16 -to'the timing o'f_ containment failure. It would be later.in f 17 - _ time.' That would have a little bit of1effect on the decay I
- 18' heat levels. -But'I don't think that that's the thing.that'.s
[ -
- 18 , driving.the. steaming' rate that-is affecting the outleakage of
[-
.[r
-.20 - the reactor building,.as: opposed'to'what's going up the stand- -
. eb .
3 by -- through thei standby gas treatment systiem.
~
5
~
l 22 fin-addition,.we-don't'see the standby gas-treatment
- - 23 system';as providing significant_ filtering. What it's really 24 (doing is providing you withJanLelevated release rather than
'\~ t - 26 .a ground -level ot*slightly _ elevated ~ release.
4 + 9 ~9*M*M g N MTS O WQep he p ps- e e-.g ag, m 9WwS. -fg 9 a 4 m-. 4
. MMijl; 9i7l- , J1: ' Now, what I think the biggest uncertainty associated i (l' l 1 2
-with'this gamma failure requence is'-- is in what happens to 3
th'e--fire-protectionLsystem in:the. reactor building. In this 4- region'of'th'e-reactor building, there is a fire protection
- 5 system that'will be activated on a combination of'-- basically 6
a smokeLdetector or an ionization. And with the aerosol release, we'll get that -- and with-high temperature. And 8 we'll get'that. -And this-is something that our. friends from
' ~
Oak Ridge have analyzed, is that it's possible that once we 10 ~
. failed this building that we-will have a sprinkler system 11
( operating in:here, and that could have a tremendous effect on a 12
.the retention of fission products.
() I 14 We haven' t actually analyzed that sequence. However,. it introduces other uncertainties, and that-is you have a E-j 15 great dealEof hydrogen that is being generated by this system. 8 16
= When we're in the dry well here, we are inerted. When we dump
- O
' 17 e into the reactor building,-basically what happens in the TC
-3 18 I sequence and the.TW sequence,.we blow steamEinto this building 2
- 19
~E and exhaust it of its air,-but we also, during this melting -
-m 2 ~ 20 E period, put'a lot of hydrogen in there.
' ; 21 j Ifithe sprinkler system does come on, what will I n.
.2. happen is the air will be brought back in and you will defin-
.n.
itely have a ' highly flammable- tondition; and with the kind of !- 24 integrity, the strengthfof'this, it's:not going to survive the 25 hydrogen burn.- F
= - 4 <-
++ c.. ,-
=
MMr j lL ' 9.3 8.
~
1- So, I think in that case, where you have the sprink-j.- -: . (_ 2 lers' operating,.you'11.have some period of sprinkler operation, 3- L some period of effectiveness, a hydrogen _ explosion that then 4 isvery:large in comparison with the strength of this building 5 to be'able to withstand it, and then a great deal of uncer-6 tainty as to what happens after that. 7
~
Now, the particular sequence we have analyzed is 8- not'that sequence.- ~The-one we analyzed is a case where we 8 failed this. building, we have. allowed the blowdown panels 10
-from this area to' blow down into this areaf We've considered the standby gas. treatment system as sucking out'25,000 cubic 12 -
feet per minute.through a combination of t$ese two. i
- 13 And when there's'a difference between the' amount of II steam coming in and that 25,000 cfm, we've allowed that -to ,, t
_j 15 leak outside the building an'd take-with it fission products. 3 !
-j 16 That's the gamma sequence we analyzed here, you'll see, 8
17' g- because of the combination of fairly large outleakage and be-3 > j -18 cause the filters are not that effective,-that the retention T
! I' '
in the reactor building is not particularly large.
.E I'm going _to. skip'the discussion o the TW sequence, I- 21 ~
other than- to make a few -- just a few comments, because of 22 - l the time. 23 - MR. WALKER: These are transient sequences, and 24 you're' talking about availability of the power protection 's .- 3 system, also availability of the standby gas treatment system. _.~.a._._,m._.-,,._~.. _ _ l
'MM jl 1939 97 1! -
'What is your situation with respect to power O 2- -availability, fire stations?
.3 MR. DENNING: Obviously, that would depend very 4
much on the. transient that you had going into the sequence. 5 rem not_.sure -- I would imagine -- I'm not sure I can answer 6 that. 7 Peter, do you know the answer to that?
- 8. .MR. CZYBULSKIS: There are a variety of transients, 8
a number of which do not involve loss of outside power. In 10 those; cases, you have your standby gas treatment system 11
-available. In the transients that do involve the loss of out-12' side power, I believe-the standby gas. treatment system is on hv I3 emergency power. ,
So, from that point of view, I think it
-14 would be available, except for those seduences l where you 15 '
completely lose power. e j 16 MR. WALKER: This is not like the TLMB -- 17 ' MR. DENNING: No. 18 MR. CZYBULSKIS: Not ne'cessarily. . t 18
! MR. DENNING: What I wasn't sure about is what-the 5
20
$. implications were of loss of off-site-power to the sprinkler 21 l
system, whether pumps are required for that. 22 MR. HODGE:
'There are-electric,' motor-driven pumps 23
-- come off the-diesel generators at the plant. There's also, 24
- p. ..a diesel-driven-fire pump, independent diesel.
Q/ - 3 MR. DENNING: I will_quickly make a couple of e_; - . __ . .
98 MMajl1 9 10-1- comments about TW. It looks a great deal like TC, except j, - ,, 3- 2. that.it's very.much extended in time. In this particular 3 -case, we have the transient.that is accompanied by loss of 4' decay heat removal from the pool. We have the core, however,
~
5 -down 'at -decay heat levels, dumping its decay heat power into
-6 .
the pool,-slowing raising its temperature until we get to the 7 point where the containment pressure has reached the failure 8 level,1at 1800 minutes. 8 Then, we see cavitation after containment failure, 10 . cavitation of the-pumps, so that we lose makeup to.the 11 -reactor vessel. 12 (Slide.)
~
13 Look at these time periods, very long periods of ( , 14 time before we can begin to start the coremelt -- very long 15 time before we get to reactor vessel failure.
?
l 16 The only failure mode.we've looked at in 0
'17 l
~
this case is the gamma prime failure mode. The temperatures
'9
~ 18 in thisEcase are fairly similar to those in TC,~except s'
18 extended very much in time.
! (Slide.)-
j 21 Finally, let me make my plea :again to point out how c- 2 3 -- important some of the modeling assumptions in the thermal [ 23 hydraulic area and containment response area are as far as t i 24 driving deficient particle transport. Some of these, as we
- 3,,.them, are, of course, time: of containmenti failure, such 25' L
l s'
%'- e
- y p.t 's-p a5 s g.9 i gg -Mu-,. ,e
99
~
MM:j1 9:10 g. as in that AE sequence.. If we're overestimating the amount of 2 hydrogen production or underestimating the failure levels of 3 the containment, that could have a significant effect on the 4 . fission product release to_the environment. 5~ Location of containment failure -- you'll see more,
'6 as we get into'results, how important that is; whether the 7 containment failure occurs in the dry well or the wet well, 8 for example, could have a significant effect.
8 The reactor coolant system thermal hydraulics are 10 - subject to significant uncertainty, this turning over the 11 temperatures --'if it happens earlier than what is shown for. 12 these Peach Bottom sequences, it could lead to more retention. 13 If itloccurs later, we could very well be underestimating -- 14 overestimating the retention. 15 Then, again, the response o'f the reactor building, 3 l 16 what happens to the reactor building after primary contain-0 l 17 ment failure? Can it withstand.the pressure loads that occur i 18 as a resulte of blowdown of the primary containment. failure? t
.!.; 18 - What-happens to hydrogen in the_ reactor building?_ Is there 20 a potential for hydrogen explosion? What happens to the l_
21 sprinkler system in;the reactor building? What'are the path-5 22 ways.for release'in the reactor building. l 23 That completes the things that I have to say at this 24 point.-- 'j-gg s
,-,, , . - u-m a ~- --
- = -
100 l ~ :MM:jl
- 9 :llJ l 1
1 MR. RAJ SEHGAL: One question about temperatures. (. 2 'How come the temperatures in these scenarios are much higher 3 than those in the PWR scenarios?.
~
4 . MR. DENNING: The question from Raj Sehgal is how 5- come our temperatures are higher than they were in the PWR 6 scenarios. 7 The reason is primarily related to the length of 8 ~ time of the accident sequences. We have a fairly long period 8 of time when we are dumping heat. It's not because the 10 core outlet temperatures are higher. Actually, our core 11- outlet' temperatures are lower. But the total period of time 12 that we're dumping heat into the upper regions is longer,
' 13 and that is why we get higher temperatures than we got at the I4 PWR analysis.
15 MR. RAJ SEHGAL: Is.that true in AECs, also; you're a g - 16 melting very fast?- 17 MR. DENNING: Yes. That's a fairly fast meltdown, 18 but that is really.the same reason-in that one, too, in 1 E I8 comparison to the PWR sequences.
.=
20
.MR. VOGEL:- With regard to MARCH 2, are we tracking 21 where the water is yet?
22
-l MR. DENNING: In the BWR?~
23 - MR. VOGEL: In BWR or PWR. 24 . MR. DEMNING: I wasn't sure what you meant. We've s- 25 always tracked vuere the water is, but I wasn't sure exactly
MMt]lf?9:12 101 I- 'what'you meant. 2 In the BWR controls, controls occur largely on 3 ' water levels within the reactor coolant system. And we do -- 4 MR. VOGEL: I mean, during the course of the 5- accident, where the water. is , it doesn't seem to be ever 6 called out in.the reports. 7 MR. DENNING: I'm not sure. Give me an example of 8 what type of thing you're thinking of.
'8 MR. VOGEL: Where'the water is during the course 10 of the accident.
11 MR. CZYBULSKIS: Which particular water, the primary 12
. system? Containment?
13 MR. VOGEL: When they'get mixed up -- I don't care 14 -- the total water is what I'm after. 15 MR. DENNING: What you're'saying is you would like o
- j. 16 to see how much water is on the floor, how much water is in o
II the -- I 18 MR. VOGEL:. Yes, how much is back in the reactor i 18 I vessel. E
. 30 l MR. DENNING: Certainly, all of that is tracked and 21
{- could be provided. 3 22 '
. MR. VOGEL: I think that would be almost as import- ,
8- ant as the fission product location. 24
-MR. DENNING: Also, for your information, the energy
- 26 balances are printed out as a function of time in MARCH-2, to
- 102 4
MMidl.9:13: I~ . to. aid 1the user-in knowing where various energies are-going.
~
2- -MR..VOGEL: Now,-let me ask a..related question. 3' During the time'in which the hydrogen is' generated, is the 4 ' bottom' of: the ' reactor ; vessel' filled with water?
~cnd 9-
.5-6~
.q-
.8-t 9-10 11~
12
-13 14 a
5 15 v [. '16 4
- 17 -
18 r i g 19 g ir #~
? ' 21 -
a M
.E-
, 23
- a. 24
.j y
-Q,f 26
----.r,w,,_j._._..._,.._.._,_s.,_,.
;;q C )c .
f^ 1- MR. DENNING: Yes. For.the in-core period, yes, 2 you are right. There is water in that region.
'3 MR. VOGEL: Is most of the hydrogen generated 4
during that period? 5
-MR.-DENNING: It depends upon the accident
~
6 sequence. 7- MR. VOGE;L: In, for example, TC? 8-MR. DENNING: I-think the answer is no, although l 9 an' important atnount of . hydrogen is generated. 10 MR. CZYBOLSKIS: Obviously, the timing of tne
.i 11 nydrogen generation varies. In the case of TC or TW, you'll.
12-get significant nydrogen generation as the water level 13 l
; drops and the core proceeds to heat up and melt.
I'! You will get additional nydrogen generation O 15i i aftier the core slumping begins. In the case of the AE i 16 ! sequance, you[will get very level' hydrogen generation up to 17 i ~the point before slumping because the core has been i la ! uncove' red and there.is very little steam in there. , 19 When the core starts to slump into the bottom 23 head, you will-get large amounts of steam generated which l 21
' flow out past the core and lead to significant amounts of 22 l- oxidation, hydrogen generation.
l 23 , MR. VOGEL: One of'the things that was puzzling l' } 24 'us as we. read some of enese sequences was unless you call ! 25 upon the water in the bottom of.the reactor vessel, there TAYLOE A SSO CI A TES 1625 I Street, N.W. - Suite 1004' W ashington, D.C; 20006 (202) 293-3950 k.
w- , . ~ . . : z. .q ..s. . o, r . i: ' GL%
- z.
Q. '. .. { ' s J K _%^ isn't.enoughfsteam_available by a large amount to react
~
'l -
, :2 with the zirconium to give the-hydrogen which is predicted.
3 bdt. CZYBULSKIS: ,That is very correct.
. s .
4 MR.'VCGEL: This ineans' you are coiling out water 5 from the bottom ofsthe reactor. vessel.
- 5. X
. g MR. DENNING: Right. )
7 MR. VOGhL:' I have ne further comment, since I ^ ( i' 8 '
~
~have got the microphone. And'it's more o A a general :{
i
- comment.
We-are seeing a lot of.-- th'e results of a lot of g 10 codes and so on. It seems to me for.this global peer review 11 tobehelpful,weneedtohavesomeprecedingspecialisg[> 12 U
. peer' review with perhaps reports to this committee. tk 13 For' example, a problem of when the containment l'- fails and so on is an extremely significant one and also g
15 one for the specialists. As I look around at the l' composition of tne' peer review ~ group, I don't -- maybe I do 17- _somebody a disservice, but I don't'really~see experts in 18 - that particular area. 19 And this we have also the problem of having six 20
.or seven codes back to back.here. It's not'the fault of 21 Battelle,' I am sure, but we really. don't know wnat's in the 22 codes-yet in many of them. And we really need an earlier 23 Jongoing peer review of each one of these codes as to the e '24 pnysical principles.modelec, the coding, whether the codes- ,
25 are properly interfaced and so on, for us to really be able T A Y L OE- ASSO CIA TES
,_7 162 5 I Street, N. W . .- Suite 1004 W ashington, D.C. 20006 !
(202) 293 3950 e .c 4.M e d , J-.i.J lpsm e4,6s., A %. 4
105
~
r. l 11 toldo something meaningful. _ f.- 2-3( o MR. COOPER:- A couple of small points. It might 3 be worth checking thel carbon bed for aerosol filtration
~'
-also, whether or not it's going to get' clogged.
~
'5 The'other thing is.when-the HEPA filters fail, 6
you'll have a greatly reduced resistance; therefore, you 7 should get according to_the' fan curves in bacs a v 8 substantial increase in airflow through them. 9
'MR. PETRANGELI: Petrangeli from Italy.
10 I would ask if you considered during the TC 11
. sequence the possible' activation of standby 12 gases in a liquid control systems --
13 MR. GENNING: The liquia control system? 14 MR. PETRANGELI:- Yes. CL 15 MR. D8NNING: Sie assumed that that system has 16 failed to get into that sequence. So we didn't consider 17 the effect of it. 18 MR. PETRANGELI: _The second question is, if you 19 consider the same system and those of the control rod 20 cooling water for the other sequences, lixe the AE 21
, sequence, not a shutdown means but a means for adding 22 cooling _ water, I believe the difference is not so much 23 because.the flow rates are not very high but'there is 24 evidence that' operators will tend to use the system in case.
25 of core. cooling problems; as an example,-tne Browns Ferry
#~
TAYLOE 'A SS O CI A TE S 4
~162 5 I Street,- N. W . Suite 1004 I W ashington, D.C. 20006 (202) 293-3950
~
o n4 _ _ Ap ; 106 l1 accident, the accident with what operators tried to do.with g ,
.2-
' ~
~tne sources'of water.
'3 IMR. DENNINGi Yes,1we certainly recognize this. I;
~ don't' care let Steve Hodge'get the microphone from Oak 5- Ridge or he'd spend the rest of the day talking about SASS 6~ : analyses on that.-.
2 i ~ It was not our intent to'look at the' 8 probacilities of sequences of tnis type, out to look at the,
=' - consequences of these sequences once we got into-the fai$ed
' condition'..
' 11.
MR. REYNOLDS: 'In'your listing.of key-12i - uncertainties,'I wondered if the gradual slump ~in MARCH 13- versusia'sudder . slumping of the core is.a key uncertainty. 14' .Does'it matter.very much to-any of these sequences whether O 15- you have a Jgradual or very' sudden. slump? 16 . MR . DENNING: Yes, __ it does matter whether we have
'17 a gradual or very!sudde'nislump. It affects thel timing of
~
la
, containment failure. in the AE ! sequence, and it- af f ects -the -
I19 turnover of;the temperatures in the system.in-the other 20J sequences. 21' .And'it's a broader.questionTthan just that, _the 22 _-graaual; slumping ~versus the coherent slump. But there's a 23 :wnole' spectrum of different conditions!that are.possible 24 for slumpingf benavior: that would' affect th'e thermal
- 25' jhydfaulic' conditions ~ in th'e reactor coolant system .
. And
~
~
jd '
'TAYLOE A S S O C I A T E S .-
'162 5 I Street, N.W. - Suite 1004
, , W ashington, D.C. 20006 -
(202) 293-3950 M'.^% .,.,U >-M_' * ^
'*g +8',*'ud * *MP .WtW4_W -
3--9F ==' " '
,? '
# g I 'N .
9* (
107 I apparently, cesium iodide would be -- retention would be f i n 2 very sensit'ive to tnat. 3
'MR..KASTENBERG . Just an editorial comment. If I
.4 recall,.nowhwere in the manuscript'that we received did you 5-mention that the reactor-initially is inerted nitrogen.
6 somewhere you might'ention m it. And if it has no effect on 7' .anything you-have done, you should' state it. 8 MR.. DENNING: That certainly is an oversignt. MR. KASTENBERG: Just editorially. It's not 10 there. 11
. Mt . KULHMAN: I am Mike Kuhlman from Battelle. I 12 - -will be talking about the primary system retention and 13 deposition of the Peach Bottom sequences which we have l'
.lo'osed at.
15 _ g glige 3' 16
.The three, sequences are, of course,.the AE, TC, 17 'and TW. I want to talk first a little bit about the RCS 18 thermal hydraulic Characteristics as they affect retention 19 of materials in the primary system; talk briefly about the 20 release from core for these three sequences; and finally, 21 about what the retention factors are for the primary system 22 and wnat the characteristics of the release to the 23
; containment and co tne suppression pool are for each of the 24
. sequences and wnat --
there is currently a lot of interest 25 :in what'.tne particle- size is which is - emitted f rom the _[* j TAYLOE A SSO CIA T ES
'162 5 I Street, N. W . - Suite 1004 w ashington, D.C. 20306 (2J2) 293-3950
y _. 10e li primary system. 2 The things which are important in the thermal. 3 hydraulic characteristics Jare, of course, the temperature ; and flowrates because these determine the driving force for 5 condensation, for example, for the volatile species. The flowrates determine your residence times in the primary 7 s'y stem, wnich is - and the constant flowrate also 8
. determines the concentration of the materials.in the 9
flowin'g gas streams.
-10 -(Slide) 11 In general, the Peach Bottom analyses from MARCH 12 were similar to'wnat we saw in the Surry analyses. By that, 13 I con't mean the actual magnitudes of.the temperatures were l'-
,5 the same. They weren't..This means similar thermal 47 ,-
~ 15 ' response. We saw the typical, as ' Rich was sh'owing in his 16 codes, temperature volumes would heat up, surface
- 17. temperature woula heat.up and cool back down.
18-You'd have deposition on the surfaces during the 19-early portion of tne accident, subsequent reevolution of this material from some'of the surfaces as they heated up 21 anc it moved downstream, so it wouldn't be able to be 22 - recaptured by the same~ surfaces which have previously 23 emitted them. In that sense, they were similar-to what we 2* - saw for Surry.
~
25 -The gas flowrates we see in these sequences --- ^ TAYLOE ASSO CI A T ES 16 2 5 I Stre e t, N. W . - Suite - 10 0 4 ' W ashington, D.C. 20006
'(2J2) 293-39 53
.~ , - ~,
7r,
.- . - . -n c
109 1: and I_will present a few rough curves of what these
- g 3 2 flowrates:look like because they are very important -- are
') 1similar in. magnitude to what we saw in Surry. But the timing is different. And this has very significant 5 implications on the primary system behavior and in terms of retention of these different species.
. 7. And finally,.I would like to point out, since we 3
are using different coremelting model, it does lead to a 9 different, thermal history for the core. By that I mean the 10 temper'ature profile of'the core as a function of time is 11
'different from before.
12
~
This leads to difference in tne calculated 13' release-rate'from core and also leads'to differences in the l'~ flowrates through the -primary system. So we will see how 15 tnese things affect the calculations of the-TRAPMELF code. 16 (Slide) 17 Ihe timing of various events of interest in the la primary system for AE, the large break very quickly 19 initiated melt,-the vessel dryout is.the' time'in which the 20 water has essentially been-all. boiled out of the primary 21 system'or actually out of the lower head of the pressure 22 vessel._ 23: And the vessel failure times are indicated nere. 2' What I wanted to point out is for each of these three 25i sequences, we have what I would call a. flushing out of the y." j TAYLOE ASSO CI A T ES ,
- 162 5 I Street, N.w. - Suite 1004
.e W ashington,'O.C. 20006 (202) 293 3950 L ,_ as.:_ , - . L..
c m, _
.q 110 1 entire primary; system which precedes chis dryout. This is c
! i 2 .the flush.which is due to the dribbling of material from
~
'3 the core into the lowerchead, enhanced. rate of steam
. generation.
5 And as you.will see in the next couple of
. slides,.there's a.very much enhanced steam flowrate which 7
brings about sweeping out of all the material in the 8 primary system into the containment, gives it a very limited residence time in.the primary system. 10 Another thing to note:here is the duration 11 between vessel dryout which is essentially the end of 12 significant flows in the primary system and the time of 13 vessel failure. Here we've-got-a period of approximately 80 l'
~ minutes in which the molten core is just sitting in the 15 bottom of.the' reactor vessel, continuing to emit.
16 It cools out during the slumping process, then 17 heats back up and begins'reemitting material. The materials 18 it's emitting,'of' course,-are the low-volatility species, 19 the aerosol-forming species, because by the time you reach 20 tnis, . you have already; essentially; exhausted your cesium 21 and iodine and what you're going.to_get from-the tellurium 22 inventory. 23- (Slide) 24 'I snow a couple of flowrate calculations. These 25 come from the MARCH analyses for.AE and TC' sequences. This r' s '~ ,
, TAYLOE ASSO CIA TES
_162 5 ! Street, N. W . - Suite 1004 W ashington, D.C. 20006 (202) 293-3950
. . _ . ___ _ _ ._ _ . . _.. a a 2 . ~ . . . . _ . . . , - . - _ , .--
m' .m.m -
, s
_' ~- . _. _. , , . _ - . - -_. ... # _ . .x
. :,2; ' - . . ~
^
111 p. 11 is flowrate/in pounds.of' steam-perfsecond.from the. core. So 2' you start /off-with in'the neighborhood of 10 pounds per (A(.[
~
3 second-for1the TC sequence. During the slumpingcprocess 4 you're upfwell'over a couple of hundred pounds per'second
' ^ '
5 ~ for afperiod'of:about -- in the neighborhood of several 6- ' hundreds-of seconds to a;th'ousand seconds. 7' Then after that process is, finished.and you 8- reach the vessel dryout time, you then have a very minimal-
'l
'flowrate whichTis due only'to gas expansion.
' 10 :
For the AE, you-even have'a more exaggerated 11
-case lwith la- very low' flow, then'a very sharp spike of flow 12: which lasts'for about 600-some-odd seconds.
13 So what you'have during this period is aLvery, , 14 efficient sweeping 1out.of tne. primary system contents of
.O- 15
_anything that was suspendeo in'the system up to that' time. 16 (Slide)- 17-Taking, in7a sense, the inverse of these
- 18 flowrate curves, we'have residence times here for tne:AE -
, 19 ' sequence.- As you recall, tne spike in the: previous' curve 20 was-down in this neighborhood ofLtime, 1,500 to 2,000 1 21 seconds.-
- 22
-You have reasonable sorts of residence-times for a ', 23 aerosol retention and cesium iodine retention-here, some 24 fluctuations in'the flowrate. And here's your increasing 251 flowrate due to the core continuing a slump on a 4 .
't?O T A-Y L O E ASSO CI A TES
_ 16 2 5 I Street, N.W . .. Suite 1004 w ashington, 0.C. 20006
~ (202) 293 3950
_- % ,.e - e,3 s -m +.-s s- ' e = - ' ' ' "-** *
, -Ws 112 1
node-by-node basis-till finally down a very high flowrate,
" Cj 2
~
'very;short residence times; finally,1the_ vessel dryout. And
'3 you have, for.all intents, a residence time wnich lasts up to.the' point of vessel. failure. This is for the AE 5 ' sequence'.
6 (slide) 7 TC is a similar shape curve. ,You can see again-a you're in the many hundreds-to~ thousands of seconds
- initially (during the low flowrate. Slumping occurs. Short to residence. time. In the TC we see a small flow following 11
- this major slumping. Then you get down to a stagnant 12 primary system.
13 (slide) 3 l' If we look'at the CORSOR release rates, these
- e. ,
'# 15 are predictions based on the 077 to release rate 16 coefficients,:with,one. notable exception: . the tellurium 17 release rates are now being calculated.taking ady'antage of 18 tne information provided by Oak Ridge and which relies on 19 the extent of zirconium oxidation at each node in the core.
20 The unoxidized zirconium.is a. strong getter, if 21 you will, for the tellurium and results in a much reduced 22 tellurium release rate until any particular node of the 23 core is oxidized beyond 90 percent oxidation zirconium, at 28 - which point the rate of release is. enhanced by about a 25 - factorfof 40. 1 r l g TAYLOE ASSO CI A TES 162 5 I Street, N.W. - Suite 1004 W ashington, D.C. 20006 (202) 293-3950
# g ..+w+ x
. Au =:i .= a ; -u . I =3
~
y ;=.3:y agg .- -
, ,, wa.;.u = -2. . : ,
.g i
~ b
%M ;
^
3_. , ,:-: ; ' r s
.w For.lthe AE sequence,1-;the rele'ase rates.-- allImy:
~
- '.l- I
., .~
N b; 6 2' " times-here are inizeros)measuredifrom' time of coremelting,
~
a TThat's.the only timescale that: matters here at the primary
~
13
"- c- 4 :
.s ys t em . -
7~.
~
~ 15 ' You can see.the' essential exhaustion of the
'*' ' ~
cesiumiinventory:and the iodinelinventory relatively early. 27 1
~
Tellurium isi. emitted at~very' low rates relative to waat we - 81 ; calculated.for Surry.c 9- LThe ' aerosol ^ release' rate, once we -get started-
.w '10- nere after<about ~ 800-900 seconds,!is not exhibitin'g
-31 enormou's fluctuations. IThis~ period is the period of:very
- 12 minimum retentio'n timeTin'the primary system,.which I 13 % owe'd youion th'e previous: graph. So you?have' hundreds of
'18 grams (perLsecond beingLemittedI by thELcore.
, h.
15 fYou willisee at-therend of this talk that this r16 - ma'terial'reacnes~its'exodusjfrom the primary' system very
~
17 near'its primary particle size due.to-the--extremely short 18! residence time. Therelis?no way to; effectively retain this' 19f material.. 2oi .Down here from this poi'nt on we nave material 21-
.whichfis emitted into essentially ~a stagnant primary '
22-system,fwnich gives yo'u a fairly.long retention time,- good, 23 possibilities for-coagul' tion of growth and gravitational a 24t"iremovallof.the.. material in the primary system. 25 - (Slide)
~
, .y
-pd.- .: T A Y L O E . ASSO CIA TES
'*' ^
~142 5. I Street, N.W .' Suite 1004 W ashington, . D.C . 20 006.
--(202) 293 3950 1
e y m ;,p s a s. a w 5 -, d - a h--v s * %
.- . L :. =:- . . . - . , . . ~ -
7~ . 114 1 MR. CASTLEMAN: This is a very minor point, but I .' /5. assume you mean total iodine here? 3- MR. KULdMAN:- Yes. :Just an' iodine release.
' Mike.Hazen, Stone & Webster.
~MR..HAZEN:
5
~
I would like to ask a question about the release
' A're those release rates constant in time step to rates.
7' - time step?. 8 MR. KULHMAN: .What we're doing is taking CORSOR , 9 cc,de predictions wh'ich are based on, I believe, for AE, I 10 celieve we had' quarter or half-minute time steps in MARCH. 11 At'each' time' step the core release rate for each node is 1,2 each of tne 240 nodes is calculating. This is done 13 .-throughout the course of the accident. l' We then examined the' cumulative release from the G 15 core and break it'down into 20 regions of essentially 16 -constant -- whien are treated as having a-constant release 17 rate over the period. We integrate these-rates t.o check it 18 ; against what the CORSOR code predictions are at the end of 19 each of these time intervals to'make sure we are within 20 about a percent. 21 ! (Slide)
.l 22 Let me just show you oriefly for the otner 23 sequence of interest, the TC. Once again we-have very 24 sign _ficant aerosol generation during that very low 25 residence time period. Then, interestingly enough, the fAYLOE ASSO CI A TES
[Y f 162 5 I Street, N.W . - Suite 1004 W ashington, D.C. 20006 (202) 293 3950
- = ==
- :: ~ . - - . _. =. ~ ,a; - _= .: , = . . . . . a 115 l i
I aerosol genera' tion drops ~and then begins to come up once 1 2 again.just_ prior to vessel' failure. 3 Again, as you can.see, the cesium'and iodine Linventories are about exhausted prior to this very high flowrate' regime. 6 (Sliae) 7 TW really hasinothing very new to say. By way of
- 3' information, what is-emitted from the core is of interest .
9 L'These are the fractions!which are emitted prior to vessel 10- dryout, ~and I will show you in a second the key species. 11 emitted-prior to-vessal failure. 12 The only point I wanted to make here is that you 13 Jare really, as you can see, emitting 80 percent,-roughly, l' of your more volatile species prior to' vessel'dryout, and
' 2;3 .
13 you have essentially your complete initial inventory of the l' other materials still available-for release. 17 (Slide) . 18 ' hhat these figures are"are for cesium. These are 19 masses in kilograms of the species which'are released by 20-tne time of vessel dryout,' released.from the core, not from 21 the primary system necessarily,'and pr'ior'to vessel 22 - failure, dryout andifailure times. The releases of.dryout 23~ and vessel failure time are-indicated here. 24 Just to give.you an indication of relative 25 importance-of tne material which is released during that (~~ TAfLOE ASSO CIA TES 162 5 l' Street, N. W . - Suite 2004 W ashington, D.C. 20006 (202) 293-3950 q r e-+=- a e--
= w. . -m .. = -
x,. -= , 116 15 : stagnant phase of.the sequences, here we have more than m
- 2
- half of.the aerosol material =is generated and pumped into a 3
-3
~stagnantiprimary. system, whereas;for the TW and TC cases we
-hav'e a much reduced'importance of that period of the 5 'acc'ident. ~
Again, the Impre volatile species are already out 7
.of the system by the time of: vessel dryout, a-(Slide) 9 As we~have seen'with the Surry analyses, once 10 lagain the_ aerosol _ composition 11s dominated far and away by 11 the' ten. Fission-products are on an expanded scale by a 12 factor of 10, so it says no more than 3 percent of the 13' aerosol initially is ande up of the moly and acout 2
'l' percent barium. _And-these fractions are.actually lower O 15 during the-course of the accident. sequence. .Again, just to 16 point out tne importance of the non-fission product 17 inventory for the aerosols.-
18-(Slide) 19 Beginning to look at the TRAPMELT predictions of 20
- wnat is retained in the primary. system for the AE sequence.
21 By way of reminder, the key areas-for retention are the 22 core region for the aerosols only,-the steam separators.and 23 ' tne' lower annulus region on the'way to the jet pump intake. 24 JThe total columns represent what has been 25 ' released from the core as a function of time. The total f' i 1
~
' TAYLDE ASSO CIA TES
. 16 2 5 I s tre e t, N. w . - S uit e 100 4 W ashington, D.C. 20006
-(202) 293-3950 m <#M ,-,,.<e% _ % +- a & 4 *-
w - --- : , ;- - ,a - w :. : 1-117 b. L L r . .. l'
' retained-is th'e mass of that material which is retained
[D
- s. ~ '2_ - somewhere in~the' primary system.
3 This. time represents the vessel dryout time 4 approximately. You_can see nothing.new has taken place with 5-
-the vapor ~ species.- .Yet-the aerosol retention really begins 6
to_take off'as-you.begin to age that contained aerosol. You t. 7 will see tnat perhaps more dramatically when it is a expressed in terms of a -retention factor for the primary b l 9 system. 10 (Slide) i 11 This is just a -- f 12 Castleman,-Penn State. MR.~ CASTLEMAN: l 13 ( On that previous viewgraph, you had one column ( l' of numoers going up,and down. Are they not cumulative 15 numbers? That's cesium hydroxide, I am a'little confused. I l' was thinking of the cesium hydroxide retained. I guess you 17' see the same. thing in several-columns going up and down. 18 (Slide) 19 MR. KULHMAN: i should have pointed that out. 20 This-says that the time 570 seconds aftert melt-we've-21^
-returned ~33.5 kilograms-of'the cesium. hydroxide. he have-22 emitted 81.7. The emitted columns are,.of course, 23 Lcumulative. The retained.is a function of time. This i
24 . material'is free to.reevaporate.
~
j 25 MR. CASTLEMAN: So you are accounting for
. TAYLOE ASSO CI A T ES' 16 2 5 I Street, N. W . - Suite 1004 W ashington, D.C. 20006 (202) 293 3950
. . - - +
._ . . _s _m . . . . . - . . . . - _
IlG t #1 reevaporation.-
- r.
2 MR. KOLHMAN: Right. Should :this naterial be 3' _locatedLin the steam separators, which for the AE sequence,
- you get'only-up to about 1,000 degrees F., I believe. If 5
_they get above'that,.then you can see the reevolution of T 6 the cesium hydroxide _taking place. 7 For the cesium iodide.as well, you see the 8' naterial is retained more. effectively initially and the , 9 masses 1present on the surface actually decreases with time
'10-for a bit. The aerosol you will never see that, because we 11 have no resuspension mechanisms acting on the aerosol.
12-(Slide) 13 ~ These are the same numbers but expressed 14 -differently-for the AE sequence. Again, this is time of O 15 vessel dryout, so this is.tne essential 1y stagnant period , 16 of the accident. 17 If you look at the retention factor, wnich I 18
.should define as being the' total-mass retained in the 19 primary system divided by the total mass which has been 20 released from the core at that time, so it's not.really a 21 - true efficiency, but'it nas a similar flavor.
22 -The retention factor for aerosol is not really 23 changing greatly enrough this period, but after vessel 24 dryout you see the trend from 30 percent retention up to 70 25 percent retention which is. experienced just prior to vessel-( j, TAYLOE ASSO CI A TES 162 5 I street, N.w. - Suite 1004 w ashington, D.C. 20006 (202) 293-3950
.- . . .~ ,
= - %, m ___ _ .~ . _ ., ... a 119 1 ' failure. 'And most of this'obviously is taking place.in the 7
( 2 core regioni since'the material emitted into the stagnant 3' system doesn't reach any.other control' volumes in the 4- ~ primary system.
.5 :For the cesium iodide, cesium' hydroxide, it 6 turns outLthe lower annulus;is'where this material 7
~ ~
ultimately resides --' ultimately meaning at the time of. d ; vessel. failure. A lot of:this mater'iallwas deposited 9 previously in the steam. separators, reevolved as'they 10 . heated up, and then subsequently deposited in the' lower 11 annulus. - 12 In terms of a bottom line figure, these 18 and 13 -19. percent retentions :for cesium and iodine and 14 percent 16- retention-for tellurium and 70 for the aerosol, O 15 characterize the primary system. behavior for the.AE
-16 sequence.
=17 Again, with respect to the. tellurium, we should la keep in' mind that only~about a third of the tellurium 19 inventory is' released in the core.
20 MR. KELLY: Jim Kelly from the University of 21 Virginia. 22 I notice that the numbers in this table are 23 significantly different from the numbers in the handout.'Do 24-you discuss the changes in the model that were respons_ible 25 for --
'TAYLOE ASSO CIA T ES 16 2 5 I Stre et, N. W . - Suite 100 4 W ashington, D.C. 20006 (202) 293 3950 L
r w % .. m- ..x. _ a _ ~~ . _ ;;2. .- _ _....u ' 2 - 130 I MR..KULHMAN:-No changes in the model. Changes in T. 2. . the inputito the model. What you have was performed with
'3~ two different.setsLofLinput.--One wan, there was a different 4
emission rate of materials used in those analyses. And 5 secondly, we used a different-thermal ~ hydraulic input data 6
-for the. set of-information.
7 MR. GI6SEKE: .Il-forgot.to mention in the
;s' iintroduction that we have-gone back to some of these and ichecked over the~ assumptions;and tne numbers and have
~
10
. revised some of the figures that appear in Chapterj7, 11 .primarily.
{ 12 We have sent'out in the mail revised Chapter 6. 13 _And you probably have-.that.- Plus, I believe, a lit'tle bit l' of Chapter 17._also. 15- We have available today, or will have, ainew l-16 Chapter 7, totally new Chapter 7. It includes the numbers l 17 'that are being' presented, as well as - _well, it's the
\
~ \
la total story-on Chapter-7'that wNr have to'date. And that i 19 will be handed out.. 20 21 4 22 23 24' 25
-('". TAYLOE ASSO CIA TES 162 5 I Street, N. W . - Suite 1004
~ W ashington, D.C.' 20006 (202) 293 395J i l
;je
vigw2_ g. a. - -. . , , . _ _ _w- . ___ a. _ - . _ - 2 ,. i] LTA3 195i ^ e : 121 MMgjl-'11:1
# '1:
; MR. ~ SHERRY : Rick Sherry, NUS.
k
-, . ty 2- ~ Mike, fin' comparing (the tellurium ~ retention-factors 3 to the cesium' iodide and cesium hyydroxide, tellurium reten-
~
-4; . tion is lower for these sequences. This wasn't.the case for 57 the Surry ' analysis, where, in-general, the tellurium reten-6 ition factors were much higher:than the cesium or iodine.
7- Can you comment.onlthe reason for this? 8- . MR. .' KUHLMAN: .'I'think:I can.- 8 'This data.is all' fairly' fresh and needs some 10 t digestion. Yet,:a' couple'of' differences from the Surry 11 analysis.-- 12 First of all, Surry, we were releasing tellurium
- 13 at:the rate which we had published in 0772, which is a 14 considerably higher rate than the rate of release for these f
't 15 1 accidents, which take into account, in a rudimentary fashion
.{ 16 at-least, the! inhibition of tellurium release due to the C.
}3 17 zirconium clado B8t we = are releasing tellurium'at a much
}1 'Is ' lower rate now.
- 8 18 We're also releasingit really relative to the 3
I #
.y cesium, cesium and~ iodine. Tellurium is. coming out of-the II core later in the sequence. And there's not as much tellurium E 3 22
~
f coming out, also. Of course, it's the first-order process. [ 23 - So,;as'the tellurium,-the rate of-it being driven 2 24 toithe surface is. dependent on the-concentration at any time. 1 ! I . ~ \L,,/ 25 .This tellurium release is occurring more, I~believe, during l: l l i'
-. ~ - . - . . .-- .- - , . . ,
MMgj 11162
'I that' low low residence time period'of the' accident, near the p
2 high flow rate, Since,the coreLis hotter at that time, more 3 .of the. tellurium release is occurring at that 1500 to 2000 4 So, that window, where you'have'very low residence time. 5 ' material', .then, is less subject to the chem absorption even
'6
~
~
that-it'would be had it been released concurrently with 7
' cesium and iodine.
-8 ~I think it's all'a question of timing. Again, when!this. material.is released at the worst possible time'for
'10~ it to be released, which is as the core is hot, as some of 11
~
the material is slumping from the core to give you the high 12 flow rate. 13 MR. VOGEL: I think, just-from the way you phrased 14 your reaction to the tellurium question just now, leads one 15 to suspect here you are really believing in this tellurium
;[ 16 behavior.
0 17 g And-before we let tellurium behavior get embedded 18 too thoroughly in the lore of reactor safety, it might be a 18
! good idea to recall that, first, we haven't seen the Oak Ridge data, but I'think -- I'm familiar with most of the experiments, 21 5 and a ' lot of the. experiments on tellurium are on a very small s
_j 22 scale. 'I think there's going to be several more chapter and 23 verses on the tellurium story before we really come to a firm
,: agreement -on that point. l (f g '
The te~11urium story is by MR. SILBERBERG: Yes. l
e :a p 3.wp , g .y. w w n - ,
. ,[MNfjl@ll:3 , 0 1 y 123 11 -no meansrfinal. 'We certAinly-don't want.to;give that' y j, . .
,E =2 But .what the Battelle people are 'doing are work-2
, Timhressio'n.'
3 2 ing:rightJupito the razor edge,~if'you will, of the state of
.a 4 itlieLtedhnology as we.know it.today.
Unfortunately, that's
. 5- .where we are.'. We,are:trying to.make, if you will -- make use 8: --ofLthatjstate of the art..
7 Now, weihave,a' tellurium report that we will have
. al ..available later-today,;this afternoon.. And you are all 8' commended to read it tonight.
10 7 Andcif-therezare questions,-Bob.Wishner!from [11 ; LOak Ridge has graciously consentedLto answer such questions-12 - citt this stage. And_of. course this'will'be~an evolving. story 13-on Itellurium. 14 MR. VOGEL: Yes. The reason for:my comment, some--
-5; times the separation between the fact'andzhypothesis in'a b 18 - complicated analysis like-this gets lost.
17 little worried that hypothesis was drifting 7 ,,,.a 18 . over into fact here.
-t
' I8
! MR. CASTLEMAN: -I would.just>like to second the-
.E I
~
A point that Dick Vogeleis making, because I noticed on page 5.6 [
=_ 5 :.
21 in the document that it was stated that the tellurium' release-jk 22- 1.is.not well1 quantified.- But then,.you go, later, 'to page-632, 23 - fandiit makes it sound like this whole problem is solved'and i } .fr 24 it isiquantified,Dand I think it's a :little misleading. k,) 26 - MR.: SILBERBERG: iSee what happens.when you get a i h: %
.c..-+ . . . ... .. .- .
~++ ~ . -. . . . . . - . . . . , . . -_ . . . . . - . _ . . . _ .
MMtji' 1134! 124
- 1. littl'e' data. It.goes a long way.
. ,. m 2 (Laughter.) - 3 MR. JOHNSON: Johnson, Oregon.
'4 ' Mike, could you clarify the way you handled the c5 cesium hydroxide there? When you= talk about its reevaporation
.6 - or reevolution, is this the cesium hydroxide?
7 MR. KUHLMAN: Yes, it's a good point. 8 MR. JOHNSON: Why would you think it stayed cesiu'm 8- hydroxide? In that= form, it's one of the more reactive -- 10 MR. KUHd4AN: We have incorporated into the TRAP
~
11 : analyses a chem absorption phenomenon as well as to the 12 condensation for cesium hydroxide. 13 MR. JOHNSON: It's a little more than chem absorp-14- It's a formation'of new compounds in the system. tion. 15 MR. KUHLMAN: Once the new chem absorption is r'emov-0 l 16 ' ed from possibly ever reevolving,.so there's two mechanisms 0 ~ 17 ' operating on cesium hydroxide, it can react with the surface 2 18 - You'll see some examples of and be permanently-bonded there. t {;. 18 that in some of the other sequences. Also, there's another 1 [ 20 portion of.it which'is allowed to condense and reevaporate. 21 fHow realistic that can be, I don't know. { 3 22 If it is condensing,.I agree the odds are good that [ 23 the condensed material.is going to react with the surface 24 ^g; .when it' condenses with cesium hydroxide. .Q y
. MR. JOHNSON: Absolutely.
vus . w.e -= _ % , ;y- m._.. ..a w.m ,_, m_ m._._,.._______._.. . WM* ]44ft[ ~ ~
~
~
125
. - s-.'.
it : MR.-SILBERBERG: ?JAre you'saying'you don't allow for I["
-q 12- 'cesiumLhydroxide' interaction with the_ surface now?;
3 - MR.:KUHLMAN:- We have two modes'of cesium hydroxide
- 4' Linteraction;with the: surface. One:is chem absorption, which
;5
~
is_ irreversible. .One is. condensation, which is subject to s .; subsequent reevaporation. 7- MR.' JOHNSON: .You-have to be~ careful'what surface 4
'8 you are condensing on..
8- MR. CASTLEMAN:' ?Maybe this is jumping . ahead 'a. little 10 ~ : bit of'what you are. intending to cover, but~I noticed, in a 11~ ifew of.these points raised by.. cesium, iodide'and cesium hydrox-121 :ide'on page1720,. Table 7.8, where you-give:alfinal distribu-13 " tion of'the-spec'ies, both; cesium' iodide-and cesium hydroxide L14' released to the environment are almost'the.same fraction, .19 15
-versus .-20, which also surprised me a little bit in view o.f 3
{ 16 the -chemical difference 'and the fact th'at cesium : hydroxide s 17
.- : presumably -- chem absorbs, and cesium' iodide-doesn't.
'l 18 Is'there a.reasonsfor;that? Is that because what 1.
! 18 gets'releasedis'moreassociatedw$.ththeaerosol? Or is it
- g.
V 8' just ' insensitivity to these mechanisms and chemical dif'fer-21 . Lences? 3 22' MR. KUHLMAN: I was surprised at that myself. I 23 thinkIthat was for the-AE-sequence. For".the TC, you see a
' 24 -
vast difference :in the' retention 'of . the two. _w
# - ^
MR.-CASTLEMAN: Why'is that? I'm not sufficiently,
+ -yg.-+~---s...s ,,a n . ~ . - . -
-- w ; a ; . . -
- MMajl-126
-1136 1; MR. KUHLMAN: I haven't had the time to look at all
. 2 the data and track all this. material through the primary 11 system.~ There are a-lot that are not immediately apparent 4 aspects to this data as to what is really taking place, and 5 I frankly can't answer that question.
6 MR. JOHNSON: .Aren't the-time relationships almost 7' the same here, the thermal hydraulic characteristics? 8 MR. KUHLMAN: 'Yes. We're talking about just one 9 sequence; the cesium and iodine releases from the core are 10 simultaneous essentially. 'So, you'would expect them to 11 experience the same conditions as they flow through the 12 primary systems; yes. () 13 14
!!aybe I'm missihg the point.
.MR. JOHNSON: I'm just going back to Castleman's
;5 15 point'.
4
] 16 MR. KUHLMAN: That makes1the TC hard to explain.
f 17 MR. JOHNSON: Well, maybe that's a starting point. 9 18 MR. ROE: I had a question on presentation. I look
# 19 at a. table like this. I like to try to add up numbers to 20 get to one,.or unity. I have a hard time interpreting this.
21 { A question, then, might I.suggest, depending on 5 j.- 22 the ancwer to the question: Would it be useful to have a
- 23 table presented in' terms of fraction-of initial inventory 24 presentation of these different nuclides? In that, you can 3 then keep track of the distribution,in a fractional sense,
n- . + . . . .
.- .. . . - . . . . ~ . . ..
MM;jlf 11:7f
' 127 1 of each species throughout:all nodal-volumes of the system,
- 3
=2 -in fact,-and what also: reaches their environment.
.3: I don't see too much of that in here, and I was 4 ' wondering.why not. .Or do you have a different philosophy 5- that you're .trying- to follow?
6 MR. KUHLMAN: Trying to keep our page charges down. 7 Tables are sort of helpful for-a lot of things when you're
.8 . dealing with four species, five-volumes, and 20 times and 9
\ five! states for.each species, they become pretty cumbersome. r So, we have hammered' around a lot of ideas just within our 10 11 grouhtotrytoexplaintosomebodyelseinthegroupwhois i 12' not familiar with these what we.are doing.
'13
. We, obviously, have'n't come up with the ideal solu-r 14 i
tion tio this yet.
\
But.we do present'it in this fashion here l .2 i
.g 15 5
to give a retention factor for what is emitted from the core
.\
-] 16 into the primary system, because that is, to me, the only t
i o 17 t number here which has some physical meaning ' to .me. 18 r If I have to multiply this, in turn, by the fraction 3 19
-. released from the core, you're not quite -there yet, and perhaps g ,
.20 you would' ultimately like'to see an efficiency out of what the
{ 21 lower annulus actually saw, how much did it retain, because 3-
- 22
{ that. turns out to be really an important parameter, because if 23 the upstream conditions change, your lower annulus would be 24 7 .seeing more material, it would be capturing more mat'erial. (.'- ' 25 It's a dilemma. I don't know of any better way, L'
a w:: . c -
-MM:jl. 128
'11:8
~
.1 other than' throwing in: tables and several different types of
{d!. 2 units, which is what we've tried to do at this point, where
~
3 we have the masses retained and masses released as a function
'4 of time:-- retention factors. And at the end, we have a 5 summary. table, so far, that has the retention mechanisms.
6 If we went to microfiche, maybe we could incorpor-7- ate all of these at once. 8 If you have.any suggestions for a table format'you-9' would like to see, I'd.be interested.in seeing it. 1 10 MR. REYNOLDS: Are some of'the low retention factors
~
- 11 that you have for cesium iodide and cesium hydroxide due to 12 the rapid evolution of steam or the rapid flow of steam? And ;
- 13 is that due to-the slumping model that you are using?
14 MR. KUHLMAN: -In part.
; 15 MR. REYNOLDS: How would it change if you had I
l 16' dumped all of the fuel down coherently? 17 MR. KUHLMAN: We will see,within the comparison i
- between these numbers and Grand Gulf, some of the sensitivity 18 I Maybe we ought to wait until g 19 =to.that mode of core slumping.
r g 20 we get to the Grand Gulf and discuss this point then. I r
' 21 think it's a good point.
{ 22 MR. REYNOLDS: Does this indicate that this
.l.
23 represents an advantage of the slumping _-- I'm sorry, this , g 24 represents a disadvantage of the gradual slumping because (~; 26 There likely would be great your retention is low. L
,_. -. ,, _~
. T EMMsjl. .
gg 11:9-1 disadvantages of coherence slumping. This would be an 2 advantage of coherence slumping. 3' MR. KUHIJ1AN: Yes, it would. 4 But another thing that coherence slumping brings 5 about is an enhanced aerosol emission rate when the core 6 reaches higher temperatures as we keep a core, which would, 7 in this model, allow the slump -- on an individual node basis, 8 if you keep altogether until 75 percent, for example, has 9 reached the molten point, your temperatures will increase, and 10 aerosol generation rate will increase as well, which will, in 11 turn, drive up your aerosol retention factors. 12 (Slide.) 13 For the AE,. if you look at the mass retained of 14 cesium iodide and cesium hydroxide in the steam separators, i 15 which are the Volume 2 and the lower annulus, which is 0-j 16 Volume 3, it has the shape you would suspect from looking at 0 l 17 the tabic, tabulated values, where the material retained in 2 3 18 reaches the peak and then rapidly reevolves that material, 19 which is then captured downstream in the lower annulus region. [ 20 So, that was just a profile of what the retention, 21 as a function of time, looks like. 5 2 22 l (Slide.) 23 Translating that into what the injection into the 24 containment or the dry well looks like, this is the cesium 'q , 25 iodide curve here; cesium hydroxide here. And the aerosol is
.. a -
=W4 jl ll:10- ;
-123
'l' .the dashed curve. So,~you can see we're pumping in roughly; j,
2 250,l300' kilograms of aerosol into the-dry well during the 3 icourse of the accident. And this all occurs during that 4 flushing phase of the sequence. 5 One other thing'I wanted to point out is that,
~
6 remember, there is'that stagnant period, which is indicated 7 here by no emissions into the dry well. This would be follow-8 ed, then, with a spike'at the end as the pressure vessel 8 fails and the material which is still suspended there is, 10 then, injected into the dry well. 11-(Slide.) 12 TC -- we don't really.need to examine this except, (_j 13 again, here's the vessel dryout times, everything subsequent 14 to this is into the stagnant period of the accident. 4 15 l3 Once again, the volatile species emission has ,
-[ 16 already stopped by that time.
o l 17 (Slide.) 3
! 18 In terms of the retention factor for TC, you have
! 18 here 11 percent of the cesium iodide; somewhat higher per-E 20 l centage for the cesium hydroxide.
21 Tellurium here reaches 80 percent retention. And 5 18 this, again, I think is due more to the timing. This is a f 23 much longer sequence than your AE was. You have a very short 24 period prior -- at the flushing period. Here you hade a much I .\
' " longer period, with the core melting -- lower retention factor
-a ,, . . , , ,
e i e e
.s~~,'
( ,., .e& l-k l l-t t -
^
I I = I i i 6
,- 3 s */ .
I r
~
l i k-I I l. s 1 i I
\ ,
t 1 , i . u,. i
, - . ~-- .. . . ,
..w . ~
-11:11- 131 i
s
.l' ifor.the aerosol, partly due to the fact that the emission
? .e, {c, 2 rate of aerosol is-lower' during this stagnant period of this 3- accident. J C2 BU. 4 (Slide.) 5 You have a similar--- 6 MR.- SHERRY: Rick Sherry. 7 Mike, your fission product and aerosol release rates, 8' whichLyou predict through the course of the' accident -- do you 8-account for any. reductions in the release rates due to changes 10 in the geometry and decreases in the material ' surface-to-11 ' volume ratios? ' l 12 l MR. KUHLMAN: No, it's not taken into account at 13 ( 311,. Rick. We're using only the March predicted temperature 14 for each node of the core. We'do.have a flag to indicate 15 whether these nodes are in or out of the, core, as a function 5 n' { 16 of time, whica could be used if we had some way to estimate ' ! 17 the effect'of bottom geometry on the emissions for any given 1 18 node. { t 8 'I8 In the interests of time, I'll get to sort of the e
\
20 bottom line here for these 'three sequences. J 21 5 (Slide.) l x-j 22 There's a couple of things this table could be i
\
23 called, fractional distribution of the species emitted from 24 the co're at the time of dessel failure. - The four species
\~; a we're principally concerned with, the vapor is meant to i
, , - . 4 , , , . , , , . . . - - . =*
,-me*-
,e
, - , ,, - - +
,. __ -. m._ _
m P orM$ 11}12i = * - 132 9' -
-q . c . , ,
+
n: Ms 11' Lindicate<theifraction-of the material which has been Md LI' ~ deposited'inithe_ primary system somewhere due to either
\8 condensation or chem-absorption, 4'and 6 percent for the f f4 i -cesium >hydroxidef a'nd 14 T for the: tellurium. And this
~
- L5 3 - would La'11L be due t'o chem absorption - aerosol term here --
~C ?is meant to indicate that portion ~of the species which'has W ,
7 been/ deposited on surfaces . af ter. : c'ondensing on aerosol-s .- .
.8,
- particles,.'which:are thenTretrained. An'd-here again, simi-t I8 ~
Llar behavior. for 'the two cesium species and throughout the 8'- -sequences. 11 -
'Then, the' suspended'~ column indicates the fraction 4'
42f .o g7 the material emitted from the core,'which is still sus-I - 13 p' ended in the primary system that'is not retained at the 14
-. time of vessel fa'ilure. 'And this is that puff release, if b I8
<you will,'which accompanies _.the1 vessel. failure. And the dash I
'I '" just indicates-mechanisms that'aren't available.
8 'gy .
- o. Here, just=the? aerosol is not depo. sited as a vapor.
I fr The TC indicates strong difference in the method <l c 1' 8 _by which.the~ce'sium iodide is retained versus the cesium _l- -
' I- I
- r. - -hydroxide. This I did'look atDin. detail.- This is almost 1
3 L21 entirely.due,to chem.absorptionyof the ces'ium hydroxide, which z,.
] _ 23
'i s not'available,;to the' cesium iodide.. -
similarly-in ths 1W case the chem absorption is what resulted
' 8C j.L < ' from the' cesium; hydroxide retention' from the vapor state.
~
.s ..
4.l) '
?. _ y,
-The other obvious' figure here is the t'ellurium chem p w n-..- +. .- a .- .- . .
, = .. . .
MM2jl 11:13. 133 1 absorption;for the longer-melt sequences, it's very high
<~
2 in'both cases. 3 MR. VOGEL: . Mike, I.think there's a problem in 4 talking across disciplines. 5 To a chemist, chem absorption is not a very strong 6- interaction, and this is what was bugging Carl -- I think 7 with' tellurium -- and particularly,.you've got chemical 8 reaction.. 8 MR. KUHLMAN: .Right. _It's not meant to imply that 10 that stuff is available to be re-evolved at all. 11 MR. ZUMWALT: I have a question about the cesium 12 hydroxide. You said part of cesium hydroxide was not con-13 sidered to be strongly absorbed. Did that enter into this 14 picture? h 15 MR. KUHLMAN: It is included in here. I would 3 [ 16 suppose approximately .01 here would have been available for 0 17
% re-evaporation here, just deposited due to condensation. The i
18 f rest of this material is irreversably deposited, reactive i 8 18 with the surface. I # a MR. ZUMWALT: Is this due to having different 21 { surfaces? 3 U MR. KUHLMAN: No, it really isn't. It's due to { 8 different rates of the two processes. There's a deposition 2- . velocity calculated for the~ cesium hydroxide. Then, there's
$..O N' 25 also a condensation rate calculated. There's two competing
~
. w.- .m i. , . . . +
w .c. , ,, . . - . . .. . ., .
, iMM:J1.=ll:14i i 134 .
.' 1- ~ mech'anisms for cesium hydroxide retention. ..
f'.x
;2 .MR. RITZMAN: :Let.me. follow that up now. There's 3 rot a' reaction mechanism between vapor deposit, cesium 4- . hydroxide,with the metallic substrate it condenses ont right?
'a - There's no. solid-state reaction?
8
'MR. KUHLMAN:- That's right.
I" And I see that as a source of-potential problem. O MR. RITZMAN: 'In reality, they probably would? MR. KUHLMAN: Probably would be, yes. 10
- MR. WA!KER: Let me ask you'one question about II - ~
.your numbers.
12 Take the first one, ' for -instance, cesium iodide.
; ~
13 Und'er AE, what it says.4/10ths of it is in the reactor sys-14 tem,'in a depository or suspended--- the rest of the stuff is
-15 out in the' containment'.
-[ 16 -
- MR'. KUHLMAN: It says 4 percent'of it ---
0 4 9 17 e MR. WALKER: . About 40 percent in those three columns l' 1 . i j 18
-- the other 60 percent.is out in the_ containment?
-1. ,
MR. KUHLMAN: That's correct.
- E
' 20 t That's.the thing,you can. sum to one.
If you find j
- - 21-
- i. ;f -what's missing,';it's.in the containment.-
.]., '22 (Laughter.), .
. MR. '. WILLIAMS : . David Williams,1Sandia.-
24 ' In the Surry. analysis, I remember I was quite j ; concerned.' As I understand.it, in both:the MERGE and-c
=
s-
~,.- ^ .[ . > m ~ .--m w ,A.._,...._., y_,,..,. , . . , , _ _ . , , , , . _ . _ . . , , ,
r r MM:jll.11:15 135 c. 1 the TRAP-melt analysis, there is no-accounting for exchange
\ - 2 of either heat or mass,= including radionuclides, between 3 There are some control volumes due to natural circulation.
4 correlations for deposition and heat transfer through natural 5 convection processes, within a control volume, not exchanged 8 within'the control volumes. And that.since these control
'I volume definitions are somewhat artificial in some cases, 8
.they don't correspond to rigorously defined physical boundar-ies that would actually block the actual circulation processes to This could lead to serious error in any sequence with a force
'II flow, where the scheme evolution was low.
12 The same thing would apply, here, to parts of the 13 sequences where we have nearly stagnant conditions. I4 Has anything more been done to address this 15 question, to determine how important it might be? 3-l 16 MR. KUHLMAN: This is a good point that you bring 0 17 g up, and it is something to which you would expect the code 3 18 f . predictions to be sensitive. 1 l a If one has misnodalized the primary system, say,
.2 20 r you could be very, very sensitive to your error in your II representation of-the system.
{ 22 The attempt has been made in these, as you pointed 28 - out, to. treat chose-volumes,which are well mixed within~them-24 rc .selves, as separate ~from one another and treat them nearly as N . ... r gg successive volumes through which everything passes. And they _ ,_u..._ _ _a.,.. - _ . _ _ . . . -.
MM:jl 11':16: 136 s I well-mixed in any given volume. But if that's not the case,
.~.
1 2 you have misformulated the problem, and the degree of errors
'8 is going-to depend on'how far you are from reality and whether 4
anything of significance is occurring. 5' Of course, tihis is part of what I think the sensi-6 tivity study aspect of this work should be addressing -- and 7 pretty 'ca'refully -- because this whole problem of . proper 8 treatment of the flows through the primary system and, indeed, through the containment is. .something you can only address 10 th' rough , I think, a well-done sensitivity study. And that's II'
~not something that we are including in the current work.
12 (Slide.) 13 I only have a couple ~of more-s'11 des which I wanted J I' to.present in terms of particle size distribution which exit c j 15 the primary system. [ t 3 l 16 For the TC sequence, as you recall, there are , I' several distinctly different flow regimes'and distinctly 3 18
*: different retention periods for the accident during the 19
! :in-vessel melting phase.
! 20
't What is shown here are mass distributions of the f aerosol. .These.are actually not in centimeters. These are 22 -
~!'- micrometers.
23
. The two times which are shown'h'ere correspond to 24 - ~
.just prior to the' flushing out of the' primary system. You U - 25 ~
'have your.very large particles. This represents'a'.well-aged
_ _ _ _ _ _ . - . _ _ . _ . . . . . . _ - , . . . _ . . . _ __ _ _ ~ _ s u . -
- -)
q MM:ji 11:17 137 1 aerosol which has'had a fairly significant source'for an fS , j 4 2 extended period of time. If you-can't read this, "the mass l
- 3. median dynamics," wh'ich is an aerodynamic diameter, is the 4 2.02 micron signature. You have 2.13. Again, this is just
;5 prior'to the sweeping out of the primary system.
6 The period during the high-flow regime, the size 7 fdistribution is represented here, which gives you an aerosol a mass-median diameter here of .02 microns and a much smaller 8 sigma G, wherelyou are:getting closer to what the primary 10 ~ particle size distribution was, where we had started with 11 _. .05 micron diameter particles.. By the time they exit the 12 primary systemlduring this period of the accident, you only 13 : -haveforthes$quence','.2 microns. 14 We've seen some cases where the diameter is even 15
) smaller during the high-flow regime. .
4 \ [ 16 (Slide.) a i l 17 To loo'k at these on another basis, look at the time
.i 18 during the core melting. Here you have the aging period, the a
19 j low-to-moderate flow rates, flushing of the reactor vessel, surges in flow abd then, again, the stagnant portion of the
.. a
.j 21 accident.
3 22-These are aerosol mass median diameters. These 23 are presented here for what is going on in the ' core region, 24 but they are not greatly different from what is actuall . 25- . showing up in the relief line.
MM:jl l11:18 138 I We have some dots in.the Grand Gulf sequence which t-
"' 2 show what's going on ~in the core and what's going on in the 3
relief line. 4 What you see primarily is a displacement. So,
'5 when this material hits the relief line, you get a large 6
spike. Then,-it comes down, tracks underneath this, because 7 a lot of this material is removed-prior to getting to the A-relief line.
. 9 That's the case-for TC. You can see a wide range 10 of aerosol sizes exiting the primary system, as you would 11 expect for^an aerosol experiencing such a wide range in 12 retention time for residence-time.
() , 14 (Slide.) This is for the AE sequence. E o j 15 Again, the initial low flow sweeping out, and then 8 16 finally the growth to.a fairly constant size up here towards 0 17
- the stagnant period of the accident.
18 - g I. thought these~were interesting results of the 19
- l- analysis in terms of what potential mechanisms are' going to 2 20 I be acting on these particles as they leave the primary system.
T 21 g That concludes my presentation. l -n
. MR. SILBERGERG: Thank you, Mike..
23 I think we'll have a few questions. Then, I think we will break at this point for ( 2
' lunch.
_. c ._ . _ ,. a. - ; ._,-.....:,....,. -, . __ - D-
_ _ _ _ _ - _ _ _ _ _ _ _ - - _ _ _ _ - - _ _ _ _ - = _ _ - -
'MMr31 11:19 139 1~
There ie still one other speaker for Peach Bottom. 2-MR. ZUMWALT: I have a question. 3 yem Zumwalt, from NCSU. 4 On Figure ~73, the particle size distribution of 5 aerosols suspended in the dry well, this is probably --r.maybe 6 early data. You know-the figure,I have in mind, it's two 7 parabolic curves. 8 Anyway, I was wondering if the curve was mislabled, 8 in that the shorter time period is listed as having larger 10 particle size. 11 MR. KUHLMAN: ~ That's what's going on in the dry wel3, 12 which is taking what'comes out of the primary system. I could i 13 see.where, c.arly on, you could,have'this larger material being i 14 injected into the dry well. I don't know what sequence that 15 is that you're looking at. 3
. l 16 MR. ZUMWALT: This is'AE.
17
- MR. .KUHLMAN: I would expect if you are, say, less
' 18
-than a thousand seconds since the start of melt that you would i
18 { have' larger sizes., Not until you get out to -- once you get i j 20 to this high-flow-regime, you then get something more closely {. 21i approximating the primary particle size. So, you can, in 3 j 22 . facti,.have a reversal. It's not.due to aging in'the dry well.
- 23 It's due.to the changing characteristics of the source to the 24 ' '
dry well. I think?that's probably what's going on. t Isn't thht right, Ken? 4 4
.-e -w
, .>*v
+ 3 WOM4& W 7'+ ** * * ** '* * * " " * * *
** * ' ~ ' ' '
l MMajl 11:20~
- 140 1
MR. LEE: That's right.
'2 MR. COBBLE: James Cobble, San Diego State Univer-3 sity.
4' I don't want to nitpick at this time what went into 5 preparing the numbers for the data base. There's one general 6
~
-omission -- maybe because there are no numbers in the data 7
base. But one general omission which might affect two or 8 three of the regimes you Were talking about for transfer of 8
-fission products -- that is, the known interaction at very 10 high temperature steam to form molecules which are essentially l 11
- hydrated in the gas phase that essentially increases the 12
- solubility.- You will affect the processes you call weak 13
($,)' chem absorption 'and possibhr lead - to teulerium and 14 ruthenium transport. 5 15 2 . Those numbers can be estimated from the data which v.
! 16 -
are available in the solubility of compounds in high tempera-17 tures. 3 y 18 Has anyone made an attempt to see whether those are
'I 18
{ significant processes? It also depends on the density of the E 20 ' l ste'am,'as well as significant temperatures. 21' Has anyone looked at that?
.3 U MR. KUHLMAN:
f We've not made any attempt to look
# at that. I don't know whether the people from Sandia or 24 0ak' Ridge 1have examined 1that question or not.
i N' 25 To my knowledge, that answer would be no. s
+ .5 e .S- % emb.4. . c - e - a .
p '
-MM:jl 11:21( 141 1 MR. COBBLE: In parlance it's called carryover, 2- maybe'a boiler word -- essentially. enhance the solubility 3 ' reaction with high-temperature. water molecules. It's well
~
4 'known -- sodium chloride. We now know it for-sodium hydroxide 5 Russians have done work on' metallic oxides. - So you can make
- 6 . estimations to see what's the type of process.- You've got
- 7. an enormous amount of steam flowing'by.
8 - For example,.your weekly chem-absorbing species -- 9 Lit may.be another steam exchange' process. . . 10 MR. SILBERBERG: We will break only one hour for 11 ' lunch, and I would like to return promptly at 1:35. 12 (Whereupon, at 12:35 p.m., the.-meeting was recessed,
- end 13 to resume.at 1:35 p.m., this same day.)
~
t.ll 14 5 15 v [. 16 0 17 d. 3 ui E. - g- 19 I 20 I 21 5
- M
- f. .
24 -
\
25 I
* # E M
- 9' %43 A aJia-W'p9' e9 % e p & NN weme F .ry,t,, M taep .,y , qey e q 4 ,' g , ,, g,, 9 g L. _ . . _
. / A-~ '-
W ff
- i48-l' AFTERNOON SESSION
- sg
$ 2 (1: 40 p.m.)
'3 MR..SILBERBERG:. Let's call the meeting to cruer, (4- please. '
5- Ken Lee will.make the closing presentation or
~
6 the next-to-the-last" presentation on Peach Bottom as f ar as
}-
7 containment transport. e (Slide) 9 MR.. LEE:- - The. primary containment is the last 10 barrierffor. fission-product. After fission. product leaves 11- the greactor coolant ; system, the-primary containment or the 12 secondary'. containment is the last' barrier the fission :
.13 product has.to pass through_before: escaping to the
<l* environment.
, O- 15-For calculating the transport of fission product 16 in' containment system,~as we-did before, we have-been using "
17 the'NAUA coce.
, 18 Just to. remind you_what we have on the board, 19 e
~all the. mechanical'~ aerosol deposition-mechanisms,-
20: gravitational. settling, ciffusional deposition.- And as a . 21= particle growth mechanism we have aerosol' agglomeration.and
- 22- steam condensation onto' particles. And as:we talked.about-
- 23. this morning,-we h' ave diffusiopharesis.
~
!24 ?And.of course,ythe source ofathe material to the-25 containment.as'wellfas-leakage out ofJthe' containment are !
': TAYLDE A SSO CI A T ES 1 z (3~ . ].
V
~ 16 2 5 - I S tr e e t, N.. W .
S uit e _' 10 0 4 - , W ashington, D.C. 20006
- l-(202) 293-3950- T-
- ..j-v..'..e 'h '- , ; . w.; . ., , 2 4 3 3 - L . ..
. , - y . ,-..
. . . -,s.+ ..e... . --
. - , , , - J. < . . . . . ~ . . . -.-.--
143 1 considered. 2-(Slide) 3 Again this morning -- 4 MR. ZUMWALT: I just have-a question. I am 5 wondering where-one can find more information about the 6 NAUALcode. I woulo like to make a suggestion of naving an 7 appendix on it. + 8
.MR. GIESEKE: There is a description available.
9 Perhaps you'could talk to Dick Vogel, who is the 0.3. 1
' 0 distributor.
11 (Laughter) ; 12 MR. SILBERBERG: If we can get a waiver from Dick 13 and from the Feueral Republic of Germany, which I suspect 1* - we could since they are.already'available -- O 15 1
- iR. V0 GEL: ' Vic would.be easier. ,
l' MR. SILBERBERG: I' guess what we;would do, we i 17 would just.make a' distribution to the entire audience here. 18 ~ In fact, that request has beenLmade before, and I think
~
I' it's probably a good idea. 20 MR. V0 GEL: I am a-little bit appalled at all 21 this attention. I didn't know, A, I was custodian of the 22 NAUA' code.
~23 MR. SILBERBERG: I have a witness in the room 24' '
that~says that was the case. He used-to work for me. 25
.MR. ROE: Let me request this go even farther. We b TAYLOE ASSO CIA TES ~~
1625 I Street, N.W. - Suite 1004 w ashington, D.C. 20006 (202) 293-3950
..,_.,4.._ _- . . . . . . . - ~ .- . - - - - - - -
~.
144 l q . nave, I notice, a writeup on TRAPMELT. I do not have one on ( 2- -MARCH. I think everyone should have descriptions as they 3 exist on the codes that are being used here.
" MR. SILBERBERG:. As you should know, Don, the
~5' various codes are in various degrees of documentation 6
. level, if you will. That does make it somewhat difficult.
7 But as soon as we can, we will try to get such a set. It 8~ will:probably start with maybe half the set. And in~some' cases we are, in effect, kind of 10-navingsto play catch-up with'it. With that, it's not 11' necessarily a justification. What w'e have to do is to 12 figure out at least for Battelle if we can get 8 days in a 13 week and'more than 24 hours in a day. I think we can 14 probably get caught up. 15 MR. VOGEL: I would-lide to add another thought 16;
.to this, since.this subject has been opened up. And that 17 is, it seems to me in a reactor safety area we nave le departea a little bit from the traditional method of 19 operation in tecnuical fields.
And thera seems to be a 20 dearto of journal articles submittea which are then 2 reviewed in the traditional journal sense and-then can be 22- .used-as references as a departure point. 23 Ano we have,iI think, some eight or nine codes. . 2* It seems to me every one of those could be.a journal 25 article and reviewed. And then after they have gone through
.(-
0
. TAYLOE ASSO CI A TES '
1625 I Street, N.W. - Suite 1004 W ashington, 0.C. 20006 (202) 293-3950
= . . . . -
.. ., . .. ... - . - , - - - . . ~
m . _ _ _. <.= . -
' . J ' ,
145 5. 1 1 m, ;such:a-procedure,-we-can say,'okay,.that's that, and use W 2" :that-from,thereJon. 3' ' MR.'SILBERBERG .Obviously, there there was a.
' 4 t' ! question of timing.
That normal process, in;effect, does really1golon. But?it'goe's.on'over a much longer time scale. 5 I (It?s'aitime. scale tihat.is. set by'itself and you can't, if-c7' youlwill',[ dictate that time scale.;.And unfortunately -- 8-T it.'s also happened. in" reactor saf ety studies in. the:past. 9 It goes: iniparallel~with and then usually goes beyond'and
' 10~
sactuallyycomes sort of at the end'of the process. And it's-11
.something we have toLoe. aware-of.
But let me note that Element.1, even though
'13
_obviously.~it-would-be nice if i't"were'available today,.will
- l'-
- be - available in tihe July-AugustL time f rame, and Elem~ent 1 I 15
;wiit.,1along with, if.you will, sets.of-~ descriptions-of the' l'
- codes. To'some extent, al lot of the key phenomena that are 17'
^
inIthelcodeS -- that are'in-the' codes, that aren't in;the 18 ~ cocesD---things:like that;will'be; included'in this
'cocumentation of1 Element 1. . So that that process will be a
~
20 lot _more visible,
, 21- In. fact, iflit's< worth it,.since.it's come out 22 t .now'so'many times.today,.if-we.can maybe< tomorrow at this i-23- time,fwe couldigive'an. outline of'what is goingtorbe of
^24-basic' ally.' the chapters: that are going . to be ini the Element ~
. 25' .
~1-document,'at i least-wnat.the-basis of-it'is, i' __
[. c. L- ,M
'"~
'A TAYLOE ASSO CIA TES -
162 5 1 Stre et, N. W . - Suite .100 4.'- * [ . W ashington, D.C. 20006 - l?" -(202) 293-3950 l l ~,,a & Aw.g g m.s
- 00ge e. e + w a 9 9 A-
- %M-9 W 94 BVM.4.,~Wd*,4-- ' '
C_; _' -'.
-- ._. ._ ~ --
3: 140 1-So, you-know.---in fact, people-are working'on h.m g 2- this rightinow. .
~
3 MR. ' BE.RNERO: I think it's worth adding, tco, a code that leaps to-mind, the MARCH code, was first 5
' documented, if I recall, in-1980<through a publication from
-6
~
4RC NU' REG. And'there is.a very large literature on the ; 7- MARCH code,.the review of.tne MARCH cod.e. 8 And I just checked.with Battelle-Columbus, we do 9
.have-a draft version,, a MARCH 2 users manual. But'it's an N ' enormous tome. It's a.very big thing. Ano we wouldn't.
'~114 attempc to flood you'with that. That's available for
- 12 pursuit if anyone wants it.
The literature is abundant, 13 task forces' reports, ACRS meetings in'this room and tnings 14 ~ like that. G" 15- But really, I second Mel's suggestion that the 16 = review of wnat that Element 1 contains, the scope and
~
17:
~ content of it anc on which it was-such a fundamental l'd -
comment from this group,~,we need your comment _on that. 19 ' because it's a necessary part of this. 23 MR.-VOG8L: One'of<the problems is that MARCH is 21 a moving target.- Je now have' MARCH--2, 22; MR. SILBERBERG: That's rignt.
'23 MR. BERNERO: --.Theres a national resolution that 24 we're going to get rid'of1the numbers after the decimal
- 25. place. :-It's'goi'ng to.'be MARCH 1, MARCH.2, and~I'll' shoot t
[- j.8 .- [ ~sp.' 7 T A Y L- 0 E . ASSO CI A T ES . j. 162 5. I Street, N. W . - Suite 100 4 W ashington, D.C. 20006 - ! -- (202) 293-3950. s f-l 2, u. . a p .-a_ .i - . _;._ A - , =. m
- , w- _a _ . _
147
..'the next' guy that moves.
~
1
~<
~
2- MR. VOGEL: What about' April? E 3 (i.aughter) MR. SILBERBERG: In fact, I think that what we 5- ought to'do, ana I think I am going to assign this to the 6- ' Battelle staff -- I think it's very simple --~is where 7~
~
-documentation clearly. exists out in the open literature on 3- the codes is the ones that-Bob Bernero referred to, we'll
-9 .make a listing of tnat available. Where the documentation
- 1') - is not available, we'll try to get these-sets together for 11 you before tne next. meeting.
12- 'Tnen that, coupled with the Element 1 cocument, 13 is about.the best' package one could have. l' Again, let's be clear that there is not a 0 - 15 complete standardized users manual, as we understand users 16 manuals, for each ot these codes. There just isn't. I am 17 not sure'there are users manuals for some of the other la codes that are being used in some of tne companion studies 19 out in the inaustry yet either. So that's a little bit of a 20 problem. 21 But please-bear with.us, and we, appreciate your 22 Concern.
. 23 I Know this wasn't part of your presentation,
. 24 but now you can go ahead.
- 25 MR. LEE: This morning I think:we talked acout i s
j- , !. j TAYLOE ASSO CI A TES
" 162 5 I Street, N.W . - Sutte 1004-
)
W ashington, D.C. 20006 '
-(202) 293-3950 i
, we . ,- .+ a.~em.-e...... co.,,x . . , . . , ..,+w.r~.s < ~ -
}
146
;, ~
'l
' flow paths of fission products for both.AE and TC. As you u? 2~
; know, the flow paths depend a great deal on'the actual 3
-event,Emostly on containment failure time and bottom head 4'
failure time.. So containment' is not as straightforward as
~5-was the case for PWRs.
6
-(Slide)
'7
'Tais.is wnat I call calculation procedure for a
modeling tnese transfer fission product in containment for 9
;AE sequence. The NAUA code takes the input from'the 10 TRAPMELT, and then'the fission product.goes through the 11 suppression pool. For that-we have the SPARC code, which 12 was developed by~the Battelle-Northwest Laboratory. And-13 then after that we have a wetwell. For that w'e.use the NAUA
-14 code'again.
15 . And~then as the containment fails, the flow path 16
.is'such that fission product-goes into-the crywell 17
'cirectly,_ bypassing the suppression pool. .And as you nad .
la the bottom head failure, you have additional source from
-19 the core-concrete interaction code, which. again was 23 ceveloped'by Sandia Laboratories.
21 (Slide) 22
-Ihis is Peach. Bottom AE. gamma'drywell cas~e. And 23 wnat we.have here is the total mass suspended as a function
-24 .of-time. 'And as is usual , you'take:a' source as your
- 25 icucemelting start .. .
That starts around 12' minutes. And then
; T A Y L D Ez ASSO CIA TES 16 2 5 I S tre e t, N . W . - S uit e 100 4 W ashington, D.C. 20006 (202) 293-3950
+v i- = &* s ,,m.-t- F - + * * "-M ""'*b # #
149 1-
, 'at 3S-. minutes containment faiIs.
.s '2 So the mat'erial leaves the drywell out in the-3 bottom until the bottom head fails, and then you have 4
'another source' from the core-concrete interaction that
- 5 starts remaining in>the reactor coolant system.
6-That:gets.into the drywell and then remains 7- suspended'until everything_ leaks out to the environment. 8 (Slide)- 9- I think one gentleman had a question this-10 morning about enis graph. This is a little bit different
- 11 than.what you had for the same reason explained this 12 morning. But this is the particle size _ distribution as a 13 function of time, what I have is radius in microns and l'
again gram per cc. And then dependent on the actual time, Cc. }s 15 you'cou'ld have a large particle about 1 micron radius early 16' oo n . And then depending on where you are in the reactor 17 Coolant system, spena a/ lot of time explaining.the Change la ef particle size shifted to smaller particle size. 19 This particle size doesn't' nave to necessarily 20' agree with what Mike nas, because this_is average of what l
' 21 you have in the containment. But the trend is such that it 22 shifted to a smaller particle size.
23 Later on,'when'the core-concrete interactions l 24 start, you take' additional source, and-the particle size 25 shifts'back, depending on the size of_the source you take
. i
- f. T A' Y L O E A'SS O C I A T E S 162 5 I Stre et, N.W . - Suite 1004 w ashington, D.C. 20006 (202) 293-3950 7; - -. : . . . .
150
~
1 at that time..
. / 7 .
7 2 MR. COOPER: Ken, perhaps.we could nave this kind 3 of' size distribution:shown in a cumulative form. It might be'more informative inisome ways. At this point, that's the 5 differential size distribution. The axis is probably
- 6. - slightly wrong because we don't know tha't's the 7
concentration. he~really want a concentration per size 8 interval. It may be a little easier to work with' cumulative L9 distribu'tions, I-thinK, to interpret this kind of draft 10 than-fractions. 11 . gg'. LEE: You mean straightline? i
.12.
MRb COOPER: Yes. 1 13 MR LEE: We could do that. I don't.know how that 14 can:: help ~you~ better. 15 MR.iCOOPER: I can't tell from this what the t 26 median is, for{ example,what the standardd'eviation is.'And
. . t 17 it's clear that the vertical: axis.is not quite right in i
is terms'of labelf,ng, because it's got to be per-something.
~
19 - MR. LEE:. You're talking about per radius? 20 MR. COOPER: Right. 21s MR. LEE: Well, I will'tell you the area of the 22
. code.is.notLin irect proportion to the information tnat 23 .you neea. Okay.- Size spread, ~ you can pretty much tell by 24 -the spread ~of the curve. But that's a good point.
25- (Slide)
<^2 *
~(~j TAYLDE ASSO CIA TES 1625 I Street, N.W. - Suite 1004
- W ashington, D.C.' 20006 (202) 293-3950
_ . - _ _ _ . _ Am --,-.-.-,.--G---- -- -- -
151 7-1 Talking about. matching mean particle size, k; \~ 2 that's exactly here. Tow,Ethis'is average particle radius, 3-but what I have'is'really geometric mean particle size as a 4 function'of time. And again, we see the same trend here,
~
5 littleLbit large particle size early on, and then it drops 6 down, and:then again shifts to the larger particle size 7 range. 9 MR. COOPER: And this is'now on a number basis 9 rather'tnan on a mass basis, number median diameter? 10 - MR. LEE: Exactly..Right. 11 MR . : GINSBERG:. Ginsberg,;Brookhaven. 12-On the slide where you show the total airborne 13-concentration or tne total mass versus time,.a couple of 14 slides back, tnere ih about aE6 order of magnitude in the
.@ - 15 airborne mass in an extremely short period of time. Is that 16 on the basis of some sort.of-'an equilibrium kind of" 17 Calculation, or is that on'the basis of a' rate. process?
Id MR. LEE: Which one are you talxing about? I'- MR. GINSdERG: that one. 23 21 22. l' 23 l 24
-25 i
i j . j~ TAYLDE ASSO CIA TES
~
l 162 5.1 ' Street, N. W . - Suite 1034 w ashington, D.C. 20006 (202) 293-3950
.WD.
/t
- - - . - -. . . , - ..... - - .- . . - . . .. . , . ~ . . . ., . ~. .. ~-
MM13 152 rgl' l' the core concrete interaction.
.m x, 2- MR. LEE: Ckay. Containment fails at this point, 3 so you have a tremendous flow passing through the'drywell.
4 MR. GINSBERG: Go further in time until the 5 ves'sel fails, down at the bottom.. The~ vessel fails down 6 at that minimum, right? 7 MR. LEE:- Right here. That's right. 8 MR. GINSBERG: ' The vessel fails; from there on 9 you have1about a :six order of magnitude increase in the 10 airborne mass.
-11 MR. LEE: Right.
12 1 01. GINSBERG: In a very short period of time. 13 The question'is - 14 1U1. LEE: You' ve . go't alot of material there. h 15 ' You're talking about 4000 kg. - of : aerosol particle out of I i l 16 ; the core concrete. At this point.I think core concrete o l 17 interaction is ended. i
~18 MR. GINSBERG: Is that total release based upon I
g . .
- p 19 . some - sort of rate process or is an equilibrimm calculation i
- f. 20 ~ of'some kind?
]f3 21' MR. LEE: - The rate is calculated by this method 22 :I mentioned. The rate is in there. It lasts approximately 23 five. hours in here.
24 MR. SILBERBERG: It's a rate process driven by 25 the CORCON calculation for the core concrete interaction. e a- mmw-- ..asei. y e6e ne. =4 he D --y- 59 W a W - = r- - P**#- -
5MM13c 153 crg2 1 ' MR. VOGEL: What;was.the composition of the con-2 crete? 3- MR. - LEE : It.was limestone, wasn't it? 4 MR. CZYBULSKIS: -Limestone, high limestone.
- 5 MR. VOGEL: May I protest that? We got a sur-6 ' prise when Te L looked at this Surry concrete composition.
7 When VEPCO' looked at it'for us,;it'was not high limestone, 8 it was'very low amount of carbonate. -It was a salacious 19 - material,-and we think, considering whe~re Peach Bottom is,
~
10 : thatlthis may well be true for Peach Bottom. So I think 11- specifically thel composition of _ the concrete should be
- 12 checked, since these are calculations specific to a particular
. 13' reactor.
14 MR. SILBERBERG: I assume that that is something h - 15 ' that ought to be easy to get, right?' { 16 MR. VOGEL: -It,wasn't-easy to get. We had to 0 - l 17. -go to the utility. They-had a'little struggle. ! 1 18 MR. GIESEKE: As far as the Surry calculations are
-g
!' 19 concerned, or were concerned, I think there was--it was mis-A l - 2 . leading because I flipped a coin:at one point and put-lime-
-21 ! stone in the text which I pointed out in one of the meetings g.
]. H- was.to beltaken out and exchanged with basaltic based on
'n ~ what'was actually.used in the calculations. We never used j
. 24 the limestone concrete in the Surry calculations. I just C.
- 26 ' stuck'it into~the' text because.I didn't know what it was at
't s
_e Oq=- p9 (W y* 1r-@%*WP' WN #e
- 4 WF-4 O --4" P 'q%%d' 4'M *-N'a "
. ,-a
< ?MM13 154 rg3
'I the ._ time I- was writing that section.
h 4;j 2 MR. VOGEL: That's very creative. 3 (Laughter.) 4 .- MR. GIESEKE: And I forgot to take it out until
~
5 a little' bit later after a' version had gone out. It was
~
6 . removed from a later issue of the report. 7- MR. VOGEL: Okay.- That's good. Then we didn't 8 :havelthe sparging of the CO2 , 9 MR. GIESEKE: You did not at the Surry. The cal-10 culations were done-with the' basaltic concrete. 11 MR. VOGEL: What have we got here? 12 MR. GIESEKE: In this case, all the information 13 ~ we've been given is that it's limestone-rich concrete. 14; MR. VOGEL: Have you seen the real analysis? h 15 It's' worthwhile checking. ~ I l- 16 MR. GIESEKE: We got that information from two f 17 sources, one being ourselve's, and one being Dave Powers, il l 18 MR. VOGEL: I think you yourself should look at 5: 19 the analysis, because.-we were surprised. [
't:
20 MR.'SILBERBERG: ' Dick, I have'a suggestion, if l-21 { you don't mind, since we are now getting contributions from 5 22 various sectors of.the industry. Would the EPRI-wish'to. f. 23 take.it upon themselves'to provide clarifying information 24 for the five plants that we are studying relative to concrete 25 as a' point of contact?- It might save us time. L --~ _ g
J n
. E413. -
155 lrgy - m 1 MR. VOGEL: That's an: interesting suggestion.
/^s -4 ,
2- (Laughter.) 3 MR. GIESEKE: One is already done. 4 MR. VOGEL: Let's take that under advisement. 5 By~no:means would we be' inclined to take your suggestion 6- light 1y. 1We'11 see if we can do it. That's easier -- to 7- - ask Jomebody to do something when you don't know how to do 8 it-laurself. 9 .(Laughter.) 10 MR. BERNERO: IDCOR is using all the same reference
- 11 plans, except for Surry, so IDCOR may very well pinned this 12 - down.
13 MR. VOGEL: If they have , .I don' t know it, but we 14 can check. 4 15 MR. LEE: I don't want to spend too much time on . 4, 2
-l 16 this suppression pool.modelling, which was used in this cal-0 -
" 17-
' culation.
~ 3*
18 _(Slide.) r a , i 19 I think sometime this' afternoon, er even tomorrow, 5 f 20 we are going to have a session devoted to talking about the
-{ - 21 suppression pool,.but'nonetheless_the mechanisms included
.5
~
' 22 in the mode 1, I'11-list it here -- condensation of steam,in-
?
23 brtial deposition diffusion, gravitational settling, and -- 24 this is kind of unique -- particle growth due to pressure
/
D 25 decrease as the bubble goes up. And that will create the L m
..-.---.-..-.--...;n. - - . .-
% v ~ v v g 4 p v -
---.-n- . .
, Dag13l ,
~
^M '
.rg5 - l'I'- - _
n ~ '
-156-
' 'a.
, -i t
!? 'negativeldeposition.jAnd[then,'in: addition, there's;an in-ap '
jf L. y7 je'ction? impingement mechanism describing the impaction:right
<2 y
s . Jafter.:thetpipe. exit core exit,'and.then.the solubility of i
~
, ;~ 4'- particles-is Jconsidered;in the models. ;
*~
L5h (Slide .-) - 4 6 LWhat[youshave is a-calculated decontamin'ation }'- 'n _ ,. . ..
. factor 'asi a L function-Lof: particle'. size f for dif ferent -times l7 (8, ifor the-;AE sequence.-!Ag'ain the pool depths thet are used for- I 9$ this/particular. sequence;isLfourLfeet, and' bubble diameter,.
~which ~is i found 'to _ be . very sensitive -to the" final result, ~we
~
- 10' 4
k e .
<11 used .75 cm.-and our aspect ratio.was_ assumed to be 1.3.-
' 12 (The point'I'_mLtrying to make is that the decontami- . -
I
- 13' nation factor is'a-rather strong function of particle size. .
, 1 14 "In fact, the factor _ larger than:10 -- we'just assumed thati
~
! -[2- l 15 - [it'sL105,. So depending;on what kind of particle size you . 4 4 f
~
[~ L 16 -have1when you go through :the suppression pool, and. the decon-r 1 o. tamination' factor based -onitheitotal- mass. really detiermines
'17.
- -l_
1. l 18 the'. ' size dif fusion of ? the aerosol: you' have .at 'that point. 1 k
'5 19 MR. CASTLEMAN: 1Could: you sayl just a little about .
{ xj.
'20 1 the physics'that went;into.that..just so I_can. appreciate ~ j
] ..21- 'whatsthe calculations. ara?- l
- 5- -
E
]- -
, - 22 '
' MR '. LEE:-.I think1I would rather leavecit'to -- ;
} 123 MR.ISILBERBERG:- Thatlwill be; discussed,'actually,. 24 (after.the next' presentation.. 1s ; (@ . 26 ' .MR.= LEE:. But this particle. size.-dependence"is
$ g.
+
g L - . $f
- 5. N4 .U'
,t
,,,.k. . 4 J_i.
4
+
d
> -4 . ,,
~ , .- - - , .
v4 . a ,~ . a , ,
.,m,-' .#
o-
g . MM13~ 157 rg6 1 clearly due to the' inertial impaction mechanism. .Again, N 2 the'same reason we talked about this morning, the Stokes 3- number plays an important role here. And then, depending 4~
.on the condition, steam condensation plays an important
-5 role here. That's another reason why this bubble size is 6
important.. And then pool' depths -- the deeper pool you have,- 7 I think you expect.that decontamination factor will increase. 8 But I would point'out that this factor'is defined
~
9 to be'a concentration of going in divided by a concentration 10 of' leaving'the pool, is that right? I'm always confused with 11 this. So, this is like we're talking about three percent re-12 lease and this is like more than:99%'-- like 10 we're talking p, 13 about 99.999% retention of collection.
%d/ -
14 MR. KASTENBERG: Before you leave that, I have a 15 que'stion. This sequence is AE gamma or AE gamma prime? 4 l 16 MR. LEE: AE gamma. prime. f 17
.MR. KASTENBERG: So those other' figures were i
18 - mislabeled when you sa.id AE. gamma? s 18,
! MR. LEE- .We just have one AE sequence: that's all E
'20 AE'. gamma. prime.
*f 21- MR. KASTENBERG: After vessel failure, you don't 5
- j. 22 . ..get-scrubbing of the pool, right?
I 23 -MR.-LEE: .That's'right.- 24 MR. KASTENBERG: So--
,, j 25 MR. LEE: This is before-the containment fails.
& , g,g rg7 158
~b 1 TheLeontainment fails at 35 minutes.
~
._/
2' MR. KASTENBERG: If you took an overall decontamina-3- tio'n factor-- 4 MR. LEE: You cannot read it from here because 5-. this is the factor -- the function'of' particle size -- but 6 DF based on~the mass, you have to multiply this with the
~
size-distribution you have. Now, overall decontamination
-8 . factor you-have to integrate over the mass as a function of g _. . time.
.10 l MR. KASTENBERG: Do you have.any idea what that-11 would be?
12 MR.-LEE: Overall, I think wt have about 30 DF -- a-DF of about 30. So we're talking about -3% of the material
} :13 14 going into.the pool escape the pool ~. Now we will'have a h 15 little. bit ofilow DF for'TC sequence.
t MR. KASTENBERG: But wait. After the vessel fails, h 16
~
S' '17 you don't get any scrubbing, right? The relense goes to 2 18 ; .the drywell? h g 1g MR.' LEE: Right. I" . . 20 MR. KASTENBERG: ;-And then out? 5 21 MR. LEE: Right..I'm just talking about this for l22 .' time..
.l-
.m MR. KASTENBERG: I-understand..But past this time
-24 would this be significant?
) ,+ .
31 MR.-LEE: I'm not talking about -- it's not going u- _ _ . , _ , - _ _ . -
, , g.
LMM13 - i l159 , rg.8: j '
- 1 to-be significant for this sequence, because we receive-
,x 2
.alot of material at'the. containment failures.
3 MR. DENNING: I'd like to clarify something. 4- You kept saying "after. vessel failure." It's actually after 5
' containment failure which occurs before vessel failure that 6
you no. longer have scrubbing. It's pressure pull.
'7 MR. KASTENBERG: As impressive as these numbers 8
. are , '. when you look at the'overall risk,'it's not going to 8" ~
have an impact. 10 s MR. SILBERBERG: For AE. 11 MR. KASTENBERG: .For AE. 12 MR. LEE: You're right. 13 ' - MR. CASTLEMAN:. Despite the fact that 14 you don't want to go into the physics now, could I just ask 15 ' one question? You have'a' bubble diameter there of three-3 g 16 qu'arters of a cm.. Supposing that were half the size, or 17 twice as' big, how much might that affeet these~ numbers?
-18~
I don't have a feeling whether the three-quarters of a cm. -- 18
!. MR. LEE: We probably have a' sensitivity analysis 5
8 result. D'o you have anything like that?- f-
=
21 Debra -- do you have any? 3 22 lf' MS. HAWKINS: I've got some results of the bubble 23 size. 24 g MR.- LEE: It's rather dramatic. Also, the ratio -- ! 'I - 25' l MR. COOPER: Just to-suggest that the model that s _ , . . ,_g- . , , -,.m'. -.%, .L4. s
MM13: -
- l: 160
-rg91
-11 you..have mightibe-compared against the models that are
. \y],}l 2' . traditional, the scrubber 1 literature, and also against 3' .some-scrubber results; I'm a littlefsurprised at-the high
~
14'
~
collection efficiency-for one-micron particles and half-5 micron particles in this~ case. 6- MR.~SILBERBERG': Higher for: one' than' a half ? 7 MR. COOPERi What surprises me is that they're 8' sollarge Ta't all. 8: -MR. LEE: I have had alot of problems because 10 -my DFs are-so low.
- 11' (Laughter.)
12 - MR. COOPER: 'You have solved them-in that case. 13 ' MR.EVOGEL: Perhaps.Dr. Cooper isn't aware that 14 EPRI has an on-going program on scrubbing of. aerosols in 15 Battelle,. Columbus,.and we are in the' process right'now of
.] 16 scrubbing cesium ioside aerosols. And hopefully these will 0
l 17 be available for use.- 1 18 MR. LEE: .This is just an example of calculated
. .a 18
[ total. solid particulate material at'various;. locations -- E
-M .(Slide.)
21~
-f~ --at different times. And I think to understand 3-f' 22 .better, fI .thinkE we j us't listed some of the important . events 1D' in here.. 'Again,_ melt starting:_ time is 12 minutes. Then you
~
24 start getting this material. And right before containment, C: - 2 fails, you've'gotJso much suspended . in drywell, and about
. - ...:.a . - ..-. ;
a..- .- . ,a ... - . .e . ,. . . . . . .
OD113 ~ 'e 161 org10 - -a-1 the same amount of material deposited in drywell again that
<~r.
2 alot of'the material has gone to the suppression pool and 3 captured, and very little escapes ~into -- suppression pool 4 .toireach the.wetwell. And, as the containment fails, the s ' flow path changes. Normal particles go through two suppressior, 6 pools,-so the amount of material captured in the pool remains 7 - constant. And then, I guess head failure time is here. You
.a have a really low source coming out of the primary system for g- the same reason Mike Kuhlman talked about this morning.
to And then, as your head fails, you start getting 11 this material from the core concrete interaction as well as 12 from the reactor coolant system. pg 13' And.right here,:I.think that's a rather long time, A./ 14 at the end"of accident I have about-20 hours in there. The
! is . material ends up being located either in the drywell or I .
g' - 16 . in the pool. The wetwell -- I guess you really don't have
'O
; 17 anything lef t in the .wetwel'1 -- and then escape to the environ--
.1 18 ment.
.g.
-f; 19 And as you pointed out correctly, the final re-20 -lease amount does not really' depend on the efficiency of the
- 3, End tape . 21 - ' suppression pool'in this -- for this' sequence.
3-3 22 2 n 24
-. .g .
4Cd 1m
*j 9 1' di %#.SM N9' d. d i .ae M.nqMW4DS .&4 4 4g4 g e-h% +%g- % M'g
Oc, y -- - W i' % > 162 ff l ' (Slide). A n, 2 _ Let'me go through.oneLmore thing. Okay. This is 3
-for different species here. / Again, 16 percentof the cesium
~"
.iocide at-the1endtor an accident remains.in-the' reactor
'5' coolant system, andia significant amnount of material is capturec'in tne pool,-and 14fpercent end-up in the drywell,-
7 mostly.'by'gravitationallsettling, and tnen about 21 percent a goes:out to the environment. 9
- MR. REYNOLDS: What pool is that?
.Md.~
l LEE:. Suppression pool. 11- MR.-REYNOLDS:. 'Okay.
-12' ~
MR. . LEE : ' And we'_ve got a pretty much similar
~
13 -: number for cesium hydroxide.- In case of tellurium because
~
o l' cof ' source tiining, we 'have Ja- dif ferent number. A lot of the 1 15. -tellurium gets r'eleased-in the drywell at the time of 16
; core-concrete interaction, and~also.as.Mixe Kuhlman again 17 mentionec in the morning, even_after the tellurium system, 18 . tellurium gets released later,' so tellurium doesn't have a 19 chance to go through the suppression pool.
20 So as a result, you-have a rather'nigh release 21
, fraction. And if you direct your attention to the 22 blacxboard there, in the case ofJeesium iodide, I guess it 23 1 started with.one here,-99: percent goes through the reactor-24
.coolanthsystem. .Out of 99. percent, 18. percent stays there,-
~
-25 ;which is the same.numberLyou have in the viewgraph.
= p ;)'
T A Y L O E ' A SS O CI A T E S-
- 162 5 I . Street, N. W . - Suite 1004
, W ashington, D.C. 20006
?(202) 293-3950
.S0
- u. . y .._ . ~ .;_ a - -.4,- .a. ;, -, - - --
l~ I
~
163
'l Out of.that,-I guess the remainder goes into the . o.
[
, 2.
drywell,-and the-47 percent has a chance to go through the 3 suppression' pool. 40 out'of 81 goes to the suppression 4 pool, and the remainder goes directly to the drywell, and 5 thenit: pretty' much~goes to zero. Then you come out with
.21.
7 So that is the flow path or the way of the
'8- cesium iodide. And you get a similar figure in the case of 9
cesium hydroxide and tellurium. 10-
}lR . VOGEL: (Let me maxe sure I understand what I
-11 you've done on the tellurium. Are you assuming CO2 sparges 12 tnrougn this melt from the, limestone?
i 13 MR. LEE: I$cannot' answer.thatquestion. 1" MR. GIESEKE: Yes. s-,
'15 HR. VOGEL: Okay. That gives you a higher release i
16 than ifLit doesn't sparge through from the limestone. 17
\'
MR. SILBERBERG: 'Yes.-
'18 k MR. LEE: The~ core-concrete interaction tellurium i
- 19. gets released at-a lower $cate because of the cesium
- 20. nydroxide, cesium and iodine. Another reason is that in the 21 case of-tellurium, .it goes through this path or through 22 this path, so you really don't nave any --
23-MR. GIESEKE: 'et L me make a point. From what.I 24 ~ understand:from Dana - -I'will try to speak for him, I - 25- guess, since he is not here. I will back up a little bit. e, ( .. T A Y L O E - A S S .0 C I A T E S p 1625 I Street, N w. - Suite 1004
' W ashington, D.C. 20006
-(202) 293-3950
. i.h
._ ._v.. . -. -..-.-- - ; -
. . . , . , . +
.b- ,r Q:q y .
' L.: - - - -
7g4 x
~
1 I have been to the pres'sure< vessel. There'are s p .N ---
' 22
' J?- mechanisms,Las-evidenced by.what~we' learned from Oak Ridge, 3I that'will help keep the tellurium tied up.with the. melt.
faut apparently, once the' tellurium with the rest of the a_ _ 5' malt { material is=on the. concrete, the-sparging process and
~ also the chemistry is such that according to Dana, there is 7-sno mechanism tnat..he-can see-or no chemistry that would.
s 8 Ihold:tne. tellurium in that melt and the sparging process 9 carriesTit-on through.j slo. I .am- just quotingiwhat he told me. 11 MR. VOGEL:t The last time.I talked with Dana,
~
;12.
- that' conclusion onrthe-relea'se of tellurium was based on 13: [the esaumption~that it behaved like-sulfur. And I recall '
l
, -:14I 'that there is even-selenium between-sulfur and tellurium. -
.15 '-
So-ILthink-that's,1A, a shakey hypothesis.unless it's IU confirmed by experiment. 17-
, L Anc1 secondly, I thinkJ that when you've got a 18 rather large mass of.. molten' material, that they' effective
~
i 19 Scale on thelrelease of these-volatile materials'is very }
- 23 important',';and'also whether you dofor~do:not have sparging.
21;
~
MR. SILBERBERG:, Even if I.had a basaltic 422'- concrete,- I'still-have'sparging,.although less.than 23-
- limestone sparging, due to-just steam from.there. .
Right? , 1 24
' MR . ' .VOG E L : . .Yes.
25 , MR. - SIC.8ERBERG : -But'it would be less.
-C.
!. p. -:
T A Y t. 0 E . A S'5 0 C I A T E S : M' '
~
162511--Street, N.W . : Suite 1004; *
;W ashington. 0.C. 20306 s'
?(202) 293-3950
.n . ,
' " . d
-4 M N
. =
g4"ed W Y** 9 11' 8 - .
- e. ? , " ,
% 4d' 47 N NP-^^* -^ '"I' E * ~' ' '
^
E ;- - - - . - L 165 1 MR. VOGEL: Yes. Am I correct that there is no
;f%
(;) 2 experimental. evidence on the tellurium release from the 3
' melt? Does-anybocy want to contradict me?
.MR. SILBERBERG: I think that's correct.
5- MR. LEE: he are going to talk about TC gamma 6- : prime-this. time. Thetflow paths of the fission product is 7: such-that they get into-the suppression pool first -- do 8' youthave any Figure 7.4, page~ number 7? 9 MR. KASTENBERG:. -Yes. 724. 10 MR. LEE: . Page number-724. Again, that's the way
. 11 we. have' been using computer code',tne kinds of flow. path of 12 tne' fission product.
131 So the first. thing.we want to do is to use the 1"
' SPARC. code to calculate the decontamination factor and then
&* - 15. . modelcne wetwell. ^And as.the vessel fails, tnen I.think
'16 the'floi path'is directed to-the drywell, again bypassing 17- the suppression pool.
18 ( S l'ide )
- 19 This is the calculated decontamination factor 23 for TC. The same form that you saw before.-The pool depth 21 is~ a-little bit larger than we.used before.
But basically it's'the same informati'on. , 23; -MR..KASTENBERG: How do you use.hAUA in:a wetwell 24 when'there's wa'ter in therd, when the pool is there? 25, MR. LEE: We just use the space above the pool.
..T A Y L O E ASSO CIA TES 1625 I Street, N.W. I,S uit e 100 4
.W ashington, D.C'. -20006
, , (202) 293 3950
# ,m e rg =m W . w* M
_m. T' ? "*"***<'*iY'W *# ---
- d LM - a&
166 1 2,
.bGt. KASTENBERG: But can-it handle the fact that 2 you have a water surface'that's hot?
, j 3: MR.' LEE: No, you can't handle that. We're 4-7 talking about another. aerosol deposition mechanism.
5
. MR . KASTENBERG: .Yes. Just from my understanding 6
of ho'W NAUA works, I wouldn't think you would model that
'7'
. interface correctly.
8
- MR.' LEE: Specifically, can you think of any ,
9
' specific lrather than -- wnat I did was everything settling 10 - on the floor, which is water, I-just shoved it into the
' ll-suppression again. That's one: thing.I-did.
.12
-If you have any suggestion or any specific 13~
mechanism ~you can think'of,- I think there might be some l'
.- , important thing we missed.
O U,) c/ .15 MR. COOPER: Two things we can.think of that 16- might,be important. One is that if~the liquid is 17
-evaporating, it will again-produce a virtual-wind that is
~could oppose any settling tendency.
19 And secondly, that the breaking bubbles will 20 after a while begin to put up some of the materials
- 21.
suspended in the liquid back up into the air. Whether those r 22 are significant I am.not sure. 23 MR. LEE: OXay. Obviously, we have not considered
. 24- .tho'se n poss ible.' mechanisms , but we did use the NAUA code for 25 modeling the-wetwell.
r ..
. f(. p--
TAYLDE ASSO CIA TES 1625 1 Street, N. W . - Suite .1004
. W ashington, 0.C. 20006
,, , (202)'293-3950
=.
,9' *
***f' - .
. . - ,, .. - ~
167 1
'(Slide) q q7; 2 Airborne concentration of the particles as a 3
-function of time.in the wetwell.
(Slice) 5 Particle size distribution as a function of time, the.same format you saw for the case of AE,~but again 7
-we're' talking about-pretty small particles.
8 - (Slid')- e
~
' ~
I want to pointfout at this point that
'10 Ldecontamination factor we have for this case is lower.than 11 ~-
-what'we'have.for AE. Again, the overall decontamination 12 factor-based on mass, we're-tal.<ing .about decontamination 13 ' factor'of 7, which is rather low, which means about 13 14 - percent of'tne material escapes the suppression pool.
15-And one of the reasons we have such a low 16 decontamination fac' tor is that the particle size dispersion 17 of.tne incoming aerosol is very small. And at that point you are tak'ing a lot of mass at a rather high-rate because
~
is
~19 of the-surge in the flow'we talxed about this morning.
-20 so at some' point you have a decontamination 21 factor,'largerJthan 100. But that is not really a deciding 22' -factor in-calculating the;overall deconatmination factor 23 during.the entire period of when the. suppression pool 24 operates.
25 - . MR . VOGEL: dow do you model the transport of'tne 7.~. * {J TAYLDE ASSO CIA TES 16 2 5 I S tr e e t, N . W . - S uit e 10 0 4 -. w ashington, D.C. 20006 (202) 293-3950
..,2_..,_._.,._: .,a _. _ .--
o-168 s
- b aerosol through the pipe'to_the suppression pool?
b i-
\; , 2
.but. LEE: What I did1is I think before I.said we
'3
~
nave an impingement zone right.after the exit of this gas.
- We looked at that, and what' I' did was to calculate the 5- velocity,-gas velocity, out of that exit. I added all this
~' opening' area,.what,115, 44, in tee quenchers, and the 7'
diameter of the hole was'.4 inch diameter. 8 So;I'added.all these holes to calculate the
-injection velocity. and-the resulting Stokes numoer was not 10 large'enough to change the calculated factors. Again, 11 particle' size-was just too small.
12
- y. con't Know whether I answered your question.
13
~
MR. COOPER: I think the question was how you 14
~
g .nahdle the possibility ot deposition in the pipes leacing (bdk 15 to the pool. And probably the' answer is the size 16 distribution is-so fine and the residence times so short 17- Tthat the major toechanism might be diffusion, but stuff 18 that's a few tenths' microns is going to get through there 19 very readily. g 20 MR. VOGEL: Is that true, that the flowrate is 21 ' tpat high? 22 MR. LES: Right.
.23~
MR. .KOLdMAN: 1During the period of: time that Ken 24 is? talking about, the residence time on the primary system 25 is well under 10 seconds for these gases to come all tne 1-TAYLOE ASSO CI A TES 162 5 I Street, N. W . --Suite 100 4 W ashington, D.C. 20006 - (202) 293-3950
t-%l.*._
'M %
~169 I- .-way?from the core through.the-relief lines. We do model
'2 LdiffusiveLdefinition in all'the pipelines on the-way, but 3
even'so,. there's.not much time for that.to take place
- ' '- eitner.
'5- :Really,-.in effect, it'.s going.to be an a-major 6
mechanism in-the bubbles and any of the pipelines along-the 7- way. ~ 8 MR. LEE:!-'Okay. Now we are in the'drywell. That's 9
-airborne' mass as a function.of. time again, and it depends la . - on.the source timing. --You have a suspended mass'as a ,
'll function of time. Again, these peaks represent source out
' 12 ; of'the core-concrete interaction, and this is the cumulated 13 mass leaked out to' the environment.
'14:
-(Slide) s ,
15-The-same plot except that now we have different 16 -species-in there. 'I think one thing ~I want to point out is 17-Lagain in this~ tellurium release, you have a rather high 18 ~
- release' at- the time of core--concrete interaction. So as in 19 the case'of AE,-again you expect that the final. tellurium 20'
. release traction is not going to be as low as cesium or 21' iodine.
22 ~
-(Slide):
23 Mean particle size as a function of time. Again, 2"
'rather small particles early on, and then maybe larger 25, . particles ~ later ..
fp;
'? - . TAYLOE~ A S S O C I A' T E S -
11625 !! Street, N.W. . Suite 1004
,W ashington 0.C. 20006 (202) 293-3950
. . - . . - +.~,....a_ z.=_-.-.. .
. - =
[g _. 170 1 21 (Slide) ; (~ . W g 2 Oxay. -This'is it. Bottom line. This is the
'3 number you-have.in'the morning. The stuff rettined in the 4
reactor coolant system, and.then 54 percent remains in the L5 pool,.'a'little. bit.-in the drywell, virtually nothing in the
'I . wetwell,-and 34' percent out to the environment.
(
+
;This tellurium is'still 32.
If you hao this up, 8- it's..not going to be 1, which means not all the tellurium 9
..was' released, .especially the stuff which got out of the 10 core-concrete. interaction, c 11-I~talxed with Dana Power. -We didn't get all the
- 12 tellurium which was available. So that's the reason why it
~
= 13 won't idd up.
m 14-But in the case of AE, everything got released.
-." (-- !
15-Some 70 percent, some 60 percent got released. But t'his is 16 the final result anyway. 17~ Now we are going.to talA about TC gamma, which 18 means we will give credit to the reactor building, the 19' . secondary containment. 23 So what 'he did was to take this and run this 21 NAUA code once again to model this reactor-building.
- 22 23
- 24 25-TAYLDE ASSO CI A TES 1625 I Street, N.W. - Suite 1004 W ashington, D.C. 20006 (202) 293-3950
^j e ,4
,, _. .s- _,-,z s, .~...G.. . .m.- . ~ .
~ ^ '
r - t ]f " :f :L , , - {TA 195? ' v
, , . 171
./MMijl'fl5tl..
;1 (Slide.)
f hf 4-
- 2. So,-this-is suspended mass as a function of time
- 68' and standby. gas-treatment' system is operating at this point.
'4 - And'really, this is-some of the mass released outside, plus 5= .
- thA-mass ~ going through.the standby gas treatment system.
6
'(Slide.)
'7
'At.this time, we have. reactor building here and
. :8 standby, gas treatment . system,: and then environment. So, ! 8- instead of' point 3, right now we have 20 percent. 10- Now,-why'is it"not zero? 'Obviously,' standby gas 11 treatment system can take only so much, because we're talking 12 about a-lahge flow.- And in.this calculation, I think we used 4
. 13 25,000 cubic feet per minute, going through the system. But
-- (M :
. %s -
14
' because.we assumed'the-HEPA filter in the system fails after
~
15 collecting.l04 kilogram, which we got from Steven Hodge -- 3 j 16 but, again, we're talking'about 4,000 kilogram of material 17 ouf of the core concrete interaction period, an additional 18 400 kilogram .out of the reactor coolant system during the 1
-[E- ' 18 '
metal' release or initial-release perio'd. We just cannot handle that-k'ind of a' mass.
. 21
- So, the- rest' of the flow will be directed to the 2
'l 22 : environment. You:still'have about, in this case , about --
. 23 '-
you add'them up, 13, 14 percent reduction in the final nu.nber. 24 g ' But: basically we're talking about.the same characteristics.we
- G' ' ,
25 saw for-TCLgamma prime. . se. M w,.e v Sw e+ nwMs.- .- e." w=m_sym -e -a** k e .'wm sw
,* ,. w *
--+e- "
m . . _ - .
~
. 3, ; -r y , n : - y;* ~ >
, z ' - - - - '- - - ' " - '
, :>1Mg j l7 /15 i2 : g
-p.
21 And IL thinkf the~ next fspeaker .will summarize the
;c;.
22 wholeIthing, comparing withLthe previo#us result. - :8 If you have:any questions, I will try to answer. 4 4- - 1MRi SHERRY: . Sherry , -- NUS . ~ ' '
~
5; Ken, did you look at the heating of the SGTS. filters
-6 sto 'thatiamount of cesium and ' iodide. deposited -on them?
7,
. MR . LEE:. No,-I.have~not looked at it.
t
'8 i :Here was:just' simple calculation. I think-the only 28 thing-I.did was,-like=I said,. the flow rate -- we have 25,000
-10 -cubic' feet perLminute, and then just filled the filter. At.
- 11 ' the: time, the f'ilter collected so much -- in reality, that is .
i 12' not what's going to happen. b 13 I think as you collect the material ~, I think the 14 flow will be restricted,~and then collection efficiency, while 15'- it's 99;9' percent anyhow, but it's going to be like a 3 1 j ll: 16 . tiransient period. -And then you will find it clocked. ; o - i 17 :
~
I don't know. Maybe Steve or Rich -- I don't know 18 ~ whether I answe' red your question or not.- But, no, we have not j a
'18 considered a'ny. heating effect.
.E #
yL I think this morning we talked.about missed mist 21
] . eliminator. .We have never looked at that.
h 22 -- MR. E WORMAN : Stone'& Webster.
> 23 -
; Can I find :that chart very helpful in understanding 2 24 the-cesihm' iodide' data?- Would-it be'possible to add the h, 25 ,
. tellurium to that? ---I.tried to use the. table and follow it.
r (o
}r.
+
; m.;_a_..s_ _..a _ _%,_.., .
^
173 LMMs jij ;;15:3 a.
- 1- MR. LEE: -Iti's rather. complicated. . I:didn't have E s r2 -timeito make~~a vi~ewgraph of that.. I-think I will do that for
'3 different' action, for different species.
, , _4 MR. WORMAN: Just'for_the.AE, if you could put the 5> stellurium in-parenthesis,'it would be very helpful to under-
'6-
-stand.
"7- MR. LEE:- You.could do the same thing for TC, also,
. 8~ which would.be a little bit-more involved, but I guess you 9 can keep track'of which.one goes where~.
10 L - What'is-not in'there -- timing is not in there. 11 . That'is still-not~ complete. _. But at least you can get an idea,
.12 LI suppose.
fQ .13 ' ' MR.. CASTLEMAN: 'During your presentation, you showed
%w 1s.2 BU~ . 14 -us a number of viewgraphs of aerosol radiants versus time.
h 15 It showed rather dramatic oscillations. .Yet, every time you , 3
-{- 116 showed us a distribution, -it was a. itice smooth -- what icoked -
0 . ~ l: 17 like .it would probably end up. being a logi number distribution. 1. j- 181 Doesn't it surprise you, with all the settling, agglomeration,
. g :.
19 and flow exchange from one vessel to another, that these
.i f '20 ' aerosol ~ distribution would show some. multiple beats, at j '21:
~
.least by' nodal distribution?
I think ' there are times 'where you have l ~ 22 MR. LEE:
-' 23 - =multip1.e'or. double-modal distribution; you just have so much 24 linformation-with this kind of calculation.. But we did see t
- - 26 - .just happens.to be --
1 4
. . , se,
~ n - y.. r - , - ,
m ( + 3 - .re-'+. A .. a--s + g =9 + e. --
-! MMa jl 15:4' 174 1
~ MR . CASTLEMAN: It' looked like everyone you showed P.- -
2
.k.. Jus was so~ smooth. You'.re got these wild oscillations going
.3
~
-on.
4 MR. LEE:. The mean size we calculated was not based 5-on log fiumbers. It was. computed from the disputized spectrum. 6' MR. COOPER: I understand what you're saying. We 7 ought to have back-mixing or something that would give us 8
~
two. modes -- otypically, the combination of condensation, 8 generation, and sedimentation, depletion, gives you the silt-
- 10 preserving size ~ distribution. . They often look pretty much 11 like what Ken has.-shown.
12 MR. CASTLEMAN: I've made an awful lot of measure-13 ments. Usually you see some, manifestation when you've got I4
' settling and settlement going on. It just surprised me. It 15 looked so smooth.
a [ 16 Some-people look at it.
)17 MR. LEE: Another thing, in this calculation, I
-18 think, really, the aerosol mechanisms, like agglomeration, f-i 18 l- doesn't play an important role, because the-residence time,
# - again, is rather short
, that.the flow turns out to be .an 21 important factor. Particles just-don't have any time to go 22 f .through this aerosol agglomeration, as well as deposition 23 ..
mechanism.- 7.;;-s MR.'SILBERBERG: Do you want to pick it up if it's
'hN - 25 relevant at Grand Gulf?
O pn g +-W q m.. .-se- e es+ w.wv., .4 +wg a., $w,s* --.- <- ,,. .. e =w-se c - A
..- . , ,- - s _ .
g , i e er 4e d
- l
.MMtjl115a5 175 1
. MR. ROE: A very quick question.- Is it possible to
.2- draw some conclusion here. relative to WASH 1400 in terms of 3
this being sort of b'ottom-line? 4 MR. BERNERO: He's about to, I hope. 5 MR. ROE: Oh, he's going to do it? 4 l6 :MR.-BERNERO: He does it in here. He's got bottom-7-- l'ine disease. 8 (Laughter.) 8 (Slide.) 10 - MR. GIESEKE: Just to refresh your memory, I guess. 11'
. of-the WASH-1400 release, we have then s6mmarized here --
12 we.re. going to be looking at the two categories that we feel
. 13 are pertinent, which is the PWR 2 and 3.
-14 (Slide.)
15 Comparing those with the calculations that we made. e
] 16 for the sequences that fit that sequence description, as you ,
O 17 can.see, tne' effect that we noted or Ken noted- as he went 2
. 18 through the.early release from the feels-for the cesium iodide 18
! or cesium and iodine categories, where they go through'the 5
l 20 suppression pool and get scrubbed, as compar-.d with the 21 ~ tellurium release, which come:3 along later frem the vaporiza-
.22 . tion of the core concrete-interaction, which.has more direct l_
23-
; path into the containment, boosting the tellurium above the
. 24
.... BWR.2 ca'tegory for the AE-gamma prime case. The TC gamma prime 25 caseLis-a:little bit more well-behaved in terms of the-timing.
W w a p% n-, e 4 9-v. m u.m. n . <p.,+,a.o s- ; 9
.s . .. . --. . __a-... - -
MMajl:15:6; ^ 176 Q . , ,
- 1 lit-is' still below the BWR 2.
fy a X ;' -
- 2. (Slide.)
e.
~
3' - The other case' that was discussed today is the TC 4- gamma case that.we just went through, with the auilding~ intact L .5 at least through the - filters up until the time they failed.
- 6 We were'a little-bit higher'in'this-particular case'and very a
7 . consistent, asfit was, with the TC gamma prime. case, much more 8 consistency between the species. 9 MR.' KASTENBERG: -Could you tell us, in both BWR 2 ;
- 10. - and 3, what are the major-th'ings that give you the-difference.
~
- 11 Obviously,. suppression pool is.one. The use of an hour code
'12 and so on -- which'are the most dominant things that.give you.
13 ~ the difference, [l u 14 :: MR. GIESEKE: Between.the 2 and-the 3? h; 15 ' MR. KASTENBERG:- Between WASH 1400 2 and 3.
-. 16.- MR. GIESEKE: We'll have to expla'in all the o-
- 5. 17 differences between those categories fully. I don't know the l
-18 details.
1 [. 18 MR. COOPER: I think the question was why did the 20 -calculations give different results, what are the primary
.l.
1[ s-21 : mechanisms-that are giving you more or less emissions at the 22 : end of this? I'.
' 23 MR. GIESEKE: That's what we've been talking about 24 ~for.the last two hours, I think, 25 ' Going through all the steps again --
A e , t . o 4
*
- y ,e 'a' ~ ww e y or =6 4- -ra~+.- 4
~ ~
' pf ; [~bf - . -
% i ;W fp -W ~= ~'=-~1
~
t hM a j l f l5~ s _7 p ,
- r.
177 m ' - ;n _ - _
~
- l,
;.7 _ , ,
=
_ ') ,-
'I 1+.. "~[ 1MRRKASTENBERG:t :Just in a capsule, what are the * ;.gQ.j- .b [i ' '
II: mostLdominantLfeatures?- In other words,.if you wanted to hone 53 ' 4
.g ,
some'of the calculations,.where would you focus? Which ones 4' ' are not that? effective?- ; y
' -5. iMR.,GIESEKE: What are the most sensitive issues?
q - [6 I I.MR.= SILBERBERG: Yes. m .. ;c .,.
; 7:
MR.' GIESEKE: ~~As1I' mentioned, as I put these up, I n ' - 8: ' think?the : timing -is' of. utmost importance, particularly in the
'- _ tellurium release, . where it either. does or does not go - through
~
s 10 ' the supkesision. pool. And'.also the containment, the primary iib ~ containment failure! timing 'is'important, because at that time
' 12 :
-you iceased to have the = effectivene's s of your pool. And so ,
t 513$ ~ r -
' that?will- shift those numbers upf and down. l
'14-
;You see the pool -- wel'1, for;ihstance',-in here we :
. i
. 15 4,,,,.!whenwegothrbughthepool,welcatchessentiallyallthe i w '
l- -16 2
. material that-is-passing through the< pool. .But atz the point 8 ' 17-
_g Lwhen the" containment fails, then you bypass, and that's the 3 : t, 18 '
'4 crucial iss'uei because-then whatever hasn't.gone through the .
.I' pool: basically goes on out :much .more - easily.
sg. fr - : So, the l timing, -I think,, is very important. 21 -
-MR; SILBERBERG:~ Maybe somebody.else can help, but a
.5 c, 22 I.think whatLBilliislasking is, in BWR.2, other than tellurium', -
23
; basically yotO re coming out lower?for TC gamma prime and - AE
- 24 '
i p 4 gamma - prime,. - L. df'
" sg ~
Okay.; What brings ~you to that i~point. .And in BWR'3,
, - i
?
9 f
^
"'y e-
- ~ '
k, y..., , w , n;a - c ,~ - .w isi:..n ',dv .
;*GV,~.. - + -
9 ::g - , ,.y . .. . . - , . ... , a _ :.- ;. . '.- [MMaji ;15:8! 178 1
'you're. coming:out higher for.the TC gamma. And again, what f}'
1 2 brings you to that point?
- 3. MR.JGIESEKE: Just the results of the calculations.
4 I wouldn' t attach .too much -- I wouldn' t. hang myself because 5 ofiany 'of those numbers.- They're a little bit up or a little 6 bit = down.- There are a lot of uncertainties in the calculationo 7 -an'd a lot of things:that contribute,-and-those all may go up
- 8' !and'down, so I wouldn't attribute any particular significance.
8 -MR. DENNING: Can I comment, also?
~
. 10 I think that -there -are a number .of very important
' ll processes here. You get large uncertainties on the'them to 12 get really. good -- where_are the areas-of uncertainty that 13 are potentially major. contributors,'as well as what are the 14 -
~ differences'between the WASH 1400 analysis results and the
.15 '
results that we had'here. Obviously, one area is in the
.g.
16 primaryLsystem, I. We didn' t take any credit in WASH 1400. It O '
-$' ~ 17- wasn't the major. contributor here, except-for the aerosols.
i
- 18 Actually, all Jim has really shown is the volatiles.
f. 1 ' 18
-[' The involatiles, we would see bigger. dif ferences from WASH 1400 t.
20 l had they been shown, due to the primary system . behavior; but j 21 therees'still significant uncertainty as to what the primary
.y.
] 22 i
-system retention is.
23 Another contributor _is the timing in the AE
.24 sequence.of containment failure.
,.., Containment failure has been 26
'later, and;more of the initial release during coremelt to the
- p. ,. .-.J.
~. -
mm . , . ;
n- -
. . _ . . _ il
- MM2jl"l15 9 179
.'I ' suppression' pool, it would have had a bigger effect. It had
.% 2 Aj a' major effect on what went through the pool. If there had 3
been more1 time for stuff _to go through the pool, it would have 4 had a more dramatic effect.
' ~
'If we look at the TC and TW sequences, the 6-
" suppression pool contribution was important, very important, but" the DF Lthere was not really as large as many people would 8 - think that fit would be. !And I think that later we may get some comparison be' tween SUPRA and SPARC, that there could be 10 significantly greater retention and suppression pool in .the TC and 1N kind of cases than we' currently have. We get back to the uncertainties again. - We have the quesion of where does 8
the containment 1 fail, is:it in the dry well or wet well?
~() 14 That certainly?is a major difference between what we're work-4 15
.ing on and this.
. ]--
I Also, I'd'like to say one more thing about theeBWR 3 0 17 2 and the effect of the reactor building. I think the effect I 18 i a of 'the ' reactor building for some sequences will be bigger than we saw with the TC sequence.
.g f ~
f In' addition, there is this possibility of the
; '21' j : sprinkler system working for a' period of time; but then c
'n -compounded.with the uncertainties of the hydrogen generation 23 Land-potential for. hydrogen explcsion, there are an awful lot
' 24
- of' areas that have to be examined in more detail than we were
~
bh 25 ' cnd:15 able to.
.re i
1 _ ;MR. SHERRY:' I think I would like to address this 7.. i; ) 2' question to Rich' Denning. One major difference is the 3 l higher releases foritellurium between_your AE gamma prime 4
' analysis;in this study and the reactor _ safety study BWR-2 1 'infthis category. Was that the release rates of tellurium?
6 I guess we're about-the same for the safety study and-for
~7~
this-study in that-a large percentage _of the tellurium was 8 released dt'. ring the melt. concrete reactions. 9-Is the difference in the amount of tellurium 10
-released to the environment: basically due to differences in
~
11 the time of containment' failure: between.the safety' study 12 - andithe assumed _ time of containment failure in this study? i 13 MR. CZYBULSKI: Time and location, drywell.versus
\
.s.
- .~
1*'
- we twe ll . - The safety study was a wetwell failure, and they
- 51' - >
' 15 got'more: benefit in the pool. i 1 -
16 MR. SHERRY: I'though't there wasn't any pool 17 'k scrubbing in~the safety study. 1 i la MR. CZYBULSKI: Yes, there was very much so, t 19 except-for saturatedJpools. In cold pools there was.
.20 MR. DENNING:- In AE that would be true, t
21 MR. REYNOLDS: Why doeslit matter so much whether 22- thefailureisin'thewetwellor.khedrywell?What is it in 23 _the geometry'in the patnways? In!either case, do.you go 2 4 _- through.the pool after the failure'? What's the big 25 -difference in the geometry? T A Y L GE. A SSO CI A T ES-D'f
.s
-162 5 I . Street, N.W . - Suite 100 4
. W ashington, D.C. 20006
-(202)'293-3950
7- g g- . ,
,, , m-: . .~ , . - - - ~ . 2 x
, -' '- . w
.p.
~ t , J MR .' TGIES' EKE : - After:the containment failure you
]G C - . ,
, %gj ,
;2- !
don:'t go through the pool ~. 13 MR..REYNOLOS:' -In either case then?
-1 4- -MR. GIESEKE: There was some. credit given for y.
5 = flow upfalong the. annulus outside the primary containment. 6' 'Wasn't thereLin HASH-l'400?:5 {7 ; gg, CZYdOLSKI: Excuse me, Jim. With regard to-
.the?questi~on,-:in WASd-14001we assumed that the-failure took
~
aJ i) lace-in'the wetwell and that the failure took place'at 175
~
'9 ps i . . As' a resultL of those' two assumptions, essentially all 210 111- the= melt' release.Eand/a portion of the release from the 12-core-concrete interaction passed through;the pool before 113 :
thefco'ntainment failed.- l'- ~
-MR. REYNOLDS: Was that because of the 175 versus
- 15. .the 1327 OrJwas.it because of the geometry?
- l. MR. CZYBOLSKI: A combination of the,two, as it 17 1 turns'out.- The l'ower-failure pressure and the different-
.18'
-location in1the present study results in :less of the
. 1'9 release' passing.through-the. pool.
20, MR. KASTENBERG: Just-a follow-on to this 121' question. If.. ~youl had used all tihe tools. hat you used in . 22: WASH-1400'but:]ust cha'ngedLthe failure pressure and-made a . 4 23 wetwell.fa'ilure, how closeLwould.-you be to what'you have-
- 24. ; got[usiNgall'ofyournewtool's?
A 25 1 ~MR. CZYBULSKI _I amfnot sure I exactly followed
- lY:-
.Q)
- t. - g. ,
T'A Y L D E ASSO CIA T ES 7162 5 i Street. .N.W. - Suite 1004 W ashington, D.C. 20006
-(202) 293 3950 O
t , s J ' , P' 0" ' ** *- k *03 Y; ** W } M ' " A* *
'* ' A N ?' V'" W"M N * ~ 1 *I '
.c =_ .
.m.
. . . . - _ . . . _- . _ s l@2 1
your question,.but let meL try to answer. L.. 2' HR . : KASTENBERG I am trying to understand 13- wnether the most dominant' thing;is the fact that you have
~" changedfthe failure pressure in the location of failure or 5
whether it's'all of the physical chemistry we've discuss e~d 6 all day today. That's what Ifam trying to ascertain, which 7-
.is more important.
8
-MR. CZYbOLS KI: -It's my impression for the AE ,
sequence, by far and away tne most important thing is the 10 location and' level of containment failure. That is my 11 personal impression of the results. 12 MR. KASIENBERG: Could.you comment on the TC? 13
.MR. CZYBO LS KI: 'In case of the-TC, as was pointed 14
.- ;out earlier, in' WASH-1400 we did not take any credit for 15
.the suppression' pool, and we-are taking some credit now, ,
16 ana'I believe to a very crude level of approximation. 17 That's really the principal dif'ference. la MR. RITZMAN: I want to correct'that. In 19- WASH-1400, for accent TC containment remained intact 20 through coremelting whenLthe suppression pool was subcooled 21' and a DF of 100-was used for the pool for the melt release. 22 So that made the melt' release relatively low, and that's 23 ~ the' reason TC gamma fell in BWR-3 rather than PWR-2 in
- 2' WASH-1400.
25~ I am sorry. That is' explained in MR. DENNING: .
,~ '
-( x)
, T A Y L O E; - ASSO CI A T ES 162 5 i Street, N. W. - Suite IJ04 W ashington, D.C. 20006
-(202)-293-3950 ,
g,, ge + , w ,s 4 , -.. <-
~
.~ 103 1
.tne. report, and I didn't want to confuse you by that. But M.A.
; .2
' TC gairtma prime actually fell -in the PWR-3 in WASH-1400, out 3'
'the BWR-3 is really determined, I think, by the TW gamma.
- MR. GINSBERG: A comment and a-question. A 5
-questionito the general audience, the other peer reviewers.
6 First, isn't it possible to fall the wetwell in 7 the vapor space and still take credit for scrubbing; or, 8
.second, it can fail'below the water mark and then you'll 9
lose'your water and you, don't have any scrubbing. So you 10 'have to'make'one of those two assumptions. 11 how, listening to all this, what I feel is that 12
- the. severities of what's happening is governed by when 13 containment fails relative to when the vessel fails. I I"-
haven't heard ~any comments from the other peer reviewers, *
-15 who I would assume have done similar analyses to see l
-whether.they basically agreed with the sequence.
17 I am wondering whether I am in order now to ask 18 - for a comment from other reviewers to see wnether they 19 agree on the basic -- is there general agreement that the 20 ' basic sequence and the calculations that led to this 21 ~
-sequence are basically valid?
22 MR. SILBERBERG: Anything'is in order nere. 23 MR. WALKER: You know, I think we're at'a point 24 where on the MARK-I we simply haven't done the containment 25 Levent trees with the splits that nave.to do with [- TAYLDE ASSO CIA TES 162 5 I Street, N. W .~ - Suite 100 4 w ashington, D.C. 20006 s (202) 293 3950
*~ -m m . ;m- .<i .
e
m
- , - ;~ - ,,
- .n
-c 164 l l
4 1 containment: failure modes and-the flow splits between flow
;,q .
9(jf' 12 out of the containment ~and into the pool. Those sorts of 3 fthingsEsimply haven't-been quantified as yet.
"J ~What I see here is a bunch of calculations
'5 withoutLthe basic systems work that comes out of a 6
containment event tree having been done yet. I would 7
--sugges t you ought ' to do that ' work. - Tae . uncertainty stuf f, e' maybe:you'ought to do-something like the DPD stuff in the 9
PWRs, but that's alfferent'than'trying t'o work the basic 10 systems _ problem with the containment event tree. That just 11 hasn't been done. 12 MR.- BURNS: Bob Burns, EDS Nuclear. 13 I can respond to Teo's question on the results 14 we're seeing in.the IDCOR program. We see similar types of
~
- - 15' behavior that.are being described here.
16 The major difference we have in our results in 17 the fact that the containment failure times seem to be a
- 18 little bit 1 1ater. And as a result, _according to the kind 19 of' things that' Rich Denning was describing, we do get more 23 -of the material di'verted into the drywell -- I am sorry, 21 from the drywell into the wetwell, also from the vessel 22 into the.wetwell where it stays up until the time of 23 containment-failure.
24 As a result, we get about 20 percent or so of tne type releases--you have been presenting here. e ..
~ TAYLGE ASSO CI A T ES-1625 I Street, N.W. - Suite 1004 .
W ashington, D.C. 20006 , '(202) 293 3950
~ . _ . _ _ _ _ . _ . _ . _ ~ _ . _ . _ _ . - -
- - - . - __-----m
[-( _. _ _. . . . s. c 1. g 185
~1' lMR. SILBERBERG: You get one-fifth. i A 2 1 C, - : MR. ' BURNS : W get one-fifth of those results,
.e 3'
one-fifth lower. i- In response to your question, Ted, the 5-vaporization release from the drywell in all sequences 6 except TW?does go-down into the wetwell because TW is the I' only: case we have in IDCOR where the containment failed 8 before melting occurs. l MR. SIL8ERBERG: Thank you very much. 10 MR.: COOPER: Although we may find eventually that
. 11 indeed the timing is crucial,.I think there is a lot of 12 value to'the iterative improvement process that we're 13 l- seeing with regard to the models. And one of the things you I'
g discover wh'en you do this is when you're close enough to 9 l (d' I' atop -- and typically that happens when you have changed 16 something and it doesn't make much. difference.
'II i So I think there's a lot of'value to'it. Those 18 people who do numerical integration there, you keep having I'
the step size and finally it doesn't make a d!.(ference and
'20 you come away relieved and you can step back a little bit.
21 MR. WALKER: The thing that worries me, we're 22
.only iterating on part of the problem.
'23 MR. SILBERBERGt Yes, that's true. Point very 24 well taken. Thank you.
25 ' MR. BERNERO: Excuse me. I. wonder if I could *
', TAYLOE ASSO CIA TES. ' S/ 162 5 1 Street, N.W. - Suite 1004 W ashington, D.C. 20006 j (202) 293 3950 t _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ . _ _ .
+;: . .
J e 186-
'l ~
-.o- -
raise'a point with Bob Burns.-- I don't know if he's in a 7 2 1v position to' answer it. Has IDCOR as of this time done any 3
; deep containment of entries of the type that Dee was' J referring ~to;on'the MARK-I containment on the MARK-III?
5
.MR. BURNS: Deep ~ containment?
MR.-BERNERO: I think you know-what I mean. 7 MR. BURNS: Detailed trees, I taxe it. I know 8' ~.IDCOR hasldone some, containment event' trees. I have'not . , seen them.myself, so I don't;know about them. 10 MR.'BERNERO:; Thank you.
' II MR. SILBERBERG lWe are going to take a 10-minute
-12 break', but it's going to cost us 'in our time. We're. going
'I3 to close later tonight.
.,3 (Brief recess.)
- ' '/ .15 16 17
. 18 19 20 21-22 23 24)
/
25 i- TAYLOE A550 CI A T ES 1625 ! Street, N.W.~. Suite 1004
. W ashington, D.C. 20006
-(202) 293 3950 We
! f',/f
- v .. .
O/ - ' ig7 i Per ' l' MR. BERNERO: Let's get started. I 2 40s MR. GIESEKE: I would like to move along and go 3 into the tnird topic that we have in terms of our analyses 4 to date, which is Volume 3 of BMI-2104. (Slide) 6 But we are talking about MARK-III design BWR, in specific, Grand Gulf. Just to. refresh your oemory, I nad a 6
-table of volumes, he are down to this point at the moment.
9 (Slide) t 10 This is in the process. This work is still Deing
-11 done. We have looked at the TC and TQ0V sequences. We're
) g still worxing.with the TPI. I think if you all picked up 13' li-
~ copies of material we have, tne first part of this- Volume' 3 14 4 report available tu) you, which coversfnot all of.that which
!\
we will talk about but which covers the first part of that report, which is pretty much through the: thermal hydraulic t
't 1
17~ cart, tne next part-of the report that-we're working on
-i
.i 10 woula be the release from-the fuel ~and-then the tra'nsport.
So the report :goes that 'far. . 20~ The discussions today will-go'beyona that ' 21
, somewhat. 'We<didn't have a chance to gettthat'all written-l t-
! 22 up;as yet,
~
i (Slide)
... 24 But we will be talking about some of f it, and the 25 ~
.. topics we will oe covering today will be-the sequence-
'[ TAYLOE ASSOCIATES - '
%7 1625 I Street,'N.W. - Suite 1004 Washington, D.C. 20006 '
,(202) 293-3950
. .- . a ,. .~-.:,- , , - - 4
16e
-1
-descriptions and thermal hydraulics. Rich' Denning will be
- n. 2
- doing that. And we willTalso cover release from fuel, transport in the coolant system. Mike-Kulhman will be doing 4
that,. covering that topic. And I will try to make perhaps a 5-statement or so at the end, and that will go on. 1 MR. DENNING: I am Rich Denning. (Slide) O The Grand Gulf plant is the plant that nas'been-9 selected to'represnt the. MARK-III.- how, the reason that we 10 picked Grand . Gulf was it was used in tne reactor safety 11 study methodology applications program, which is a study of. 12 four designs that_was cone 1following WASH-1400. So we nad 13- 'quite a; bit of information on it . There would have been.another logical. choice, Y:.) C~: '33 and that would-have'been the GESSAR plant, which we did not-16. co,-and'that procably has-caused us a-little bit-of grief. . 17 General Electric.has provided-us a_ great deal of 18 information on the MARK-III design. In-some cases, we have 19 nad.to_ adapt that to_the Grand Gulf. The analysis that_we'will show today,~we have r 21 recently had some additional comments by General Electric 22 on the way tney feel the MARCH analyses should be done on 23' some of these. I don't think there are major changes, but 24: there.are things that we will'have to be giving some 25 consiceration--to11ater. But I don't think they.have any ("; TAYLOE ASSOCIATES
.Q' 1625 I Street, N.W. - Suite 1004 '
Wasnington,- D.C. 20006 .
-(202) 293-3950-
_ _ _ ~. n ;-;.4 , ,:_ . .,.. 4 m -- -
_u , , g A,-
; , . :.1 _ 4, _ _ r: n.. i Jg;s; A y i - :
4;
+ .- , .
a
- 'g - !
{ t
.' significant.effect on the results taht we're going to-show m..m .
+;
- n ~. , 3 12 ,
-- by -
you today. - 4 3 The sequences-that we will be analyzing <are the. U
'4-TC,3-TPIs--\I will explain what TPILis,-but/in reality it's
<t- .
'5,' so similar:to TW it-Lreally doesn'.t:make.much difference i
'6 4
-which o.f thesetwe're:doing-in th'e.TQUV sequences, that 7
these are: expected to be risk -dominant sequences.. LIn the boiling: water reactors--there are so'many
'91 t
. 2diverseisources'of-water'~to cool the core that hte risk lo- [ studies generallyfindicate that it'.s!the' transient 11 ; i .J : sequences that are risk-dominant sequences and not the b ;12 . .. 1 . , . pipe-break : sequences . 3- -
.(Slide) ,
14 _ . Grand Gulf is'a-~ steel-lined reinforced concrete '
.s c ~W 15 t ,
cbntainment. There are'iother MARK-III designs:that are (freestanding steel: containments.-~-But' Grand; Gulf is~a
'17 I reinforced containment. The drywell'in'thisLease is here.
I
- Thefwetwell orithe: suppression pool is now-on'an annulus ,
19- . around theioutside-of.theiplant. - n 20' [ fTnere,is~ water"inside_of.a wearwall here'that l'
- gets': depressed in-the: event of:a-pipe-bieak accident that
, L
- L 22
. pressurizes the'drywell.-Land?then the flow.that goes
! 23 -
~
L- :.through ' the'. vents really goes horizontally through the , !? '24-
. . ~ .
{" S - vents.. 3- , Tne outer containment is' kind'of like-the vapor i ~ 1 5[;b :TA Y LOE o ASSOCIATES ikD.,/ T il625 I Street, N.W. - Suite 1004
. 1 Washington,~ D.'.*. 20006
. (202) 293-3950 , .
. , s ,,
=1
x -.-
- 9
~190 1
region of.the wetwell,'but it really is the containment. i This is really the containmment right here. s The primary 3 containment is1this' region right here. It sees the top of 4 the' suppression. pool-so that anything that goes into the 5 suppression pool will go up into this region here. 6
-And of course,.there is a steel liner right 7
there that'is the vapor barrier'or the' barrier.that 6 prevents fission products from being released tc the-9' environment as'long as.the containment remains intact. 10 The location of failure and tne pressure that we 11 are taking into the failure pressure is 72 psia. The 12 location is the junction of the cylindrical. wall and the 13. dome that is at tnis location right there. 14 (Slide)
$2 15 In the case of the MARK-III design, the Grand 16 Gulf design, there is a confinement building around'the 17 building, but the standby gas treatment' system is much 18 smaller than at Peach Bottom.
19 It was our feeling that in the event of-failure 20 of tne primary containment, that that confinement-building 21 would not offer additional retention. JSo once the 22 containment fails, we-have release to the environment. 23 Incidentally, this is a pretty healthy margin 24 above the design pressure, which is 15 psia. That converts 25 into 3.8 times design.
. ", 'TAYLOS ASSOCIATES
1625 I Street, N.W. - Suite 1004 Washington, D.C. 20006 (202) 293-3950 :
l l 4 ..
.l
n . ._
~ ~
r 191
^
Ll= MR. KASTENBERG: Could you show us where the vent
-r+R
.2 ~
pipes are on the down plenum?
, . ([
3 MR. DENNING: You mean where the. vent pipes are, 4 the vent pipes tnat go -- if.tnis pressurizes, how does the 5- ~ material vent?- You see these little horizontal tubes right
-6 Lthere? There are three layers of them. There is water in 7 this region right here.
8 I think that actually in the' drawing there might-- 9 be a few dots that look like concrete.in'there wnich aren't'- 10 true. This is.a wall right here. This is: called the 11' wearwall, and there's a water level in here that under 12 normal pressure conditions would~be-the same. When this 1 area pressurizes, it pushes the water level ~down here so 14 ' that'it goes below this horizontal vent, and'we relieve -
!O". 15'
- hat way.
16 tiow , as far as the accid 5nt' sequences that we 17 ' are studying, this only comes-into' play after we have 10
. melted through the lower head because we are'looking at 19 transients.
20 During-the transient, while the core is 21~ degrading-within tne vessel, you again have the steam line 22 and a relief line that go directly into sparge rings in the 23 suppression pool. 24 Did that answer the question? 25 MR. WALKER: The'75-psi tailure pressure, who did f'S . TAYLOE ASSOCIATES V -1625 I Street, N.W. - Suite 1004 Wasnington,' D.C. 20006
-(202) 293-3950 m.- . -. .. -
I _(y - 133 1 that?
- D
- i 2 'MR. DENNING:- Actually, I am not'sure what the 3 . answer to that is. That is a number'that we have really L 4 gotten'from tne industry, and particularly from IDCOR.
5 But'we don't really have good credentials'for.that, . I 6 although they may exist. 7 (Slide)
- y. Tne first sequence is the TC sequence, which we e
L- 9 have actually tal.<ea about already for the Peach Bottom 10: reactor, if you recall. l 11 And tnis time, the containment failure is at 80 l l .12 - minutes. Remember that the' system does not. scram. In this 1 ( L13 case, the core power levelLdrops to 16 percent. That heat 14' is' dumped into the suppression pool. The suppression pool l 'N k'hN 15 neats up, and eventually.the containment gets to the point l I 16 at wnich it fails.
- l. 17 he have beginning of coremelt at 118 minutes, 18 and coremelt Jis occurring in a. failed containment.
19 MR. KELLY: Question. Jim Kelly,-University of 20 Virginia. l 21' .Why is it 16 percent of this reactor and 30 22 percent of the Peach Bottom?
.23 MR. DENNING: I would rather let General _ Electric L 24 answerL t hat, if tney.would be willing to do that.
f 25 MR. HOLTZCLAW . Kevin Holtzclaw from GE. V L [ TAYLOS ASSOCIATES I L' ' .1625 I Street, N.W. - Suite 1004 ~ Washington, D.C. 20006 (202) 293-3950 h u_ . _ _ - . - L_ _ m _ ____ _
~ - - .
^ ~ ~
193
'l The difference is in the equilibrium power level
.N 2 following an ATWS eventfare different between a MARK-I and
_(( 3 'a MARE-III design, primarily'due to the emergency system 4 enat would be providing coolant to the core.- So it's a 5 ciffe' rent flowrate, different water level, and consequently 6" different moderation levels. So you.have a different power 7_ level. B 'MR. DENNING: Okay. Then we have the pressure 9 vessel failure occurring at this time. In each of_these 10 cases, we have called tne containment failure mode gamma 1
- 11 prime, although you will see when you get to-TQUV there is 12 a slight difference in the gamma = grime.
13' (Slide) 14 The flow pathway within the vessel is exactly i . 15 the same as modeleu for the Peach Bottom react'or. Again, 16 core through the ste,o separator, some fraction going
\
17 through the steamalryer, some fraction bydassing the steam 18- dryers. Ana I~ won't show you the MERGE and MARCH model -- I L 19 am sorry, the MERGE'and TRAPMELT model, wh'ich is 20 essentially tne same as what we use forlthe Peach Bottom 21 design. . 22 Now, the next results,nowever, fare.somewhat 23
- different than what we obtained for . Peach Bottom.
~
24 (Slide)' - 25- Th'is is'what I alluded to this morning as some 1 1
, ~ .
( l; TAYLOE ASSOCIATES U 1625 I Street, N.W. - Suite 1004 Washington, D.C. 20006 (202) 293-3950
-l
, _ 7 194 l' -sensitivity.weisee to the coremelt modeling and how it can .
, .7 3
- (,/ 2 affect the retention-efficient products.in the primary 3 system. So I will explain some differences here and
'4 remember them when Mixe Kulhman gets up to show you the 5' results that he has obtained onprimary system deposition.
6 These are the gas temperatures.. We haven't 7- labeled these curves. I am sorry they're.not labeled.in u your report. 9 MR. KELLY:- May I ask you a question-about the 10 previous slide, please? 11 MR. DENNIhd: Ine previous slide on the pathways, 12 was that the one? . 13' MR. KELLY: das the containment failed at the 14 time that this flow takes place? u ,' 15 MR. DENNING: Yes. The containment already 16 failed. 17 MR. KELLY: Does the suppression pool have any 18- water in it? 19 MR. DENNING: I will get toLthat part of the flow 20 path in a minute and show you that too. Yes, it does. And
- 21. that's a good point.
22 For all of the sequences that we've analyzed for 23 the MARK-III, there is water in the. suppression pool in the 24 pathway, . and there's water in the pathway. Let me go back 25 and let me show you that figure now.
.1, (P- ' TAYLOE ASSOCIATES 1625 I Street, N.W. - Suite 1004 Washington, D.C. 20006 (202) 293 39504 c- L: :L'::_ L~_
a
-~
lir5 i.
-1 <
(Slide) (w} 2 Here is the containment. Remember, we nave 3 transient sequences where, while the core is degrading, 4 your.patnway for release is from.the steam line through a r ! 5 relief line'and into the pool. And the water remains-'here,. 6_ .and actually'there's a significantly greater submergence, s7 That is .the tission products are released much farther
- b. below the surface of the pool in the MARK-III design than 9 they are-in the MARK-I design,
~
lu When~the core melts through the lower head, it ! -11 will drop into the reactor cavity. Then we-have a pathway 12 into tne.drywell and now back then again into the l ' 13 . suppression pool. So containment failure is here. We assume. 14 that when the containment fails, it-does not displace this 15 water here. 16 For a number of designs, General Electric has 17
~
'really gone to great extremes to indicate to us anyway, I 18 thinx quite effectively, that. failure will-be here as 19 opposed to cown there.- That's obviously very important to 20- tnis' conclusion that you have water always in the pathway.-
21 MR. BERNERO: Excuse'me, Rich.'If only they would
- 22 put a vent there.
'23 (Laughter) 24 MR. DENAING: Okay. But the point is that the
'255 . failure'of the containment does not displace'the water. The C1 TAYLOE ASSOCIA FES
(/ 1625 I Street, N.W. - Suite 1004 Washington, D.C. 20006 (202) 293-3950 s .. r.e s , s+ vse, ' - +m N - ~ .- A+ >
. ~ ._ - ,. - , .
, , . 5 f96
-l- release . pathways are' always into water with water in the n ~
-2~ pa t hway .^ .
T. ,
~
.3 'MR. VOGEL: In addition to that, that's below 4
(4 ' grade , isn't it? 5 ~MR. DENNING:. -It's below grace, so you fail here.
~
6 rhere's some question as to where the water would run out 7 anyway. u (Slice) 9' So,those'are basically the pathways. l 10 Now I would lixe to get back to the structure 111 . temperatures in the TC sequence and show you the difference
- 12 between the structure temperatures'we have calculated and l -.
i 13 - wnat we calculated for Peach Bottom.- 14 how, the form of it'looks basically the same. I 15 . am sorry these curves aren't labelea.. Actually, you.can 16 figure out wnat they are. - But it turns out this is the 17 steam separator here, this temperature right here. And the 18 steam cryers.on one of these' lower temperatures. 19 But the.important' thing is that this temperature
'20 . peak gets turned over much more.rapidlyfin tnis particular 21 analysis than it did in the Peach Bottom analysis. And the' 22 temperatures are brought down to significantly lower 23 - temperatures. ,
24 In the Peach Bottom analysis we saw those
' 25 temperatures in the neighborhood,-evening out here in the
' +- , TAYLOE ASSOCIATES 1625 I street, N.h. - Suite 1004 Washington, D.C. 20006
, 202)
( 29 1 3950
,peMr',.w.>~.+. m ra. s,-w-*. .*M 9'* --
Q. 197 1 neighborhood of 1,500 to 2,'000 F. And this difference of 13. L;{p 2 ~500.or so degrees is; going to make a fair difference. 3 .The difference'in the way we have treated the 4 core modeling for tne Peach Bottom and Grand Gulf is>only V l 5 in the way we have subdivided tne core into radial power [ 6' regions. In the Peach Bottom analysis we had 10 radial 7 : regions across the core with equal volumes. In the Grand 8 Gulf analysis we have broken the regions into nine regions, i .9 and they are ' not -of' equal volumes. And indeed, tne center 10 region, wnich is the highest or the region'that has the 11 highest power level, it's a fairly small region. 12 And'apparently, what is happening is it is l E
-13 melting and slumping earlier than the equivalent region 14 witnin our: Peach Bottom analysis, giving this extra cooling
- 0\_/ 15 j of stcam, bringing it in earlier.and not allowing our l 16 temperatures to get as hot as they were and keeping them I
[ 17 down cooler. So it's part of.this melting / slumping modeling 18 tnat snows considerable sensitivity to the way we do that L 19 melting / slumping model. l 20 MR. VOGEL: Rich, with such a sharp peak there, L 21 does all of the structure get to-that temperature? I would-22 think it would not. 23 MR. DENNING: Yes, this is basically -- I think 24 .there is quite a bit of -- the artists have made this peak 25 much sharper than it really is. .It doesn't peak like that. i
-(#,
' TAYLOE ASSO CIATES 1625 I Street, N.W. - Suite 1004 Washington, D.C.' 20006 (202) 29}-3950
=F"
- " ]L
- M ,,' - ' '
198 l' 'Isam'sure it's much more rounded peak than that. And this
+
2: is like 10 minutes from here to here,-so it's not quite
- 3 that, not quite as sharp as.it appears there.
l. 4- But-this representation ofuthe structure, the
-5 ' steam separators, which are pretty massive, represents all, 6- of the mas of those structures. So that's an average value 7 for all of the mass of that structure.-
8 MR. VOGEL: I would think, in reality then, the
- 9' surface temperature would be higher and'the outside 10 -temperature would be lower.
l 11 MR. DENNING:- That depends very much upon.the ! 12 thickness of the structure, of course. The steam separators 13 aren't that thick that their response time is that long. So
'14 in that case, it's not true.
l -- .;q ! .9 15 Certainly, in terms of other structures that we 16 have in here, there could be an effect lixe that over a 17 10-minute time period. But definitely not for these i 18 . structures. Their time constant is less than that. 19 MR. SEHGAL:: daj Sehgal.
'20 Suppose you contain eroding of Peach Bottom, 21 would you get similar results?
22 MR. DENNING ~ Same results. 23 MR. SEdGAL: So you would have greater potential 24 of.-- 25 ' MR. DENNING: There are some other changes that c- , t
\ ".
TAYLOE ASSOCIATES 1625 I Street, N.W. - Suite 1004 Washington, D.C. 20006
-(202) 293-3950
199 1 naa to do with the power profiles. This had a flatter power r^ (s 2 profile than the Peach Bottom. I would suspect if we did 3 the same nodalization, we would nave seen similar results 4 for Peach Bottom. I don't think that this is necessarily 5 closer to reality. 6 MR. SEHGAL: This might be longer. 7 MR. RITZMAN: Quickly, you're talxing now just 8 about core nodalization, not talking about MARCH 1.1 versus 9 MARCH 27 1C MR. DENNING: No, just core nodalization. 11 MR. SEHGAL: 500 degrees difference? 12 MR. DENNING: Like 500 degrees, yes. 13 (Slide) 14 I will just say a few things about TPI and then (D. x/ 15 go on to TQUV. 16 The TPI sequence is a transient with a 17 stuck-open relief valve and loss of decay heat removal from 18 the suppression pool. ..s I mentioned before, it looks a 19 great deal like the Ta~ sequence, and in the long term it 20 looxs almost exactly lixe the TW sequence. ! 21 he have the core degradation occurring at low 22 pressure. I did forget to mention that. The core 23 degradation really occurs at high pressure in the TC 24 sequence. It occurs essentially at containment 25 back-pressure. Here the -- but what actually happens here ( TAYLOE ASSOCIATES
- k. ' 1625 I Street, N.W. - Suite 1004 Washington, D.C. 20006 (202) 293-3950
200
~1 is we have a stuck-open relief valve; we are discharging f 2 into the sump -- not the sump -- into the suppression pool 3 and heating it up. The decay heat removal for the ,
4 suppression pool.has tailed, so eventually after some
-5 substantial: period of time, we heat that up to the point at 6 which the containment fails.
7 We-then lose our makeup water to the reactor
.8 vessel, begin coromelt. And here is pressure vessel 9 failure. Again, we have analyzed the gamma prime failure ,
,10 mode.
ll The temperatures are quite similar to the ones
;12 for the TC sequence, so I won't bother to show you those, 1
.13 and I will move on to TQuv. TQUV is again a transient.
14 MR. WALKER: One question on TPI. When you ! (Oe,7 15 release the fission products, are you sure the water is not , 16 completely gone? 17 MR. DENNING: Yes. Even though you will have 18 flashing of the pool,.you don't lose that much inventory of , 19 .tne pool. Tnere's still a lot of water in the pool. ' 20 (Slide) l 21; .This transient is, in some respects it's 22! something like TMLB' in.that you lose all makeup to the 23l system. 'The containment failure occurs -- I actually have it4 ' . these reversed as far as what I want in time. We begin 25 coremelt at 82 minutes with teh containment intact. We have . 1 O TAYLOE ASSOCIATES
- k.',.
1625 I street, N.W. - Suite 1004 Washington, D.C. 20006 (202) 29h3950 u__.___.-_-___m- ..____..-m ____m-- _ _ . --
h .,-
, - ]
201 4 I '
!1: containment failure. occurring at 96 minutes, and I will
'n' p,.(
I 2 explain that_in a second because it is an important 3 conclusion-for this particular accident sequence. That is, ! 4 _it's an overpressure failure, but:it's really a hydrogen l 5 . burn generated. overpressure failure. 6 -The reason that we get into the circumstances 7 -that can lead to hydrogen burning and the failure in this L 8 containment have to'do dith some of the details of the way l 9 the: sequence. proceeds. And let me explain those. I 10 In this particular design,;we nave assumed that 11 there are igniters in the containment. -During the period 12- when the core is beginning to uncover and melt,.the system 13 is maintained.at nign pressure. That is, in this particular 14 accidentesequence, the guidelines to the operators tell. 1 t l I i 15 them, before you-depressurize the' system, check and make 16 sure that your low-pressure Ecc systems are operating - 17 before you depressurize and put reliance on them. 18 In this particular sequence, they did not , 19 operate. So the operator leaves the system at pressure n L 20 instead of intentionally depressurizing the. system. You I j 21 start to uncover the core, degrade'the core, produce l 22 nydrogen release, and actually. to a large extent bottle it i 23 up within the vessel. I 24 There-are then' instructions that say, when-the 25 water level gets down to 2 feet, depressurize the system. f I) ' TAYLOE ASSOCIATES 1625 I street, N.W. - Suite 1004 -- ( Washington, D.C. 20006 (202) 29h3950
f _ l 803 l l 1 -The logic behind that is it's reccgnized at that point you l m l" (l2 2 are in serious trouble, that in order to buy some more time-3- to get emergency core cooling injection in, it would be a 4 good idea to.depressurize the system, force level swell,
-5 'and recover the core due to the level swell. So the logic 6 behind it is. pretty reasonable.
i 7 As far as the implications to our analysis, 8 however, were the followings wnen we depressurize the l L 9 system at the 2-foot water level, we hate already produced 10 a significant amount of hydrogen._ We dump that nydrogen j 11 rapidly into the containment, and the containment pressure ! _ l:2 . then rose up well above the failure pressure of 72 psia. 13 And we failed the containment due to this hydrogen burn. { 14 MR. HAZEN: You said you incorporated the fact ! n
-> 15 that Grand' Gulf nas hydrogen igniters installed.- Does the 16 burn occur because the igniters are there and it didn't
.17 work?
ld MR. DENNING: I tnink the fact that the igniters 19 were.tnere would certainly assure that you get a burn. I 20 _think with the kind of concentrations you would have, 21 oxygen and hydrogen, that it's quite likely you would get a 22 burn anyway. So I am not sure how important the fact is , 23 that the igniters are there, but certainly the igniters 24 would cause a burn to occur. 25 MR. KELLY: Is the containment inerted? I i
, j TAYLOE ASSOCIATES
'V f 1625 I Street, N.W. - Suite 1004
. dashington, D.C. 20006 '
(202) 293-3950 , i I
203 s. t 1- MR. DENNING: -No. The MARK-III design containment
..e 2 is not inerted, t 3' MR. SEHGAL: You assume then that the hydrogen-4 burn pressure is above 72 psi?
5~ . MR. DENNING: We calculated the pressure to be 6 above 72 psi, yes.
.7 MR. SEHGAL: -You bring the steam out the same 8 time too, don't you?' ,
i j 9 MR. DENNING: ~ Don't. forget, we're going through 10 the cola suppression pool when we do this. The steam is l 11 going to get condensed out. 12 MR. PETRANGEL: You assume that the operators , i 13 decide to make a manual. ADA' actuation now? Now, this tends , 14 to be a little provacative. Why didn't you Assume that the 15 operators decided to actuate one of the vent lines of the 16 containment.which are, I am sure, available in these two 17 reactors in Peach Bottom and Grand-Gulf? ! 18 We have in Italy two reactors similar to theso l 19 two, and we have a small containment vaporized, not the 20 bigger, . larger containmen't which.Mr, Bernero was referring 21 to. . But you can with time actuate these lines.and see the pressure in the containment goes two times the design
~
22 23 pressure. - 24 MR. DENNING: I believe what we did was 25 consistent with the emergency operating instructions for l [ b', TA YLOE ' ASSOCIATES 1625 I Street, N.W. - Suite 1004 l Washington, D.C. 20006 (202) 293-3950' i i i T
- t 3u4 t -
g
~- l ' the. operators.- I certainly agree there are other things
-I think that what we have done is 2 that could happen.
A 3 consistent with the instructions that have been given to
- l. 4- .the operator..
5 MR._PETRANGEL: So this procedure is shown in the-i
-6 emergency procedures'?
7 MR. DENNING:- Yes.. 8. 9 10 i ! 11
+
12' I i 13 i
~ 14
. ~15
\ '
l 16
~,
17
- i-L. le l
- l. 19 l
20 ! 21 22 23 24-25
. g, 4 "," TAYLOS ASSOCIATES 162$ I Street, N.W. - Sute 1004 waenington, D.C. 20006 (202) 293-3950
'fM.D
- odyd
TA 195 205 MM:j1 19:1 I (Slide.) , I 2 \ Now, in this particular case, we did not see as much 8 effect of this nodalization as we had seen in the others -- 4 in the TC and TW. And this is -- and this temperature pro-5 file is much more similar to what we had releasing in the 6 Peach Bottom. And the amount of retention of cesium iodide 7 is reflected in that. 8 Basically, that concludes the thermal hydraulics 8 part of what I wanted to present. 10 Are there any questions on that? II MR. COOPER: I guess I find myself concerned that 12 the choice of where to put the nodes in the analysis can make II
) that large temperature dif ference -- when the vapor pressures I4 have this exponential dependence on temperature, that could 15 make the choices very important. } 16 MR. DENNING. I didn't understand your reference to 0 17
- the vapor pressures.
I 18 j MR. COOPER: Of the vapor pressures of things that 19
, are going to be released, typically have an exponential --
20 very strong dependence on temperature. If we get the tempera-
*I ture wrong, we get a very different amount of vapor release.
l 22 l MR. DENNING: Yes. 23 We 1, that -- you actually bring up a subject which fs I think is an important one that we really ought to discuss k ,' 25 - here with regards to the ability of the MARCil code to
206 MM ji -19:2 I accurately predict -- particularly peak fuel temperatures and ( 2 time add temperatures; that capability is extremely limited. 3 It relites to the choice of a single melting temperature, 4 and'we are stuck with that as far as the MARCH code is 5 concerned -- other'than the ability to do sensitivity studies i 8 'by changing that temperature. ' 7 Pete,'do you want to make some comments. 8 Just a comment on the sensitivity l MR. CZYBULSKIS: 8 of-the core nodalization, we also are very much concerned l to that the dif ference in core 'nodalization and ' power distribu-- M 11 tion seems to make that much difference on the final results. 12
- It was somewhat of a surprise to us.
II In a However, let me make the following point. (C) . I* real core, there is no such thing-as a right power distribu-to tion. It's a continually changing one'with time, or it's .
- f. .
! l 14 changing with reloading and a variety of things that I can II e go on and on and on. 1 I is So, it's dif ficult to come up with anything resemb- l l 19 ling a right or a unique core noda112ation for core power
'Ir # distribution. It's something highly variable, within certain 21 5 limits,.but highly variable.
t 22
- f. So, I think that's a sensitivity that's inherent.
MR. SEHGAL:- I would like to dispute that. A lot
" of schools have' treated distribution for the core very G 2i effectively within 5 percent, and they've been doing that for '
207
-MMajt 19,3 ,
I years.
\g 2 .Every reload, every distribution is measured.
! 3 MR. DENNING: That's not Czybulskis' point, Raj. 4 I think his point is that this is a variability that occurs j 5 during cycles and between cycles. i ! 6 MR. SEHGAL: But load. patterns are such that you 7 try to maintain as> you distribute.
- e- MR. DENNING
- But-they do vary significantly over
.9 ' cycles.
10 MR. COOPER: It would seem to me that perbsps we l 11 could, in fact, work with something like temperature frequency , 12 distribution, the way we do'with particle size distributions, 13 and get something. that, indeed, is generally correct. 14 MR. REYNOLDS: Reynolds, Virginia. i 18 I would ask a general question. t I l 16 It seems that the transient cases that you are { 17 looking at all successfully go through the suppression pool 1 i 18 in their blowdown. i 18 While you're still on the introductory part of this ; l' I, 20 reactor, are there any cases such as major pipe ruptures where j 21 you could fail the dry well, where pressures early could get s l 22 so high or hydrogen burns or something could happen early 23 to fail the dry well. M Is that possible or not?
~' 26 MR. DENNING: Clearly, it's possible. All it takes
- , .. - , - . . . . _ . ~ ,
208 p MMijl 19:4
.1 is some fairly' inventive people-to think through the. things
?-(~'3 2
. -that you can do that would.really bypass that.
3 The whole question really gets down to how likely 4 is it, which really is a difficult,'a very difficult question. 3' And 'I' know it' is one that, in the GESSAR review, they have a tried to seriously.-- General Electric, on one side, has tried to seriously look at. And I'm sure the NRC will 8 seriously be looking at that.
' That, I think, is certainly the key issue regarding to this kind of analysis -- I'm sorry, it's a key issue regarding 11- the type of source terms.that are credible for a MARCH 3 type 12 of design.
13 [ There's no question if you can force everything to 14 go through a suppression pool that you are one leg up. It's la awfully nice to have that suppression pool in the pathway, 3 l 16 and you are raising a key question, which is are there are o l 17 sequences of reasonable credibility that can bypass that? I 18 MR. GINSBERG: Ted Ginsberg, Brookhaven.
'It ' . Sticking with that question for the moment, is it I ~# true that the pressure difference, the driving pressure r
21 difference to get material to go from the dry well into the 5 , 3 22 suppression pool is just the height of liquid? l 23 MR. DENNING: Yes. 88 So, it would seem the only kind of MR. GINSBERG: (, 88 mechanism would be an explosure mechanism that would
, .<. -.t .... ,.
MM: jl :19 3 5 209 I' . destroy --
? .fi .2-a MR. DENNING: Are you saying that would destroy.
3 the dry well? [ 4 .MR.'GINSBERG That would destroy the dry well -- 8 MR. ROE About the only mechanism I can think' of i t s' . is if there's some way'that water would be displaced in places 7 that it wouldn't be because of sloshing or something of that 8 type. Is that credible? 8 I notice some other cavities in and around, over 10 the wear wall. Is that a place where water could collect - II if .there was substantial sloshing? 12 MR. DENNING: There had been lots of questions ! 18 raised, particularly during a blowdown, what would happen to I4 the suppression pool water, will it get splashed up, things I8 like that. There are questions about can you get the sup- 4 I 18 pression pool water to come back up into the dry wells for 0 17 some reason, like condensation, that type of thing. 3 . 18 We haven't examined those things in adequate detail : I' to really comment on them. [ I # MR. WALKER: Benero says I've got a conflict of r 21 3 interest. ; s 1
] 22 (Laughter.)
23 Anyway, I think on these plants the sequences that 1 24 are similar with respect to the suppression pool in the con-O ss tainment bypass sequence, just like in the pWRs, where you - -_.3 __ _ ._ .,__. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ , _ _ _ _ _ _ _ _ _ _ _ _ _
- iMa j l 210 1926 1 lose the main pipeline steam isolation valves, I gQess I'm sur- -
2 prised you didn't do one of those as an example. 3 .MR. DENNING: We did give quite a bit of thought to 4 steam isolation reliability. And from the discussions we 5 had with General Electric in Oak Ridge, we are convinced that 6 that was not too likely a pathway for significant release. 7 But once again, it was not a detailed review. 8 But that's why we did not look at that pathway. 9 MR. WALKER: It just seems like those valves ought to to have about the same reliability as the suction side valves 11 on the RHR systems, on PWRs, roughly. They're thick valves 12 that failed in the same way. 13 MR. REYNOLDS : Question on the dry well. It seems (~~h M/ 14 the hydrogen burn could be faster. Could that be' fast enough 5 15 if you had a major hydrogen burn in there? It's not inerted.
! t
] 16 Could that be fast enough to give you the pressure before it o
l 17 could be relieved through the suppression pool? 1
$ 18 MR. DENNING: It's not inerted.
t 19 Dut, Pete, you better check me on this. My guess El is you're not likely to get combustible compositions within 21 the dry well in the sequences, because you'll blow in the air 2 . M you start off with. You'll get blown out sometime earlier in l 23 an event. 24 MR. CZYDULSKIS: You tend to have more flammable
.) 26 mixtures -- if I can use that expression -- in the wet wolt
( _ __ 211 MMzjl 1937 1 than you do in a dry well for some of the reasons that Rich 2 mentioned -- in particular, in the sequences that we're talk-3 ing about here, where you are, in fact, dumping much of the 4 hydrogen directly in the pool, and only some of it will get 5 into the dry well. So, the more likely place for a large burn 6 appears to be the main containment, rather than the dry well. 7 Going back -- perhaps just stopping back to the point 8 that you tried to raise earlier, a hydrogen detonation,should 8 it take place, would be the obvious way of failing the dry 10 well. 11 Again, what the likelihood of that kind of thing, 12 I suspect it's highly unlikely. that's a question. 13 Does the Grand Gulf plant, the MR. PETRANGELI: I4 third isolation valve on the steam lines, slow isolation valve 5 15 which other reactors have -- or there are two. isolation valvesi 3 l 16 MR. DENNING: Yes, two. l II MR. PETRANGELI: Not a third isolation valve? I i 18 MR. SEHGAL: They have two, one each side of the
! 18 containment.
I 20 The one I know - .for example, we r MR. PETRANGELI:
~ ; 21 have a third.. valve, which is.a third closing valve.
2 22 l MR. DENNING: GE may know the answer to that ques-23 tion.
.24 MR. HOLTZCLOWs !!oltzclaw f rom CE. On the standard
' 25 plant design, we have three valves in series. So, that was I
JMM jlfl9 S' 212' I one of the reasons why we believe that a bypass tap right [. N- 2 c through the main steam line~would be'would be low probability. 3 I. don't recall exactly what~the Grand Gulf configuration is,
'4 But they do have a turbine if they have-that third valve.
V, 5 stock valve along that same line, so you've got a multiplicity a in the valve. MR. KUHLMAN: Mike Kuhlman, from Battelle aga'in. 8 I will try to briefly point out the similarities
'and differences in the Grand Gulf behavior vis-a-vis the-
! 10 Peach Bottom behavior that we talked about this morning. (Slide.) 12 ! Again, in terms of' thermal hydraulic characteris-I3 tics, releases from core, you'll see there are no major , I' differences again. And retention and release from RCS -- it i -
.j 15 a bit dif ferent, as Rich alluded'to, in talking about the TC
- l. I I I8 l sequences and again presents some information regarding the 0 17
- emitted particle sizes.
1 18 i i one'of the key differences between the-two plant I' '
! behaviors as far as the sequences go is, you recall, .for the I m.
,. I AE, TC, and TW, in every. case.we had a vessel dryout period,
'I 21
.Here we have this only the' stagnant phase of the accident.
22 1 for the TC; these two sequences do not have any indication of i a vessel dry-out actually occurring. These are' times which l- se f>q correspond to a lowered, very low water level in the system,
* ~ kJ g but there's'a continuous low level influx of water which
-. . . . _ _ , . . . _ . - _ _ . . .. . ( . . . .
leyjl 19 9. _
-213 15 ~provides for a flow throughout the sequence. So, there is
\f 2' not this long stagnant phase of-the accident.
3 Another difference is the TQUV does not exhibit the r , 4 flushing phenomenon that.we saw in all of the cases this. 5 morning.- This TC still has that in it, so there is a-large l e very short duration injection.into the containment,'as you l 7 -would expect.from a flow curve such as this for the TC. , 8
- (Slide.)
l 8 On the.TC time scale, you have, at about 3,000 W seconds -- or actually a little bit before that, you have the II l initiation of this high flow, beginning, corresponding to the 12 core slumping which takes place. j : () -18 14 TQUV -- you will note'this is not a thousand pounds Lper second, but it's over here on this axis. You start off I 15 with a low flow and come up to a fairly high' flow rate and l 16 then tail off throughout the accident sequence, which is a O 17 bit dif ferent than we have seen in the others this morning. l I 18 f i (Slide.) j' 8 .The percents of core inventory of"the species I " admitted at the time of vessel dryout, alleged vessel dryout r-j 21 if you will, is, again, similar -- a littlu bit more in the s 22 ,. l- TPI than for one of the sequences.before. But in general, .
" again, you have the highly volatile' materials all being
.- released early,.relatively early in the sequence, followd.by
"' 26 the less volatile materials being' emitted, such that at the
~
a
- D
-e t a. -v.
F. MM:j1'19:10 211 L i i 1 time of vessel failure, we've emitted all of the cesium p
'. 2 inventory --
3 (Slide.) 4
-- the iodine inventory, roughly a quarter of the 5
tellurium here. 6 And if you recall any of the numbers from this 7 morning, this is about a factor of 50 percent higher than 8 ! any of those this morning -- due to different core heatup 8 history, really -- but these numbers are.very much in line with what we saw for the Peach Bottom sequences. 11 (Slide.) 12 The composition of the aerosol is, again, dominated 13 by the non-fission product material and is indistinguishable 14 from what we saw for Peach Bottom. 15 (Slide.) 2 l 16 For the Grand Gulf TC sequence, again, these times 3 17 e are measured from the start of coremelting. Once-again, 3 18 y this is a total which has been emitted up to the given time. i 18 l This is the amount retained in the primary system at any time.
! 20 t Note that we do have, for example, half the cesium hydroxide 21 being retained in the primary system at the end of the .
l 22 l sequence. There's a lot of retention that takes place in 23 here. It's an unfortunate choice of the times at which this 24 table had information listed. U,. 2. In this period is.where there is a good deal of the
>0hji 19:11 215 1 cesium which has been emitted a bit earlier in the accident r3 2 sequences actually getting to the cooler regions of the 3 primary system and depositing.
4 We see no real enormous increases in the retention 5 factor of the aerosol as a function of time due to the lack 6 of a real stagnant period during the accident, the in-vessel 7 melting portion of the accident. 8 (Slide.) 9 If we look at these same numbers, same data, in 10 terms of retention factors for TC, you see the cesium iodide 11 transport. It initially is held up in the steam separators 12 and is carried on down to the steam dryers as the steam 13 separators heat up. So, you wind up with an overall retention ({ 14 factor of 38 percent for the cesium iodide, and 27 percent of 15 the total is showing up in the steam dryers. This 38 compares i 16 with, I believe, 11 percent we had for cesium iodide in the Q l 17 Peach Bottom analysis. Cesium hydroxide, we're retaining half I 18 { of it -- again, most of this is showing up in the steam dryers, i 19 and it has moved during that last flow surge from the steam f E 20 l separator to the steam dryers. And this, again, compares 21 with very little values for the Peach Bottom seqyence. 5 ZZ-l Tellurium retention is, again, complete. This is 23 a fraction of what was actually emitted from the core during 24 the in-vessel melting phase, which was about a third, about ( .\/ 25 -a quarter of the inventory for the TC.
. .-. . ~.a. .-
MMijl 19:12' 216
- 1. The aerosol. retention'is fairly similar to the
-p.
I think we had 69 percent for the Peach Q 2 earlier TC sequence.
-3 Bottom.- But again, most of it is showing up in the core just 4 do to' gravitational settlement, or removing the larger parti-cles from-the distribution.
6 MR. SILBERBERG: Why is the tellurium higher here
'7 than'in the Peach Bottom?
8- MR. KUHLMAN: The tellurium is higher here.- I 8 think it's a questi~on of the residence't'ime in the system. 10 It's all'due.to chem absorption or chemical reaction with'the-I 11 . surface. 12 .. We initially started putting in data for all of l 13 :these species 'for, chem aborption -and called some of it-J
-14 irreversible : chem absorption. - But it's due to a longer-2-
;; . .15 residence time; when-the. tellurium is admitted.
~[ ' 16 - -Remember,'a lot.of the tellurium apparently is 17 . -emitted during the high-flow regime,-for.the one Peach Bottom i
jl8 : sequence,you're thinking of. So, fit doesn't . i have time to hit
- ~
~
18
.g.
the- surf ace, - if you will b 20 MR.- HAZEN: .Hazen', Stone & Webster. 21- I notice'here,,for this TC-sequence:on the MARCH.3, 3; 22 ~ .you take cre'dit'for.: scrubbing!out'some cesium iodine in'the. f 23 steam separators and dryers,'.same. graph-TC for the MARCH 1 P
. 24 :did not, hadlonly annulus space,-lower annulus.
.,a
. 25 :Why_the(difference?
++ . n a._ w. . 4- + , .,. . .w ., .. - . - , . . . . .
!p MMejl 19 13 217 1 MR. KUHLMAN: The big difference was temperatures of A
\ _. _ 2 these structures.
3 Rich was alluding to, in his talk, the different 4 core melting or core nodalization leads to significantly 5 different gas temperatures and therefore different structure 6 temperatures. 7 Even in the MARCH 1, the cell transport of the 8 cesium iodide, what nodal was retained wound up being in the 9 lower annulus, which was downstream of the steam separators. 10 Here we're winding up with the steam dryers being l 11 the next downstream volume. They don't reach a high enough 12 temperature to evolve the cesium iodide. 13 The temperatures are on the order of 4- to 500 14 degrees cooler in this case, which is sufficient to give you 15 0 l good -- if you consider 38 percent good -- retention of the'
] 16 cesibm iodide and the hydroxide. Part of this, also, is due 17 to the reaction of the cesium hydroxide.
18 MR. SHERRY: Mike, then the conclusion here that the i 18
! apparent holdup of cesium iodide and cesium hydroxide at E
. 20 l Grand Gulf is due really to the assumptions on the core nodal-
- '5 21 izatibn and not really due to any geometrical differences or 22 differences in the calculated behavior which are not.attribut-23 able to the nodalization scheme?
24 MR. KUHLMAN:
.. I would say that's probabl,y true. It 25 certainly is the way it appears in our understanding of what
c _. .-.
- u. ,.+ 3
,.-:-v e,
- s. _
.218 lMM:jff15tl4; 6 ,
y- , <
'l? ithese-results:ar'e telling;us right now. And that is a signif- '
,_+ -
,4, -
2.u J
--(hnd:1"9: icant-point.
- 3
. 4. ;
-9%
S U 5 ws 16'
- 7 f.
e
- 8p
.c- '
.g
~ 10 4
11.: -!
- 12 ':
C i,._ -
- 13 : ,
f Ns
~.14 ~ r 4
5 is s-
> f:. _16 O
'* -F 17 e
s 3 . i : 18
- ar- .i 4.
gi . 19
- s-:
^
,t, ,
4 4.
- ; 1 , -
;o;
~
'! 22 O
tb :
^^
.. +
, , .s
,..s,.
t ax. #..'_',* [
*: . q-
^'..<- .-M :. - '
-.y )
r
- ~
Y r ' ' - _. , y
~
d z 7- __ e ,
\ ,
_s _,f " G j
^ "
I .N? k M:hk k ywy_, M p-~<:2 = - . ' 5, Q:) '3.. A _. e -. e c. s,. o : ,; . 4_ .
~"' _-
~ ~*^' '
- _.__= _ . , . .
. .%f. 'h /', y .
219 . .: ,y
'(Sl'ide)
,.m 2
'V For the TQuV sequence, as you recall, we are ,
3 again back.to the temperatures in the same neighborhood as
- 4 what we had:for the Peach Bottom, which~ Rich was showing. ,
5-
.Here we are down to the cesium iodide, essentially no 6=
-retention of 'the cesium iodide in the system.
7 Cesium hydroxide is being retained with about 40 d '
' percent-efficiency. This, you'will see in a-later slide, is
~ '
due almost entirely.-- I thinx it is entirely due to chemical reaction, again due to the better. contact time'for 11 the cesium hydroxide in this sequence. 12 Tellurium-being. emitted later is occurring in a
'13 somewhat higher flow regima-for this sequence and is again 14 residence time limiteu for its deposition.
And the aerosol retention is.again showing no 16 indications of the benefits you get from a stagnation 17 period. 18 MR. JOHNSON: Mike, at the outset, is there some allocation of cesium-and cesium iodide and cesium with 20-cesium hydroxide? 21 MR. KULrIMAN: What's done here, we obviously
.22 don't think1that cesium iodide'and cesium' hydroxide reside 23-
-in the core. Take the iodine emission rate predicted that's 24 coming from the core,.and the cesium then is allowed to 2
'25 combine'with'that to exhaust the iodine supply, which is 4 %. '
TAYLOE ASSO CIATES -
. I 1625 I Street, N.W. - Suite 1004-Washington,' D.C.' 20006 (202) 293-3950
.-:a ..- -.,+.+~ + n. ~
. ,s , , + ._ _
~- , , , . . .
,. 4 .-
^ - -
" .n ~
220 I
- - . . . . what tne results-from 0772 indicates'should be the-case for
~
!:%* '2
-q- -
the temperature and CSI oxygen-hydrogen ratios that we're operating in. All the ramaining cesium is then said to be
~4 ce'sium hydroxide.
S MR. JOHNSON: Here you want to take and retain 6 all the. cesium hydroxide. That'.s what this is saying. You 7 don't' retain.it as. cesium: hydroxide. It's at some otner compounds. There=is in the overall. reaction equilibrium
'9-chain in a system like this,, cesium iodide converting to cesium' hydroxide. going-to the trap.
U' ' ' So in' reality, that cesium iodide is not lost--in I ~
- the environment. It's still-trapped wnerever you put the 4
13 cesium hydroxide, because you are going to run-that
- 14 .
reaction'too. 15 MR. KULHMAN: . 'I-am not'following the conversion
-between cesium iodide and cesium hydroxide. . 17 '
MR. JOHNSON: You arbitrarily. allocated a certain 18 fraction to-each one and not letting the'two interact. In-19 real life they will interact. Thatiis'in a moist system, as 20 you pull 1the cesium hydroxide out,-it' leaves a hole. There 21
'is still water >around. You are going to start" transforming 22 cesium iodide.
23 gg, - KdLdMAN: iYou're right, that would occur'if
' 24 you ever removed a' sufficient quantityc of. cesium; hydroxide.
MR.; JOHNSON: You'said you'l removed 90 percent. 3 u
.3 - .
/- TAYLOE ASSOCIATES D.
- 1625 I Street, N.W. - Suite 1004.
Wasnington, D.C. 20006 _ 1 (202) 293-3950
.gb,' Jde'.'- 'S.4
' 0-% '
' ' l 221 1 1
Isn'tithatJwhat it'says. retained? 2L ~ MR.-KULHMAN: '90 xi1ograms is. retained, 40 3 percent of.the. total', total cesium hydroxide which was
'4 present.
- 5' I.think realistically ---I am not'sure, but I 6
- believe.you even have to.get above 90 percent of the cesium-7 hydroxide removed before you would get any effect on the ,
8
' cesium iodide before the equilibrium would shift ,
9
' significantly.
' 10 -
MR. JOHNSON: I would doubt that.
- 11 MR. KU LH MAN : - It was looked at in'some detail.
That's all I can say. 13
- MR. COOPER: Since we.see rather different 14 behavior for the-cesium iodice-and the cesium hydroxide,' I
. ' 15 -
am" curious aoout' what we know.abou the kinetics of -the 16-reactions that form these two. Do-we know that there'is 17 enough time, for-example, for us to move toward the
~
18 equilibriumJ that you're postulating .before we - approach the 13 ' control surfaces thatc are picking this up? 20
.bm..KULdMAN:.That's a good question. ~ That's one-21 that was looked at exhaustively, I think,'during the 0772 22 review. I would have.to refer you tocthe= appendix to-that 23
. document, which Sandia has worked since then.
- 24' ~
The conclusion of Sallach.and Ohlrich and t h . Company'out there is that'for alll intents and purposes, i - i - e($ ; : . TA YLOS L ASSOCIATES -
%^ ' 1625 I Street, N.W. - Suite 1004 -
Washington, D.C.~ 20006 -
-(202) 293-3950.
s "4 !G..- .., s+ .% .
. +> - < - 'e--
= a w- v
222 I ~ these things have reached their equilibrium distribution
-n' }'
well before they get to this. 3 MR. SEHGAL: The main thing is the temperatures 4
- are very high and the kinetics are very classic.
MR.' COOPER: It seemed to be kind of a dilute 6 system. 7 MR. SILBERBERG:-This subject, you know, is by no 8 means closed. If it needs to be debated further by the
' chemists, I think it's open_for debate.
10 MR. CRESS: . Tom Cress, Oak Ridge, 11 Calculate these chemical reaction depositions, 12
. cesium hydroxide and tellurium'with cesium iodide, do you 3
. use a deposition velocity?
. MR. KULHMA!4 : That's right.
15 MR. CRESS: Is that a. constant regardless of flow 6
. and temperature and concentration?
17
.MR. KULHMAN: Yes. It appears to be. What we are 18 using from Dana Powers' data, it's the deposition data on 19 oxidized stainless-steel in the steam nitrogen flow. 'From 20' what they coulo tell under tne concentration regime that we
'are in -- remember that the cesium-in particular is emitted' 22 over a fairly ~short time span of the total melt.
And the 23 concentration seem to be irrelevant for this deposition 24-velocity. It was like a factor of'10 difference relative
- 25 between the cesium hydroxide and tellurium deposition (f~
d',~
. TAYLOE ASSOCIATES 1625 I Street, N.W. - Suite 1004 viashington, D.C. 20006 (202) 29F3950 jet g- .4 g eA-4 . Ame-..egrep p , -W.im.as i:-p*g .g 80sidai 7- 'a ,A, r'.i N
P 323
- 3;
. velocities that we're using.- ~k MR.' COOPER: The deposition velocity shouldn't 3
depend on the concentration. 'It should depend, nowever, on
'4
~the. aerodynamic flow regime that they're in, and:that I 5
think would'be crucial. It's:probably turbulent diffusion 6 matter that molecular. diffusion except at a' laminar 7 suolayer. MR. KULHMAN:- I-believe that~to be the case. You, don't think that is turbulent?
- 10^
MR. GIESEKE: I think it's surface-limiting, not diffusion-limiting. MR. COOPER: You think it's coming into an s 13 equilibrium and coming back up? !
' 14' MR. GIESEKE: As-best we can tell1from the Sandi
" 15 experiments, the deposition velocities represent a. surface ;
16 i reaction. rate, in a sense. The-tellurium begins to approach! 17 the mass transfer coefficients, so you're getting close to i l 18 diffusion-limiting case. .But you're not'quite there yet, 1 i 19
~ apparently, and the others fall'below that. So it's-
. 20 controlled at' surface reaction.
21 MR.. COOPER: So.it'st really reaction rate-limited ' rather than diffusion-transport-limited? MR. G1ESEKE: Right. 24 MR. .KOLHMAN: Okay. Looking at the overall ' 25 Dehavior'of the primary system for the TQUV.
. O' TAYLOE . ASSOCIATES 1625 I Street, N.W. - Suite 1004 :
Washington, D.C. 20006
.(202) 293--3950
- e. pgud a h4 M yh*P'$"'APN P ' * * '-'"""#'lf P O #[ '"
224
'1
( Sli~de ) 2 'vf As you saw before, there is no retention of the cesium-iodide in the sequence. However, we do nave 40 4 ~ percent-reaction of the. cesium hydroxide passing through 5 the system for this TQUV sequence and nearly 70 percent of 6 tellurium is retained, and a typical figure of 65 percent of the1 aerosol with all of this showing up in the core 8 . regime region. With the resulting injection profiles for the 10 injections or ejections from the primary syst'em for the TC 11 sequence -- 1 (Slide) 13 -- it winds'up,1as you recall, there was an 14 g initial fairly'gooo flowrate through the system and a very. [ '* 15 low'flowrate for a while, wh'ich glave a significant 16 residence time. Then again, the flowrate through the system 17 increases and mass injection into the containment follows-10 #T cae same pattern you would expect,- which is not 19 significantly different in-terms of magnitudes _ injected 20 from the s'equences this morning for the aerosol at least. 21 (Slide) 22 The TQUV again is very'short, sweeping out 23
-behavior wnica weLaid.see.in-this sequence. And once again, 24-
-you'would have a bit of-a spike at the end of this as the
-25 '
pressure vessel faiis. Again,- this'is what is being 'CI V TA YLOE . ASSOCIATES 1625 I Street,' N.W. - Suite 1004 wasnington, D.C. 20006.. (202) 293-3950
l 225 : l 1~ 1
-injected into the pool. '
N.
'2 So that the distribution in the primary system 3
at'the ends of these sequences or at'the time of vessel. 4
- failure, for a-TC we have 38 percent of the cesium iodide 5
- retained, this being 18 percent retention due only to 6 ~
~
condensation in the primary system, 19 percent was retained due to the deposition on particles which are subsequently 0 retained. 9 This is the material that would be present in 10
.that ~ spike which goes into the pooliat the end. Similarly 11 nere, TQUV really no condensation due to the higher-12 temperatures again, but chemabsorption or chemical reaction 13 with tellurium, cesium hydroxide already affected. A 14
.relatively reasonable residence times for the sequence, and (7s, J
"' 15 again 64 percent of the aerosol is being retained.
16 The particle' size cistrioution data for these 17 sequences. 8 (Slide) 19 It's somewhat interesting to look at the core. 20 These are mass median diameters for-the core in microns. 21 You can see the aging of the material taking place. You 22 have a slower-flowing ~ portion of the accident. This largely 23 hundred pounds per second begins. You get down near the 24 primary particle size and in the core region relief line.
-25 Here you have all the detention mechanisms acting to remove TAYLOE ASSOCIATES-Si ~ 1625 I Street, N.W. - Suite 1004
- Washington, D.C. 20006 (202) 293-3950
- . . _ _ . a .- . = - - -
g.
, 226 1
- the large particles,Jbring down.your mass median diameter.
[s 2 The difference here is the result of the 3'-
~
settling out of the large particle'. s In any event, you 4- reach a fairly stable size'here until you begin to get-this 5-material-blown out:of the core.in a shorter period than is 6 ~
. characteristic-for,its removal due to gravitational 7
settling.
" Once this pulse is over with,-you are back down 9
to a similar size to what_you had before the accident, and
' 10' this would.ce coming on1down as the flow continued to bring 11 this very fine material in through the relief lines.
- 12 '
So the final pulse woulo be a very fine 13 aerosol-size distribution. 14 (Slide)
- 15 For the TQUV, remember.that was the water mass 16 monotonically decreas'ing sort of.flowrate. We have 17 initially a long residence' time to give you the large 18 particles. Then as the flow increased for a while, you 19-actually' dump to the smaller sizes. ~ And then the aging 20 begins to take effect again and then bring your mess median 21
. diameterfon up.
22 Through this region is where-the majority of the 23 aerosol. generation is taking. place. As you can see, you are 24 ~
- working with a wnole range of mass median diameters to try
. 25 to calculate your.DFs for a suppression pool.
(b '
.TAYLOE ASSOCIATES .
1625 I Street, N.W. - Suite 1004 washington, D.C. 20006 - (202) 293-3950 ~ w @*'h-iatun, es *h '@ -ad gsby -pQ. pg+ * # A* ' 1.
v- - - - - .__ - - 227
'l This~is the current status of where we stand on the Grand Gulf analyses. It's obviously still under way, a 3~
little bit behind schedule, but we are continuing to work 4 on further analysis of the results we have from these two 5
. sequences and still working on the TPI sequence-as far as 6
-the primary system goes.
Ken has not as yet'had an o'pportunity to do 8 anything with my primary system numbers. 9 That concludes.the Grand Gulf exercise. 10
.MR . COOPER: I am sorry, Mike, just to pick up a.
1 point with' Jim. 12 If the deposition velocity is independent of I concentration, tnat doesn't sound as though it's surface 14' reaction rate-limited, though I-might be mistaken. If it L5 were reaction rate-limited, we would see the deposition 16 velocity changing with the concentration. So maybe that one needs to be looked at little bit more. 18 MR. SILBERBERG: Okay, let's proceed to the next 19 topic, wnich I believe is-now entering into the realm of 20 suppression pool analysis.- I wonder if I-could ask the 21 spea4ers, two speaxers on suppression pool, if they could 22 adjust their presentations to a two-thirds' presentation 23 and roughly one-third allowance for questions, if possible? 24 25 f ~x- TAYLOB ASSOCIATES 1625 I Street, N.W. - Suite 1004 washington, D.C. 20006 (202) 293-3950
~ ' - tf9 2e
r y.. .
=-
TA~195 , 228 iMMajli21:1 I- 'In.other words, leave about one-third of your
~
' (h ..
2 allotted time for questions, if possible. 3~ MR. VOGEL: I' note EPRI is down to discuss the 4 suppression pool. 5' MR. SILBERBERG: 'No, we have a revised agenda. We r 6 removed that. MR. POSTMA: Thank you. However,'I'_will comment. Well, we are down to some things that are perhaps a 10 little bit moreLfun to deal with, some mechanisms we can all II picture.a.little bit better. 12 (Slidd () I3 14 What I would like to do is describe the technical bases for the.SPARC code.
'ii. 15 My name is Postma. I should say my'co-workers, o
81 16
= Peter Owczarski and Kevin,'should be given a lot of the 0 17 12 credit. Pete should really be here, but his wife blessed him 18 y' ' with a newborn son just two days ago, and he decided he should i
I'
.l. be with her.
~ 20 ~
EJ .MR..SILBERBERG: Was that on a milestone schedule? MI'.y. POSTMA: I-think he was a little bit late on 3
/ / that as a niatter. of fact.
-. 23 As~an overview of our work, this SPARC codes was 24 j77 deve' lope'd as partio'fla.P&L1 program on the performance of
-%.a .m
. engineering. safety features under severe accident conditions.
- , , u., . . . . _ . .
(. I w 229 3MM jlT21:2-.
' l --
(Slide) p, N; / 2- ' This-is work being-' sponsored by the NRC. We're
~
?8- .looking at.more than just pools. We're looking at pools, at
< 4- the containment sprays, at= ice beds in the ice condensor l
.5- . plants, at 'the filte'rs ---~ you've heard' some descriptions of l
? . i
- 6. filters today -- and at the containment coolers. And the
~
- 7.. emphasis is on what.happens-in severe accident conditions.
8 Particular effort here was a high-priority one. It
- ~
8 'was"a little bit out of.ou'r origina1 schedule, but we con-l L 10 centrated on it so that we could have'a tool to be used in y ll . .this study.
~ 12
!: - I would say, because of our time limitation and
. .')jc 13 because-of'the state of'the art at this time -- this'is a l'
j . preliminary model -- we attempted to.make it as realistic as l' O
- g 15 . ,, can --
L l 0 [. 16 (Slide)
~
17
- - as we could in the time.we had. BUt it's not the
'18 final work'.-
~
.g.
= 19 '
We wanted to account.for what we thought were major
)- #
t l phenomena. First of all', in'any kind of air-cleaning system , 21 you know right'off the particle ~ size.distribut' ion is going to 1-22l
}- be..very~important. There is always-a particle-size wh'ich 23 - penetrates more than others,.and we have'to account for that.
24 Pool.temperatiure -- this was one of the important
^
j.. QJ . 26 things in the-saf'ety study.- The: saturated pools were said.to
~
4 w - * *Wa+=' , % 'H , ' g
230 MMejl 21:3 1 have a -- it was recognized at that time it was simply a p ( ., 2 simplification. So, we have to focus on saturated pools ver-3 sus sub-cooled. 4 Pool depth -- we expect that to be important, 5 because the energy that is dissipated would increase. :The 6 bubble residence time would be longer for deeper pools. We
- 7 have to account for that.
8 Finally, we wanted to account for steam condensation 9 of evaporation. Obviously, if you condense all the gas, the to DF would be very high. On the other hand, if you evaporate 11 all the waylthrough a bubble, you can expect a retarding 12 effect. u g 13 Finally, this tool we developed was to be compatible 14 with the TRAP melt and MARCH codes so we could get decontamina - 15 tion factors for specific sequences. O ! l 16 (Slide) 17 As we started this work, we thought how can we go j 18 about this. There were a couple of different ways. 19
! The first was to say, well, I'm just going to focus I
20 l on bubbles and do the scrubbing in bubbles. That's one way 21 to approach it. 5 m j 22 Then, I have comments on the method we did adopt 23 later. The second method is to look at the energy dissipated 24 per unit volume of gas. There are correlations in the scrub-("' 1 25 ber literature that approach scrubbing efficiency for
[ LMM jl 231 f( :21:4' 1~ particles lin this way. 2 A'nd - finally , if we had sufficient data base, if the 3 L- other modeling approaches worked good, it would be possible 4 to apply experimental data to this question and admit that 5 we-are st'ill developing models. j_ -(Slide) l Let me backwards now and tell you what we thought ! about these. First of all, the. data base is fairly limite'd on 10 :
- ' pool structure. We're a-little bit limited because we were' II ~
studying at a slower pace, and we also did some work for the 12 Limerick study. And we were fairly familiar with pool I3 3 scrubbing. l 14 What we can say is there's really no data on saturated pools, so we're.very limit'ed-to know how to approach
~
- 3
$ 18 saturated pools if you-just look at the data.
L 8 17 - l 43 Steam-gas ratio --.we didn't find a consistent study l 18 ' I where you could'look at the effect of' steam-gas ratio and know
'l I'
what its effect would be so you can predict it for the
.E' ' 20 j r sequences.
~
21 l- . Particle-size effects.-- it's recognized that there. 22 f,. a particle-size effect, but it's very difficult from the
' 23
' literature to identify a clear trend of particle size.
.7 Finally, there were very, very few large-scale tests.
~
One large-scale or moderate-scale test was a GE. test, and I i f b _ _ _ _ _ _ _ _ ____ _______ _ __ _ _ _ _ _ _ _ _ _ _ _ _ - . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - - _ _ . _
.~.
MMajl:21:5'
~1 . think:that was significant. And it led us to'believe that'we p. )c , ' 2 - could approach the' bubble-scrubbing approach, e
3 So, we cancluded that simply applying experimental 4' ~ i data.would lead us to some problem in justifying whatever DFs 5' we cameLup with. 6 (Slide) 7 . The second thing we looked at was this energy cor-8
~
re1ation. And this we take from Calvert, who, about 10 years ago'or so, did a study of various kinds of scrubbers, correlat+ 10. l ed a cut diameter - let me tell you about a cut diameter. II It's-the size at which particles are removed with 50 percent
' II -
L efficiency.
~ 13-
;. : .! What you assume in this. method is that all of the
{/~N , I' bigger particles are removed and all of the smaller particles ' 15 pass through. That's an approximation for sure, but it i j 16 works reasonably well-for a lot of different kinds of scrub-l 0 17 e bers. i j s 1 15 And we show cut diameter versus pressure drop.-
!e This is representative of the energy expended per unit volume i .g. y.
E of gas. And you see there were various kinds of scrubbers. F 21
;- I mention.here a region for.the SGS, and that's a L l n submerged gravel scrubber. It's a device which is the: type
.t 23 of pool scrubber under development at the Hanford Engineering 24
#+ Development-Laboratory. . Indeed, it was found -- looked at in b .n terms of'energyi-- that fit pretty well. And that is a
. x e W. .a. #e.e.m -...-py_- w- . m. ,, ,e , q g. e 4,, -.e .p .*w...g er ,
~-
I : MM jl.?21r6 - ~ 233 i-
~
1- possible approach. u c im S._,/ _
,2 And as I mentioned, we would like to retain that as
( 3 a pointy of comparison. 4
' (Slide) 5 A few comments on.this cut diameter versus energy 6
method. It's very,ivery simple.- All you have to know is the
' -7
- j. pool. depth. If you know that, you can get a cut di,ameter.
8- If you know.your particle-size distribution, you simply.'look at how many particles are larger than that, and that tells l
. 10 you. And it is supported by data obtained for submerged II gravel scrubber.
12 I should comment-that-we attempted to do mechanistic 13 J l- models-for'the submerged gravel scrubber. We looked at I various retention methods in a gravel bed. And no matter 15' how we ' looked at this, whether it was' -- like the Scrubber 1 -l 16 - Handbook approaches.this with a partially filled bed, you O
' 17 e allow for centrifugal' force around the grains.. Or if we look-3 18 E ed at immersed flow, past spheres, or if we. looked at flows
-t g 19 e in the~tristese spaces between the particles, we always 2
I-m greatly underpredicted the-actual efficiency. So, it made us I 21
' feel the energy dissipation anel the turbulance that develops
[ g~ ,
- is not-negligible. It dominates with that kind of scrubber.
23 There are some real problems though. This method did
};- not account for. steam condensation. It.did not account for k,/ 26 -
. distances in geometry, did.not account.for pool temperature, c.
- ~,-n <-+ - ~4 +e .e
MMzjl 21 7 234 4 1 nor did it account for particle growth. 2 And as I mentioned, we will keep that as a point of 3 comparison with the model we have. 4 (Slide) 5 So, we have concluded that we would go to a bubble-6 scrubbing model, and we started with the standard textbook 7 model in the Fuch's look. And he accounts for sedimentation, 8 inertial deposition, and diffusional deposition. What this J 9 neglects is what happens during the bubble formation region. 10 Consider you have a stable bubble, it rises. You follow that 11 bubble and c.se what happens. As an input, bubble size is a 12 critical inpue parameter to this model. Bubble circulation is s
) also a critical parameter. If the bubbles do not circulate, 14 then sedimentation is the main mechanism, and it's not a very 5 15 g strong one.
v. 8 16 But if we do have bubble circulation, if we do 3 17
? know the bubble size, we can model the important mechanisms.
9 18 j Let me describe those mechanisms. Ken Lee already alluded to
$ 19 $ these; I can mention them just a little more.
20 8 (Slide) E 21 g First of all, we have the condensation of steam. j 22
- And what we assume there is that the fraction of the gas that 23 condenses carries with it its own fraction of aerosol --
24
,e 50 percent of the gas condensed -- we'd say remove 50 percent \- 25 of the particles, right at the inlet of the pool. Then, we
~
MMajl.2128 235 I' Then, we allow for sedimentation, which is ordinary 2 gravity settling distributed over the cross-section of the 3 b ubble . Centrifugal deposition turns out to be a dominant 5 mechanism. The bubles that rise circulate; the surface moves 6 from the top to the the bottom of the bubble. And because of 7 that you get a centrifugal force. And this centrifugal force 8 causes particles to move radially and be deposited upon the 9 wall. 10 We account for diffusion -- and here we use the 11 penetration theory of mass transfer. You assume that the 12 gas that is close to the bubble whirl moves with it, so it () 14 has a certain residence time. You do a transient calculation; you find how much of the gas is depleted by diffusion. N 5 15 g Another factor was steam evaporation into rising 8 16 b ubble s . Fuchs points this out. He quotes some experiments 0 17 { of Wieme, some old German work, who claims that if you have R 18 g bubbling into a saturated boiling pool suddenly the particles h 19 g aren't trapped anymore. We dug out those references. It 2 20 8 turns out they were done in some rather small laboratory
; 21 ; scale equipment, about 30 centimeters or so of liquid. But ! 22 i it's .something we felt we had to account for.
23 Finally -- not finally, but another factor is the 24
,~ growth of soluble particles. As the bubble goes through the \./ 2 water, it certainly has to be -- if you have s'oluble materials,
fMM:jl.21:9' 236 s I they'11 pick lup water pressure.until the vapor pressure is pC 2- thin,'to be equalized. So, we do allow for the growth of
'3 soluble, material. Cesium hydroxide'and cesium iodide are two 4- important materials.
5; Finally; we took a first look at the entrainment of 6 poot liquid which would carry with it some amount of material 7 .that had been deposited in the pool. 8 (Slide)
' Castleman, Penn State.
MR.. CASTLEMAN: 10 i Could I.ask a question'on-this centrifugal deposi-tion. Is that a well-known phenomenon? Or is that something 12
~t hat you have-accounted for that is something new?
13
) MR. POSTMA: I was going to show you a picture,.
14
-and I can' describe it a little better. I'll discuss it a r . . ,
15 ! -f little bit.later. 3 _[ 16 Let me define " decontamination factor" in case l .,
; 17
-some don't know what it is. It's the ratio.of this mass i
I8 To most.of us in air
- f. entering the pool to the. mass leaving.
t. g 19 cleaning,'we think.about penetration, fraction of penetration. l "
-r- ' That's lthe reciprocal of the DF. .Or if you want to work in 21 l fractional efficiencies,'you have to find'the. penetration is II
! 1.minus the fractional efficiency.'
For a single particle size, we assume that the 7-c . mechanisms operate generally independently -- that is t3 say C/.' s .. the penetration overall is the penetration you would i s
._. .s ., ,. . . # s
w ,.- . .
- , . - :/.
~
MMa jl: 21:10 237 i
~1 ~'
calculate for-the product of.the penetrations for several rT
~2' removal mechanisms. But,.of course, youchave to integrate 3
over the particle size! distribution to get the total mass
~
4
'that would be removed.
(Slide)_ 6
'On this first thing, let me say what we do on deposition due to steam condensation. I think I already mentioned how we handled this. We simply put the gas in.
l 9 X sub I'is a mole fraction of noncondensibles entering. We
~ assume we have thermal equilibrium, very close to the inlet.
We expect that. BUt the GE people -- Fred Moody has done 12 some transient calculations and showed, for reasonably. sized l ) bubbles, indeed, it's only, I think, .01 seconds or so he 14 . estimates. t E 15 y So, he assumed we had thermal equilibrium in the t i. g 16 small stable bubbles very quickly-after they entered the pool. 8 37 !
? So,-you can do a mass balance and say, "I have L' 18 g ,
'hermal and mass equilibrium."
. The vapor pressure in the i ist bubble here is the same as the vapor pressure of the water in g
2 2-I the pool, So, you can just calculate the mola:. fractions. 21 j .You find DF is equal to X sub zero, mole fraction of non-I- 22 . . 2 condensibles here versus the inlet. , 23 I won't go into the simple matter of defining the 24 pressure here. If you now the vapor pressure, you can calcu-r~') L - 25
- late,due to the weight of water over the density of water,
.__,_s._._...a___-. -
7 , . - . . - . fMM jl121:11~ ,
- 1 . what'the pressure is at the point'of entry.
j, 4i 2 (Slide) 8 The other problem that is important here -- this 4 shows kind'of a horizontal entrance -- if we have this type 5 of inlet, the. gas enters and typica'11y' forms a big bubble
-6 which detaches'and breaks up rapidly.
I Then,'we move as.a' swarm instead of individual 8 3'ubbles. So, we account for that by recognizign the velocity
~
8 of the bubble is the velocity with respect.to the field 10 around it, whereas the swarm velocity, which determines the II trans'it time, is~some larger-value. 12
- l. ,
So, residence' time is the height from the shattering r . level on up, divided by the swarm velocity,
- I4 f (Slide)-
A
.j W 15 Steam evaporation velocity is important, as I I I'
' mentioned. It could be important at least, because the I .8 17' g only data that ' exists -- as -it has, in'effect, on particles --
3 18 i l-
'i: you expect an inward flux for. rising bubbles.
! I' Think about 1t 1 fo'r a minute. If you'have a warm i-r " pool, the bubble enters, it's saturated at the-beginning. But j 21 as it rises,;the-pressure decreases. The pressure of water l 2 ,
22 also decreases,.so you tend to get a by-bubble.- Then, you-
,, ~
'have' mass transfer into the bubble. So, the bubble rises,
" always has an inward flux.
g1
'D 26 We account for this with the penetration theory of
m - MMajl 21:12 2B y-1 mass transfer and heat transfer. We account for a cooling 2 of the interfacial liquid by evaporation. So, we do a delta T 3 across the liquid film. Then, we calculate it,.a concentra-4 tion difference of the steam across the gas film. We assume 5 this flux is directed normal to the surface. It turns out the 6 flux' increases with height, so we have had to divide our pool 7 into a series of height elements. And I think we used 50 8 height elements to calculate things. 8 (Slide) 10 Here's something I'11 just briefly go over. I'm not 11 sure how this was accounted for in the work that has been done 12 by Battelle. 13 J
') It turns out that the magnitude of this evaporating I4 flux depends quite sensitively on the temperature. So, we 15 ; feel it's important to know what-the temperature of the pool 2
l 16 is. At first glance, you might say the pool -- a 18 g saturated pool is at,the normal boiling point. But as we 18 E pass noncondensibles through the pool and we've made a 20 calculation here; assuming that the gas is leaving on a 21 saturate, if you do that, you see, you get a fair depression 22 {' in the pool temperature, at least an equilibrium. This is 23 equivalent to an adiabatic saturation temperature, except we 24 7 have allowed for some decay heat. ( ..' 25 So, wnat we did in SPARC is we. allow either an
~
g u a. -- -
. -. - - . . . . y - . - . . - . .
MMajlt121:13 240
+-
L 1 c l' JequiIibrium temperature-by a simple valve balance or-user-to-D1 2 x;/ : input temperature of.the pool based on a more global look at 3 the.thihg, perhaps-from MARCH. f. 4 ! (Slide) L 'y 5 Particle growth by water sorption -- this'is kind 6' .of.a. standard ~ calculation'from applied physics. Soluble
'I L
particles will grow if the relative. humidity is greater than l 8 : a critical value, and this . depends on the solubility.- of the
. . ma te rial'.
10 t l . Essentially, what you-find is,.at equilibrium, the i
. 11 L vapor pressure in the solution valve is equal to the water 12
-vapor pressure in the bubble.
l 13' Using Fletcher's formula from the physics of rain ( l I'
' clouds,wedidsomeexamplecalculations.JIt'showsasa i 15 function of he dry particle radius cover, the drop grows.
8 16 i = I guess the important thing'is humidity of 99; percent. You I 3 17 e get a growth! factor of roughly;4. So, this: calculation is 1
~*
j a allowed for in SPARC. So, we do allow soluble particles to f
'[
-x I'
' grow.-
r 1. E- 20 I And I think one'of'the assumptions that fairly
.j 21
~important-in SPARC is that we,do assign.a humidity'of
- 5. ,
22 - l i
- l. 99' percent.. Really, we wanted to calculate that as a' func-23 -
tion of height in everything that was going .on,' but we had l I -
.some difficulties in getting-convergence.of the program.
~
/ 2 That's one of,the things ~we're working on.
1
, . - .x MM a j l E 21 s l4.. 241-it' -go, we did' kind of-back-of-the-envelope calculation N, 2 - fori the region near.the' inlet, and'_we estimated 99 percent 3:
- wasn't too bad for'the saturation level.
4 (Slide) 5: . Bubble shape is an input to'the model. First of ( 6
.all, the Fuchs model was based on' spheres.- You can see the 7
interface moves around and causes centrifugal deposition, 8 gravity settling to the bottom in' diffusion. If you have a distorted' shape and it turns out bubbles really are distorted,
~
l 10 you ca'n get spherical-bubbles if you have small bubbles, 11
.below about .1 centimeters. As they get bigger than that, l.
12 l= 'they get more and more oblate spheroids. So, we allow that 13 ~ A'over B ratio to be an input parameter.
)Iend21 14 5 .15 i C r
i .[_ 16 i 3 - 17 0 ? I. 18
- 1
.g 19 E
20 f. 21: { s
.l' 'N L m 24' y
g
# 26
- u. . .. _ , _ . - - _
='
- R l
~
(! ' 242. p" w , , e . '~
.. 31- 'MR. POSTMA: .(Slide) o , 3*M/
~
= fu . mml\T22- 2 .-
gj; L -
- Let'me, with this picture, describe the mechanisms.
'3- -
- I'didn'tlput'down all of the equations that we had. I r
F 4 didn't think'we had time for that. i ' 5 Let.me.show you the centrifugal sedimentation and
~
6 Ldiffusion.- What we said, we have gravity acting downward
- 7 .- on'the' particle giving.'it a downward velocity. We have a this evaporation velocity acting normally'to the surface.
8^ So we simply resolved those two. It is a deposition are, 10 .' deposition velocity times.an' area. It is an integrate over
. j.'
11 .the. bottom part of the~ sphere.. 12 Realistically, you would-like to look at an hV 13 - 'iEdividual particle in this boundary laye'r and simply p 14 ' resolve the forcesion it to see how it. acts.-I don't l' : 3- 15 think they would be'very much different.= For one-thing, 33
-] 16 - this gravitation is a minor factor. . It is.less than.20 g.
~l 17 percent.of the: total DF in:any case,Lso-we don't think this i
is an important thing as~far as gravity sediment is concerned. 18 t' 18 If'you simply make a material balanceion the: gas phase in-I l l j- - 2D the bubble,;you find-it decreases with time. It is an
*i; I ' 21 ' exponential; type of thing.. So the decontamination' factor is L 22 three' halves,:it is D sub S, which the. settling velocity,
-l-
;23: : subtract-off.the, steam. evaporation velocity, divide by L
- 24 : bubble.diameterIby..the timing, which is the time-for the afn.
? , -Q' c ,, 25 swarm. F i l---. ,,.-3. +- . . -. -
[ i-I 243 I L - mm2 1~ Let me describe how the others are calculated. . r~s l'1 2 for centrifugal force, you simply take a regular M V squared I 3- . over R, a force on a particle undergoing centrifugal force. I ' 4 .You equate-ihat to-the Stokes law drag, let that move the par-5 -ticle radially. 6 The real question comes, okay, how do you L7 calculate the velocity? We did two different-things. The 8 rathod that Fuchs uses is to1 assume potential flow around-9 the sphere, and there the. velocity varies with angle. You 10 get a little angle,? differential angle. It gives you an 11 -- incremental area. You simply integrate like you do here, 12 velocity times an area over the whole angle of the part of 13 L .the bubble. 14 MR. REYNOLDSi I didn't understand that. I 5 15 tlought you were rotating the whole bubble. I thought your y . . [. 16- -centrifugal,.the whole bubble was rotating. 1 f 17 MR. POSTMA: ~ No. If you imagine the bubble being i
-18 stationary, fluid flowing:down, splits on both sides.
,E t 19 MR..REYNOLDS: It is uniform?
E- _l 2 MR. POSTMA: Goes around like that. When you say f- 21 ,is it uniform, I'm not.quite sure what you mean. 3.
} H. MR.REYNOLDS: It is symmetrical?
23 MR.:POSTMA: It is axi-symmetrical. If-I put 24 ' "an-axis here, it would flow down in all directions. (;g ..: t--
. p< #
2 MR.REYNOLDS: Where does the centrifugal force s .. i~. 41 -
-.wr- w -v -
e
- .a - - . .. - -
'.i 234
- mm3 1- come in?
.[ -
2 MR. POSTMA: If you are a little guy, a particle 3'~ -riding:along the outside, wh'at you feel is-this curvature 4 effect. 5: -MR.'REYNOLDS: It is right on the edge? It is not
~
'6 fon the.;inside?
7 MR. POSTMA: We assume the interior of the bubble
.8 is well mixed. That is another assumption we had to make.
~ 9' That is the one Fuchs makes,the internal part of the bubble L10 . isjassume'd to be well mixed._ It has only got a little layer 11 .near'the edge that we really care about.
12 MR. CASTLEMAN: I thought you were talking about
- 3. .
'~13 - the inside.
jf 14 MR.POSTMA: If you had a discus bubble like two j -; 15 . liquids,.I think you would have to worry about that. It
} 16 would be a'more. complex problem.
O u 17. MR. CASTLEMAN: With.the gas in there I couldn't
'R 3
18 see that'was happening.
- (-
h ' 19 MR. COOPER: This is very similar to the collection 3 j 20 of particles in a curved' spike. You can do that same kind {- 21 : .of analysis either by using:as you did, the business of c 5. 22 ' centrifugal force and a drag -- coefficient' drag force. Or,
]
23 .you!can. simply use an-impaction parameter. It turns out to be , 24 'an. inertial parameter, it turns'out you get the same s~,
$~~: ' M, 26 . dimension.
y[ &-'#*94e6ys' Mg "J b' 4 W W-MW N hkd OOMa~ $dh 4A h + et t *
$ p } W- 4
Q,f - ** % -n 235 l o i I
,i
- mm4 If LIffyou do a very quick impaction parameter p,
3([ , 2 Jealculation using.the characteristic velocity which has got 3 to be the rise of the bubbles through the.liauid, _if you 4J use the-characteristic time'for the particles, which I grant 5 you 'is1 somewhat ; larger because. of the hydroscopic growth, 6- if you use in the~ denominator the diameter of that bubble,
'7 you'should'have very much the right. order of. magnitude that 8 you~ will ge't ' from this - very elaborate calculation.
8 The problem'is when I picture those numbers in 10 myLmind, that upward moving bubble,the size of the particles 1 11 being collected and.the diameter of the bubble, I get 12 impaction. parameters that are terribly small. (D W. 13 So that;my initial response is to be very surprised 14 .if this should,-in fact, turn out to be a significant 15 collection mechanism. 3
.] 161 I have done a few calculations with bubbles going
. 17 - up through liquids. . Typically .dif fusion and sedimentation.-
j 18 have dominated, and this kind of mechanism has not been 2 g 19 important.- I could be wrong. But I cal'1 your attention to A 9 -
' 'i it, to check.
- 21. MR. POSTMA: We looked at that very carefully. I 5
l 8' Lwould'have-to say.you should'go back and really do the 23 integration. .
'M If you just look at the Stokes number it turns I .
'ALC 25 out bubble size would just be to the first power
~
I think vs ,==p.me
%- ,f-w.
-a.*_-em q m , -*y
m - - c... g amS~ 1. .the denominator. But when you do-this calculation, as I 9p ~
'recal1,;you get a'D squared.down there. So~it is a lot
- Ai[ - 2
- .more important'for the small bubbles, first thing by just 14 . : lookingfat;that? Stokes. parameter.-
~ MR. COOPER: 'Whenever one is surprised, one doesn't
'6 know whether one is surprised because the truth is 7 surprising,: or-because a mistake-has been made. Yes.
8 - (Laughter) 9F : MR. POSTMA: I must admit,.I was a-little surprised
- 10 too'when I.first-did it, toffind that-this mechanism-11- . totally -- I.should say'for.' particles that have inertia, 12 it is i common for particles-_with small size diffusion
~
. - 13 ' andlthis has very little effect.,
- 14 - (Slide) 5 - 15' Let me hurry;along',-let me show-you some examples
;[ 16 of ~ calculations that we have made with this code. There is f
5 17 one more. mechanism _that Ken alluded to..that-I haven't even
*: 18 described. .That is this impingement, as you enter.~the pool.
-t
~i g _ 19 - -I think:Battelle Columbus hasladded that to'the SPARC model~
lb - m- and accounts for it.
.It is modeled very much like a cascade
- j. 21 impactor.where you have a gas jet impinging.upon a liquid g-3 c '.
- 22 . surface.-
- n We didn't put-that in. We: thought about it, but I
. M feltra little uncertainfabout theigeometry of bubbles n
ii
. entering;'whether-that could.be simply. approximated )y a
~
. '? ' 25"-
s
- - w. e m - w- s ; ,i n. s ,a.w^ . c w -
, - y .,
r -- 1 1 237 l l vn6 -1 plane.. Plus I also felt since this inertial thing was
. rTh
(,' 2 so powerful for small particles, that if you would remove 3 them by impaction you would already remove them by the
.4 circulation factor, and it probably wouldn't change the 5 result.
6 That could be wrong if you had a very shallow pool. 7 MR. ROE: I am a little confused as to where the 8 particles are, where the bubbles are, and what, in fact, the 9 removal process is. What is being removed? Where are the 10 particles going? 11 MR. POSTMA: Let me go back and show you the 12 picture. 13 MR. ROE: I thought I was following quite well, 14 then I got lost. n 15 MR. POSTMA:' (Slide) 3 l 16 - Here is the first assumption,the bulk of the bubble
" 17 is well mixed. Will you buy that for the moment? Assume 9
% 18 that.
5 i 19 If I start right at the topiou are in the gas flow, 5 l- M' okay, and water is flowing down like this because the bubble { 21 is rising. - Okay. We assume the interface on the gas side 3-22 is moving at the same velocity at the interface on the liquid l 23 side. So the gas is being dragged. So the particle now j,
- 24 finds itself,in the flow field within the gas, so it is 25 being projected outward. And they hit the wall. And we
. . - . - - ~ . -- - . .
238
- mm7l 1
. assume when they hit the wall they are gone.
7 l.; 2 MR. ROE: You have particles in the gas phase? 3 MR.rPOSTMA: .In the gas phase, that's right.
- 4 MR. ROE: That's what I was missing. Particles 5
in the gas phase. . They find . themselves then reaching inter-6
. face of the bubble. Once they hit that because.of the
.7 centrrfugal orce, they are now removed to the pool.
/8 MR. POSTMA: -We' assume the surface tension 8
effects cause the particle to be captured in that pool. 10 MR.-COBBLE: Where'the particles finally end up? 11 At the bottom of the pool or the top?' 12 MR. POSTMA: They end up in the liquid phase. 13 ' This is[just one bubble =now in a huge pool. Remember 14 t'here are some other bubbles about. So, the particles hit 15 the' surface. We assume they are wet by the liquid, they 0 l 16 simply'get carried down and they' mix in all the liquid. 17 ~ MR. COBBLE: What was the-particle size? Are we 18
.using uniform particle size?.
18 E; MR. POSTMA: As far as the model we did particle
- 7:
20 size as a parameter, so we-predict for all particle sizes, f
- i .
21-MR. COBBLE: What'I am worried about, there is a
- l 22 whole industry you know, where you' float very fine particles 23' to the. top by bubbles,lnot the bottom.
24
- -. MR. POSTMA: I guess I am not worried about where 25
.they are, as long as they are in the water at this point.
i __.-__-____e -
3 1
, . # 239-r
.mm8 -
'1-
,n, MR.' COBBLE: If they form a foam on the top, then G 2 ~
-it is kind of a question whether-they are in the water or 3
.not.
4 MR.- COOPER: They are not in the air. 5
~MR. POSTMA: ' Well, I think that is perhaps a good 6
question.- I didn't mention it, but our method of accounting 7- for.is based.on the' experiments done,~where they~actually I measure entrainment rates for-the pool. We somewhat 8 arbitrarily put a lot of material, like half of the core 10 ~
-inventory into the pool and then calculated for that
- 11 volumetric rate what:would-that be equivalent to. It turned
~ out.to be 10 to the 5th, so we limit our DFs.to 10 to the 5th.
~
13
. That-is a question I think ought to be studied a 14 little bit more. As a matter of fact, peter Owczarski has 15 Particular t searche s ,
ordered some literature searches on that. 3-
.[ .16 And that is, do-bubbles really circulate like we'say they do.
II We know if we have very small bubbles-and you have
} g 18 surface active agents, they rise as' solid spheres. That is 18 '
j well known.
. 20 --
The question is, for the bubble-sizes we are 21 We don't think it is, but j speaking-about, is thatimportant? 3 we still want to look at that some more. 23 ' fMR. PL1' RANGELI: .It is.rather surprising also to 24 ,,.that this mechanism is very important. What I think is f,
'.k) _
25 that the bubble essentially pulsates in the liquid, and c-L
< . . * . - m. # .. .
.w g ve . - - -
240 mm9 I this movement on the surface should be more important than in
'. 2 the entrainment inside the bubble due to the motion or 3 the relative liquid.
4 Did you consider these in some way, this kind of 5 -pulsation of bubbles change their shape and things like that, 6 with the size of bubbles you are considering? 7 MR. POSTMA: I certainly agree with your perception. 8 when we look at bubbles in a pool we know they are vibrating. 9 They have a way of moving back and forth because of fluid 10 instabilities. 11 You might think that would be important, maybe not 12 as important. But we haven't found a way to model that. That C) 13 is what I wouldhave to say, that is part of the energy Qi: 14 dissipation we are not accounting for. 15 In scrubbing, again, when you have' MR. COOPER: O l 16 a droplet in a gas rather than a bubble in a liquid, often 17 you find that the thermophoretic forces pretty much offset! 18 the diffusiophoretic forces. And I think if you incorporate i 8 18 thermophoresis into these bubbles you are going to find it n 20 l goes in the opposite direction from that diffusing vapor 21 force. 'And it again deserves at least to be checked, because 5 3 3 [ 22 we have got some scrubber results that show that even for 23 an evaporating drop the thermophoretic force is greater than 24 the diffusiophoretic force and you are rather than 25 collecting particles because of that, you are not collecting
~ - ~ * ~ ~ ~^
n
;.1 ;w - , : _; - -
1-q 241 mml0i ;t- fthem'. -It is' going theioppositelway, p. 1J 21 Slim and~Halds.have pointed-this out over a decade 3' ;ago, and : June Stuko atJ Illinois has demonstrated it.
'4 ,MR. .POSTMA: -I agree. Thermophoresis we-haven't 5- considered. A couple of reasons for that. one, we didn't T6 have? time. The'other is:thermophoresis would:mostly occur
-during the very initial'part of the bubble formation.
~
7 8 _MR. . COOPER: It occurs wherever.you have the 9 condensation. 10 MR. POSTMA:' It turns out for most ofLthe interesting 11 cases you have a lot of. evaporation' going on right at that 12 time. So it would probably be offset by-the large amount
~
.1 ,
13 -of evaporation that takes account of that. But'we need to 14 analyze it.- 5 15 MR. COOPER: I am referring to the latent heat.
-I .
.] 16 _ MR. POSTMA: We'should add that.
3 '17' .(Slide) e t
. 18 . I think the key thing.I want to' add here -- here r
i ~ g 19 I am showing you'the access ratio -- show'you the bubble
.g-20 shape is often'important here. These curves, you see the
[
'iJ 21: elevation is tremendous. The reason is, as you go around
's . .
-0~ 22 ' that' sharp bend on that oblate . spheroid, you have much-1 ,
23 higher centrifugal-forces and causes-that centrifugal 24 > > mechanism to be very .muc h more 'important. 25 .I think the most important thing is this model
,4/ A.,, .,. d, w .5 ..4.,.,. . - . - _ . , ~ . 3 ._
242 mmll. t predicts - -this is no soluble particles of cooled pool, 2 12-feet, 20 degrees Centigrade, and we get a fairly low 3 DF here'for this intermediate size range, even if we take
,4- this elliptical bubble.
5 MR.-VOGEL: .Is this 100 percent noncondensible?
'6 MR. POSTMA: Yes, it was. It'is one of the worst d
-7 - cases-you could do.
8 MR. VOGEL: It looks pretty bad. 9' MR.-POSTMA: lIt all depends on your . particle size. 10 If you get particles of any-decent size you have an elliptical 11 bubble. You see this is still a-half micron particle. .You 12 ~are a pretty good one. , 13 But, if you are.really loaded with particle .l or 14 .2, this model. predicts you are not going to get very high
! 15 DFs and that is what limited the DFs in some of the cases 3
l 16 shown. o MR. PETRANGELI: Did.you also consider the tendency l 17 i
*- 18 of the vapor to concentrate also in'the middle of the bubble 5-
')- is .due to the centrifugal effect; or, did you consider the 5
j 20 concentration of the particles uniformly within your sphere? 1 21 Because you htve both phenomena . if you consider the j 22 centrifugal. Because you.have an entraintment of gas on the
; 23 surface ~,but then you have an internal circulation. You 24 have also a concentration in the middle of the bubble, I r?. .
25 think.
,_w-___.,._.___,......,...._. _
243 mm12 1 MR. POSTMA: You are saying it would be more
, ~ .
2 concentrated in the middle? 3 MR. PETRANGELI: Yes, of course. 4 You have to have some vortex inside the bubble. So, 5 you have centrifugal forces towards the outside, and your 6 centrifugal forces are towards the middle. So your 7 concentration is not uniform in the bubble. 8 MR. POSTMA: We didn' t consider that would be the 9 effect.' 10 MR. PETRANGELI: You have to be consistent. If you i 11 see the bubble stubionary and the fluid is moving around it, 12 within the bubble you have movement, you have to have some 13 circulation. And these things occur to send the particles 14 towards the outside, but also towards *,ne inside.
$ 15 So you don't have any uniform concentration any 0
[' 16 more. i
" 17 MR. POSTMA: That's an interesting point, we 1
l 18 haven't considered that. I don't quite know how to calculate 5 h 19 that. I guess my own feeling was that as this bubble i l lm oscillates going up, with the one effect you mentioned that i 21 I would think would cause it to be sturdy inside. We have i j 22 not accounted for thati 'f there is such a factor. 23 MR. ROE: Is surface tension and wetting at the 24 surface of the bubble an important issue?
/ 25 Do these particles -- do they wet readily?
-. . , ~
7_ , g; o . 1 i e
;g .
244-
^
( ~ (mm13' - 1; MR. POSTMA:--IIthink that;is-a question-some'have
$K J 2; T- :1 raised. - My._own'. experience with scrubbers. leads me.to
<- ,37 (believe~that'for most-substances I'm aware of the perfect p
- j. . 4 .-- ' assumption is a prettyJgoodTone.. - The particles hit the L
L5 1 surface,they are: captured; they-don't'just stay-on the - l , .. . l so; ? surface- .without. wetting. It could;be there are come cases s~' fl w~ ., _- 7} with' unusual materials where that'would. happen. But, I am ze; .notLaware ofiany'such data. L l s'- .DoLyou have al comment _on it?
- 110; -MR. ROE: No.='You-have a strange mix of everything
- 111 . coming'out of' containment, and who-knows whatLweird things 1
12 ; _it might be. l L 13 . MR. COOPER: THat.-has.-been confirmed by experiment
}
; quite L a few times.1 hey. don' t- have to be : hydroscopic or -
~
, 14 -
1 - 15 : wetted.t'o stick if theyJare.a few microns or-smaller, and _ L:- 3; - l 14 ' ?yot. have such low velocities here thati there is no . likelihood
,.o-3 4L .- 17 fany. rebound would occur.
- i. 1I .18 < MR.'POSTMA': 'Let-me mention key' inputs to our model:
l .V l- _j :19 (Slide)-
'L 20 ; . Particle size, distribution, poolLtemperature, inlet l=
.. j. 21 fgas temperature'if you'use an equilibrium-temperature model, c5
;[
22: 1theLpressure-above the pool',-the-effect~of scrubbin, height - [ 23 : !has:toibetan input,-the average bubble' diameter and the
- :information.we have so far'is basedLon the GE experiments.
t, A,
- M >
'26 1 , .-Before GE did$thei'r; experiments,.I mus.t say I was uy
- ~_. %- 7 e
.; _. _ un . . ' nt an .-- ,. , . - - . .,
245
. mml4 : t quite uncertain as to how quickly bubbles would break up
[ 2 and what their size would be. But with fairly large 3 experiments that were done by GE, I have been convinced that 4 bubbles break up rather quickly, that they break in fairly 5 small individual sizes. We typically use .5 centimeter. I 6 n ticed Battelle result was based on .75. It was an 7 elliptical bubble, probably a good estimate. 8 Steam gas ratio to determine whether you will have 9 initial condensation, to And we also allow for viscosity calculation 11 purposes, the makeup of the gas. We consider hydrogen, water, 12 carbon dioxide and carbon monoxide. 13 We keep track of those four and use the standard
- 1 14 Wilkin formula for calculating viscosity.
j 15 We mention a few key improvements we feel in evaluations that ought to be done. { 16 8 17 (Slide) a i g gg Whether we need to predict within the code the I j 19 bubble size distribution so that the user doesn't have to e a be really smart to know what to pick. There are some data i 21 available. We can do some modeling of that. i n Entrance impingement effect. BCL has already 23 put that in. We also desperately need to compare with other 24 codes and models. EPRI has a model which we would like to s/ 25 do some comparisons with.
- p. -
246
- mm15 ' l- I know that is on Jim's agenda.
- C t
2-GE has a model and I have talked with them. I 3 think'we can make some comparisons with their model. 4 Obviously, we need to compare with eicperiments. 5 We don't have experiments we can really compare. 'There is 6.
'a very' substantial program going on in Battelle Columbus I
under the EPRI support, which is really going to supply 8 some good data. This model might be totally revised based
-I-on that experimental data.
10 We need a more realistic model for entrainment that
~II accounts for carryover, and perhaps they have some 12 experiments as Dr. Cobble mentioned, which show other materialt 13 tend to accumulate at the interface.
14 Finally, we need to know whether bubbles really 15 - circulate the way we assume they do.
- g. 16 very quickly, when you summarize what we have done, I
3 we developed the SPARC code in a fairly short time period. I (Slide) I' l f, It has a number of assumptions in it. We think we have accounted for the dominant depletion processes, the , 21 predicted DF is very dependent upon the particle size. { I 22 l think the bottom'line I have come to is the prediction of 23 DFs is never going to be any better than the prediction of
- particle size.
g, Obviously, the overall DF will depend very U 26 much on this particle size distribution, and we are continuing
~ 1
~. , ,
247 mm16 1 work on it and we hope to-improve it to be more realistic. ,~.,
, 2 Any other questions?
3 HR. SILBERBERG: Thank you, ARlen. 4 MS. HANKINS: I am Debra Hankins from General 5 Electric. Most of what I was going to cover Arlen already 6 has, so I will try not to be repetitive. 7 (Slide) 8 Basically we got into the pool scrubbing test 9 program because of the fact that in the previous work, 10 saturated pools in particular were given no credit for 11 scrubbing. We felt that was a tremendous conservatism in 12 terms of fission product retention. So, we set about in a 13 combination analysis / test program to verify pool scrubbing 14 for saturated pool conditions. 15 Basically this program involved develoment of the 3 l 16 firstprinciples model which you will find is very similar 0 l 17 to the one Arlen just described, generating sufficient test 1
# 18 data to verify that model and calculate suppression pool 19 DFs for the plant conditions under severe accidents, f,
i a 20 I think Rich Denning did a very good job this 21 morning of describing the pathways in the Mark III, and 5 5 22 basically'to refresh your memory we find for the standard l 23 plant Mark III design we do in fact fail in the knuckle region, 1 24< It is a very point, maintains the drywell suppression pools. (s 25 our relief is through two safe relief valves and g' then transits through the horizontal vents in case of pipo l
248 9'17 t[ ' breaks or releases after vessel failure.
~g~
-( (Q L' 2 . (Slide)- -
3 Then we tendedito concentrate on the saturated' pool ; l- 4 condition CAlthough if you look at the probability of 3 -various sequences for a Mark III plant, you find majority i .s_ of events actually involve subcooled pools. But, because-
-7 credit was not' being given for. saturated pool scrubbing, ;
f- s we concentrated,our test program on a saturated condition.
~s (Slide) l l
10 I've already mentioned the major elements of that program. 11 - First of all to develop a first principal model, i; I- . ! 12 we'conductes single bubble mass transfer scrubbing tests,
" hydrodynamic testing to characterize the bubble. shape, rise 13 14 . velocity and size, and finally calculating suppression pool t . !
-DFs.
16 L j$ . -
.g _14 (Slide) fI 17 The model was a combination of mass transfer ,-
L ! la of particulate scrubbing, and'of' course the hydrodynamics I h c is that characterize the bubble itself. , ( .f so (Slide)
..~-
21 Arlen has just-gone through the major mass 3 ' l l ' 22 - transfer processes, sedimentation, inertial deposition, that 23 is the centrifugal force that Arlen was just. talking about. r. 34 And Brownian diffusion. (, 26 (Slide) l ! cnd mm jl'fis
P
~
^ Ilc mm1b 249
'MM:jl 22:li L 1
.As_Arlen indicated, the DFs are. calculated, the
.h .
t 't~
. mass transfer coefficients are calculated for a particle size 3' .and integrated over particle size' diffusion to get the total 4 DF.
5 (Slide) 6 In terms of mass transfer testing, we used dry _ air 7 as the carrier gas in a sub-cooled pool. So, we carried out 8 l these tests at isothermal' conditions. Therefore, we were 8 isimulating the saturated pool, where you had no condensation 10 of the steam. But we have since realized -- and I think it 11 l is pretty true throughout the industry -- that there is no 12 ' such thing as a saturated pool at the discharge location. You 13 always have a height of water; and as such, you're only 14 thermally saturated at the surface of'the pool. i 18 Since many of'these cases we are talking about,13 to E .! I 18 18 feet of pool depth,could be quite substantially sub-cooled 0 l 17 at the discharge location. And you do, in' fact, .get steam 1 18 condensation. I-18
!- (Slide)
However, again, I can only state that's a major
}3 21 conservatism in our raodel in that we don't take into account 22
{ the steam condensation. 23 The single-bubble scrubbing experiments that we did, 24
, we took europium oxide, suspending it in the nitrogen or 88 compressed air carrier gas, injected single well-spaced bubblos
,. _ . . . ~ . . . . . . . . - _ . ~ . . . . , . -
it>u j i 22:2 250 1 into a column of water. The column is one foot by one foot 0 2 by six feet. We can vary height. We can vary bubble size. 3 And as a result of varying bubble size, we saw different 4 bubble shapes. 5 I'll be talking a little bit about what kind of 6 a difference that makes. 7 We measured the inlet stream with an impact sampler 8 to get the particle size distribution on the inlet. We also 9 measd:;el the outlet stream with an impact sampler to measure 10 the 3utlet particle distribution. This allowed us to get a 11 DF as a function of particle size. As Arlen indicated, that , 12 particle size is by far the most important parameter. ('?g 13 We varied bubble size from about .4 to 1.4 conti-
\>
14 meters. That is about the range you would expect for stable
~
@ 15 bubbles rising in a large suppression pool af ter the bubbles
! 16 have gone through an initial breakup period.
! 17 We varied particle concentrations, although we i
# 18 kept them low to avoid agglomeration effects. We were testing I
i 19 three dominant mass transfer processes. We varied submergence i N height. As I mentioned, tests were done at isothermal condi-e 21 tions, although we did do one test at elevated temperatures { 5 22 to see if there was any effect. l 23 We varied the particle size distribution over about 24 .05 to 3 micron, actual diameters. V 25 (slide)
251 MM:jl 22:3 1 As one would expect, the decontamination factor is 2 very sensitive to particle size. This is a typical single-3 bubble scrubbing run, where these are the actual DFs measured 4 for the eruopium oxide as a function. Again, this is true 5 particle size. This,is a theoretical prediction, using a 6 spherical bubble bottle. 7 I think we all tend to agree that the minittum of 8 the curve,.no matter what you do, seems to come about to 8
.1 to .3 microns in size.
10 (Slide) 11 If could find a way to get Battelle-Columbus to la up their particle size, it would make life easier.
. 13
; Actually, I think what happened is we presented them 14 with our curve and showed them a minimum. Tbey then went 5 15 about calculating particle sizes.
l 16 (Laughter.) 17 As I mentioned, we also determined DF as a function 1
? 18 of scrubbing height. As one would expect, there's an I
19 h exponential dependence on scrubbing height for distribution i 20 l of particles. j 21 One thing we observed was as you vary bubble size z 22 l -- and that's what these different symbols indicate, are the 23 different bubble sizes -- we saw an order of magnitude 24 difference in the scrubbing efficiency. And this troubled us 25 because it turned out the larger bubbles were giving better
MM:jl 22:4 252 1 scrubbing. ,e i 2 Now, if you use the classical spherical models of 3 Fuchs, as you go to a large bubble you get poorer scrubbing. 4 So, what we did is we did a closer examination of 5 ~ exactly what the bubbles were. And what we found were as the 6 bubble deviates in size from about a .3 centimeter equivalent 7 volume bubble, the bubbles tend to flatten. They are no 8 longer spherical. They become elliptical. They move into a 8 transition of more of a spherical cap. They become almost to lenticular in shape, as you go from about a 1 centimeter to 11 1.4 centimeter bubble. The look a lot likd frisbees rising 12 through the water. P; 18 Many of you have seen the films we have of these \.j - 14 bubbles. They are so far from a spherical bubble that there 5 15 really is no comparison. l 16 As you would expect, once you flatten this bubble, o l 17 you shorten the distance that the particles have to travel. 1 I 18 So, you would expect better scrubbing. And, in fact, that's 18
! what elliptical bubble models do predict, better scrubbing.
i a 20 The lines on here are for the spherical model, and 21 { you can see the trend of the data is in exact reversal of the 3 22 l trend of the spherical prediction. 23
!!ere they go to larger sizes, get poorer scrubbing.
24 7- Fractions of data shows the reverse trend. k# 25 So, that prompted us to begin modeling using an
, b +e a %
MM;jl 2215 253 1 elliptical model. 7 i 2 (Slide) 3 I'm just going to go a little further into the 4 hydrodynamics. 5 We were very concerned, especially in the case of 6 the standard plant design, where we had very large, 27-inch, 7 horizontal dents, that we might actually get 27-inch bubbles. 8 We didn't think we'd get very good scrubbing with those 9 bubble s . 10 So, we conducted a test program to determine for 11 various flow rates, various discharge configuration, what 12 would be the bubble distribution, the final stable size of the _ bubble, how long would it take it to break up, what would be (} 13 14 the rise velocity of the squarm of bubbles, and how could we
$ 15 characterize that bubble distribution as a function of height 6
[ 16 in the pool. C l 17 (Slide) 1
# 18 We found out there's no such thing as a large stable i 19 bubble on the mass flow rates one would predict for severe 20 accident conditions.
{ 21 If you take a horizontal vent, initially you can s 22 grow a large bubble, As soon as that bubble is released from l 23 the charging source, the bottom catches up with the top. The 24 bubble shatters into a multitude of very tiny -- about half-t 25 centimeter bubbles.
. R A 5f 255 MR. HOLTZCLAW: My name is Kevin Holtzclaw. Debra 2
'k: . brought me along because 'she is a little bit uncertain of
'3 I presenting this portion of the model which Fred Moody, our 4
nydrodynamics expert, normally covers. I have to admit-I am
. 5 a very poor substitute for Fred Moody, so don't ask me any 0
. tough questions. I can go a little bit tnrough what we saw on the hydrodynamics portion of our test program and how we i actually went back and tried to do some modeling to predict 8
.the things tnat we saw.-
10 (Slide) U And it's contained on the next couple of charts ' 12 l nere. i 13 The first thing in our observation, we coserved 14
- tne growth of enese large bubbles, their release, ana 10 ultimately theirl breakup and tnen their resultant sizes of ,
16 the small oubbles that we are actually scrubbing the 17 fission products froia. ' 18 Another key element is identification of the 19 velocity or the transit time that the bubbles existed 20 witnin the pool in order to allow the scrubbing of the U particles from the bubbles. 22 In order to model some of these things, we first 23 of all wanted to be able to identify potential breakup 24 mechanisms. That is why these large bubbles were breaking up into the small bubbles. If we could go back into the TAYLOE ASSOCIATES '
'(N 1625 I Street, N.W. - Suite 1004 Washing':on, D.C. 20006 (202) 293-3950 L
256 literature and find reasons for this, it would give us a lot better feel for the things that we're seeing. 3 Also, it would give us some capability of 4 predicting the ultimate bubble size and the velocity or 5 transit time, which would then allow the application of our 6 single bubble scrubbing test results. (Slide) 0 This is one chart that just briefly identifies 9 some of the mechanisms that Fred has considered in our 10 evaluation, one of the things was just being able to 11 identify tnis bubble shape and the fact that the bubbles
.12 themselves are not spheres as utilizing the Fuchs-type 13 analysis.
14 In actuality, you can do some very simple
> 15 analytical work in potential flow to identify the fact that 16 you will experience this bubble flattening process. In 17 fact, if you look at just a bubble in a flow ficid and you 18 can identify the stagnation pc;.nt and you can quickly see 19 that you need to have the velocity of the trailing portion U
of the bubble moving actually faster than the velocity of 21 the leading edge, which ends up giving you the clipsoidal 22 butble shape or, in fact, even a lenticular bubble shape 23 that we see in our actual teste. 24 Dr. Moody nas gone through a good deal of work in trying to identify different bubble breakup mechanisms,
.~
TAYLOE ASSOCIATES 1625 I street, N.W. - Suite 1004 Wasnington, D.C. 20u06 (202) 293-3950
257
'I ~
and he's actually applied some of his own names to these. 2 g,,s'got an aerodynamic type of mechanism where you 3 actually have a bubble which is acting essentially line an 4 airfoil-in a flow field. We have lift forces exceeding the
. 5' surface tension forces..
6 This then identifies the minimum bubble size 7
'under this kind of a mechanism. So what he's done is gone ,
'8 -
through and defined a number of different mechanisms to try L I to identify what you would ultimately believe to be the 10 ! i stable bubble size. -This ends'up giving us a stable bubble size approximately the same as that which we're seeing in 3 12 the test program, something on; the order of half a 13 centimeter. <
' 14 He's also looked at(a number of other more
' 15 l
classical breakup mechanisms, things like fielmholtz-Taylor 10 stabilities and items like this to again try to' identify 17 wnat you would ultimately end up with in stable bubble 14' ,g,,,, + 19 l He has done some work in incorporating this into 20 an energy analysis where you can equate the draft for I 21 l surface tension forces and combine all-the bubble breakup 22 t processes that he's come up with!in order to balance out l 23 this energy. force and identify the sizes and bubble 24 y,g,,ggg,,, , 25 (Slide)- : t
; TAYLOE ASSOCIATES 1625 I Street, N.W. - Suke 1004 -
; Washington, D.C. 20006 '
(202) 293-3950 o ae. m .a ,s m. , s+. w a v e- - - awi s l
258 l'
', In order to try to get some feel for how good 2
the analysis.is relative to things that we saw in the '
~
tests, we have compared two special literature cases that 4 is usually for very, very small bubble cases or very large 5 l cases, and gotten some relatively. good comparison with our
' calculations.
We nave also made some comparisons to the 8 testing that'Debra mentioned earlier. We have shown some of , I i the parameters for comparison here, primarily, the oubble 10 rise velocities we estimated from looking at our test-films 11 l and observations, rise velocities of bubble clusters that l 12 Arlen was talking about a little while ago, on the order of 13 about 5 feet per second for.the types of flowrates that we 14
,- would be encountering under severe accident condition.
, 15 ,
And the bubble rise velocities we calculated l' baced on some of this relatively crude modeling puts us in 17 that same ballpark, within'3 to 5 feet per second bubble l 14 rise velocities -- I am sorry, swarm rise velocities. , 19 (Slide) O I The next facet of our overall program was to put ! 21 all this together and perform decontamination factor 22 calculation. We used the model we have identified for each 23 particle size in the swarm, and then we summed the particle 24 size DFs over distribution to obtain the total Dr. That is > 25 l shown xind of schematically in our next chart. 1 (* ' TAYLOE ASSOCIATES ' 1625 I Street, N.W. - Suite 1004 Washington, D.C. 20006 - t202) 29H950 l i
259 1
. (Slide) 2 This shows a bubble swarm rising in a pool.
3 Again, we. calculate over the bubble swarm over particle 4 size distribution to arrive at a total DF. 5 (slide) -; 5 Just to snow you some fairly typical resulta, 7 these were for some of our BWR-6 analyses. Depending on l 8 what the flow path was for. going down through the safety 9 l relief valve discharge lines through the quenchers into the 10
, pool, we would e'ncounter typical parameters such as these.
11 Velocities submergence heights being fairly critical 12 parameters, using a 5.1 AMMD particle size distribution, we would calculate very signficant DFc, on the order of 1 x 10 i 14
,3 to the 4tn.
I ~' For the case of horizontal vent discharges, now 16 we are dealing with both aerosols emanat.4.ng from thorium I interactions with the internals materials in the reartor 18 vessel as well as interactions with the' concrete using II ditferent particle size distributions typical of theso two 20 types of distributions. 21 he would octain decontamination factors on the 22 order of about 100, values like on the order of 2 x 10 to 23 the 3rd. (Slide) Overall, from the results of the program, some (/ TAYLOE AS8OCIATES 1625 I Stz% N.W. - Sune 1004 Washington, D.C. 20006 , (202) 293-3950
~
260 c 1 of the analysis parameters that we^have defined that are
. 2 very significant in the resulting calculations are the 3
bubble swaim diameters, the submergence neight, obviously. 4 Arlen, I think, also indicated, and I thing tne Battelle 5 people also' indicated, the significance of that parameters 6 in the calculations earlier this morning. , ~ Also, the bubble rise velocity and particle d size. Particle size I think has been mentioned very o
~
frequent 1piodEh,asbeingveryimportant t in identifying J-10 pool scruobing factors. Based on our analyses, we havex_ "~ 11 incorpor bd into stu:11es for standard plant design, we. -c 12 usud realistic d$ contamination factors anywnere from valuesl~ 13 on the order of 600 to greater than 10 to the 4th. 14 I thin < that was'about all we had to present.- - M R'. SILBER 8 ERG: Thank you, Debra and Kevin. 16 . Do we have some question <s/~ - y g e. - MR. COBOLE: what happens haienthe temperature in the pool changes that it changes surface tensions? Oces i 19 that hel<p tiecontamination factor? 20 MS. HANKINS: As I indicated, our model is for 21 the nonconcensing cases;'so we don't really include heat + 22 transfer into it. If y$u were to.] include heat transfer 23 terms in the model, what happens inythe terms of the , 24 hydrodynamics,youwouldpredictbecauseyounavea^lowe[. surface tension you would actually end[*4p with smaller n , .,- r. r [k/ TAYLOE ASSOCIA TES , ,s 4 1625 I Street, N.W. - Suite 1004 " , Washington, D.C. 20006 < (202) 293-3950 s [ P,w g s s
,%. -[
261-1 1 ~ stable size bubbles.- So stable size would go down.
. MR. COBBLE: The other question is what happens
.to surface tension when the pool oecomes saturated-with 4
gas? Do you do experiments?. 5 If I remember. correctly, gases change surface
'O tension'of the water too.. But even if you keep the A- temperature the same, are there any data on what happens in 0
the gas tension or. surface tension? 'That'apparently is 9 important to the calculation. 10 MR. COOPER: Could I comment on that? I think the surface tension is, what,'72 times per centimeter and the 12 size of croplets presumably by analogy, the size of bubbles 13 changes with the square. root of that. 14 If you take a look at the temperature dependence of surface tension, it really is quite weak. The addition 16 of surface-active agents can change surface _ tension.- Even I there you see at most a factor of 2 change.- So I don't
-18 tnink this is going to turn out to be a big issue.
You might check. There has been a recent very 20 nice book out from Clif t on bubbles. MS. HANKINS: Clift, Brace, and Weber. 22 MR'. COOPER: Yes. You might also loon in the 23 analogy with bubbles because the same Weber number ought to 24 control. 25 Finally,.I calculate an impaction parameter and [%/ ~ ; TAYLOE ASSOCIATES 1625 I Street, N.W. - Suite 1004 Washington, D.C. 20006 (202) 293-3950;
, . . . , - . _ . _ .~. . -
262 1
-a ratio of settling distance to bubble size, and settling
.n. h 2
(;i-
-to me,-using something in analogy to what you are doing, 3
-seems to have to outweigh impaction greatly. nhat have you found?
f 5 gg, ggggggg, .Ihe same result that Arlen 6 presented. The dominant term'for between about .2 microns 7 to about 1 micron in size is inertial.
- MR. COOPER
- I calculate 5 microns.
MS. HANKI6S: For 5 microns I agree.
-10 '
Secimentation, aosolutely. MR. COOPER: Maybe that's where the disagreement 12 g,, 13 MS. HANKINS: If Battelle would give us 5 micron 14 particles, I~would love'it. 15' MR. SERNERO: -we. won't let them. 16 MR. COOPER: The thing that surprises me is that l 17 -
'both should go with diameter squared as the aerodynamic 18 clameter. They should scale on'down together.
' 19 MR. VOGEL: I might just scope what the EPRI 20 j scrubbing program is just for completeness here, j We have under way at Battelle-Columcus -- and 22 Mike and Jim.are' involved.in this program -- a worx > involving this probing, and I have a feeling that certainly
'24 cy winter we are going-to have some appreciable data, i- The Battelle so far have been. aimed at 2/ldths t
, j TAYLOE ASSOCIATES 1625 I Street, N.W. - Suite 1004 Wash:ngton, D.C. 20006 (202) 293-3950 4 % -
, -- n - , -. .
. _ _ .. u . ; ___ .- ]
263 Hof a micron particle size. As Debra says, tnis has been (7N 2, (; focused very well on the most difficult one. 3 In general, we are finding decontamination 14 factors'in the neighborhood of 4 or 5 all the way up to 5 almost a thousand. And the important variables are the 0 amount of steam in the bubble, which, of course, you get I condensation. And that collapses the bubble, which is 8 ~ ciscrete, and.that gives you very good scrubbing. We are 9 varying submergence and so on. 10 I must admit that since I an involved with the 31
' program, that I have perhaps'sl' owed the model development 12 cown a little bit, and perhaps the reason for this is that 13 I. felt we would only have to develop one model. You know 14 how it goes, you get a model developed, then you get the OC 15 ~
experimenta1 data and then you've got to go redevelop your
'16 model. So that's where we are on-that program.
17 Our own model, as it now stands, gives somewhat 18 higner decontamination factors, by and:large, than post 19 buzz does, but'I can't get very much exercised.about that. 20 I enink it will converge. 21 MR. SILBERBERG: Thank you. . 22' Jim Gieseke.has a few comments. 23 MR. VOGEL: Do you want to say something about 24 tne EP'RI program? Go ahead if you want to. 25 MR. SILBERBERG: No. Just some comparison. L 7' TAYLOE ASSOCIATES .
\j 1625 I Street, N.W. - Suite 1004 Wasnington, D.C. 20006 (202) 293-3950
_ .. ._ ____..,c_.-._- . .
l' 264 1 MR. .GIESEKE: I just wanted to mention, it was
~ .s -
.L., ' alluded to the fact that there are scmo comparisons in the
~
3 works:cetween the various codes. You nave heard that there
.4 are'three codes available for the calculated pool 5
scrubbing.- 6 (Slide) 7 These are what Dick just mentioned that they 8 have developed at EPRI as_being the supracode. I think SAI
' worked with EPRI in the development of that code. The one 10 we heard from Arlen is the NRC-sponsored work at 11-i Eattelle-Northwest, the SPARC code.
12 l I don't know if you nave a name for your code. 13 t MR. HANKINS: DCON. l4 MR. GIESEKE: I didn't know,fso I just calle'd it 1 'E 15 General Electric. 10 (Laughter) 17 What we're.trying to do'is just pick some 18 generic type of conditions that-go along.with BWR 19 ex-quencher, as an example, to make comparisons-among the 20 codes. 21-The other thing that~has been offered is the 22 possibility that some~of the experimental results that Dick 23-nas' mentioned, we have worked with Richard Olberg on this, "24 and there will be some experimental' cases identified that 25 will' also be. usec in comparing the codes.
- (oM .-
u TAYLOE ASSOCIA PES 1625 I Street,' N.W. - Suite 1004 Washington, D.C. 20006 (202) 293-3950 T
% -a e % m m,u w1w4.===*- s*w *-~- ,
*=-*+0="- ' *.9 * * - * -
- r 265 1
We started to make a go-around comparison with p, 2
's. . the codes, and we found there are slight difterences among 3
all the codes. For instance, Arlen talked about the 4 condensation on soluble particles is in that code, is not 5 in the others. There are some with injection deposition and 6 not in the others. There are some that calculate the 7 bubble size internally and out. There are some that U calculate the aspect ratio internally. Tne otner one, 9 Arlen's, you fixed that at the beginning, 10 So there are all these differences among the 11 codes, and it's going to take a few iterations, I think, to 12 sort out, to get some cases where everyone agrees to do the 13 calucation the same way just to get some feeling for how 14
.,3 close the codes are.
Nl-] LS But we're trying to work towards understanding 16 what the ditferences are and see if there are major 17 differences in the calculations or whether they are fairly 18 close. 19 MR. VOGEL: Why don't we just take the 20 experimental data ana plot it up on a log? 21 MR. GIESEKS: Fine with me. It's certainly a lot - easier for me. 23 (Laugnter) 24 MR. COOPER: We could take the theoretical 25 results and plot them. TAYLOS ASSOCIA TES 1625 I Street,-N.W. - Suite 1004 Washington, D.C. 20006 (202) 293--3950
q_- .. y
; +
266' [' r. V 1 f - ( Laughter ) r._A. i; b ' MR. SILBERBERG: Thanks, Jim. V
'3 Before closi'ng, I would like to briefly note i
4' what tne order of business'is for tomorrow and some }: 5 ( ' contributions that we will need first thing in the morning. 6 l from all of you'here. 7 At this point, I think.I will spare you all the l' 8
- chairman's summary and closing remarks, because the 9
chairman's comments at this point -- and I will save those i 10 . Cor tomorrow. Iwouldverymuchhppreciateif,startingfirst 12 thing in the-morning, we will;maybe go arond the table as 13-we did last time in January for.maybe a 2-minute summary 14 i
- g. statement, preliminary-statement. We' recognize the boiler b 15 review is still ongoing. 1 don't want to hold you'to that.
I 16 - But 2-minutesummarystatement\ofwhat you: thir$k'are key ~ l 17 I points that you would like to make at-this point and based on what you heard today as well;l~ 18 as some material that you 19 might peruse through tonight, wnat topics do you think need 20 '
' some further discussion at this' meeting, you know,
! 21 tomorrow, so that we can prioritiize our time.,in our 22 discussion tomorrow. 23 And I then will ask he invited observers to 24 f . taxe similarly 2 minutes but only to add:new inputs over 25 and~acove what the panel at the' table has contributed. We g 4 TAYLOE ASSOCIATES
- 1625 I Street, N.W. - Suite 1004 Washington, D.C. 20006 (202) 293-3950
, , , sm.4 mN. # -..-e.
g 6% we p. uw up., we . ~ -w . - w', i - , . <= J r <u . k_
2 267 1
. will try to_do that within the first hour tomorrow morning.
2 (l.- And then, based on that, we will then proceed , 3 with further discussion of key points or perhaps maybe have 4 some additional presentations, backup presentations by the 5 Battelle-people. we may even bring the tellurium issue to 6 the table.from the Oak Ridge work, or any other subject, as 7~ well as'what -- and I think th'en to close it tomorrow we 8 will' then try to refocus again and try to summarize where 9 you think the-attention should be placed oetween now and
~
next time. We are in the completion of tne BviR study by
'Battelle-Columbus.
12 e
.ith that,'I would like to thank you for your 13 patience and_your_ indulgence. We wish yo'u a good evening.
14
.ne will meet tomorrow again at 8:30. Thank you.
~15 .
(whereupon, at 5:40 p.m., the meeting was 16- . aojourned.) 17 ' 18 19
- 20 211 22 ,
23 24-25-
. -1
- /
O' i TA YLO E: ASSOCIATES
- 1625 I Street, N.W. - Suite 1004 -
'washmgton, D.C. 20006 ~
(202) 293-3950-cro J.13
. - a , . a - . .- , ,; . , .. -
3
' C,ERTIFICATE OF PROCEED E::Gs
~~\ 2
, . ~,
3 This is to certify that-the attached proceedings before the 4 ::RC CO:0!ISSION
^t*
5 In the matter of: PEER REVIEW MTG. Date of. Proceeding: f tAY 24,1983
., . Place of Proceeding: Washington, D.C.
, were-held as herein appears, and that this is the' original
, transcript'for the file of the Commission.
10
,,- Mimie Meltzer Official Reporter - Typed it
~
9 !b ' k% h a !' ) .,, Of ficiad - Reportez()- Signature 1 15 i-16 17 18 19-ao , 21 I I 22
^
23 24
/
25 TAYLCE ASSOCIATES
-REGISTERED PROFESSIONAL REPORTERS NORFOLK, VIRGINIA
- m - ,. . _ . _ _ . . , . . . ~ , . - . . _ _ . . . _ . . _
,- e ,o , - . . , . . . _. _
VOLUME I
\s APPENDIX THE TRAP-MELT CODE
- w. .
s 9 - h
" ' ~ - ~~~~= eor. . ~ ~ _
,r'T
' APPENDIX
\ _.s THE TRAP-MELT CODE The philosophy and logic structure of the TRAP-MELT code are dis-cussed in the main body of this report. Here we present detailed expressions of the mechanistic treatment for reference purposes. This treatment remains - essentially that contained in the published TRAP-MELT manual
- and the reader is referred to that document for additional insight.
Master Equation The master equation set of the TRAP-MELT model is:
- I 8 n"n N m"3 m npm I k
'I "sf*MI* () )
nfm
+
jfj j m jm
~
-jfj im m where M = Mass of radionuclide species k in volume i and state m-Sh, = Source rate of species k in volume i and state m U
S , = Transfer coefficient for transport of species k in volume i from state m to state n d FI * = Transfer coefficient-for transport of radionuclides in state m from volume _i to volume j. , f %.)
.A. e --
e--wr- . v_6-a -*.--'e*wwe a- -, m. e ,- , e-, y a,
' ~
- 7,
+ ( For a given species k and volume i, therefore, Equation (1) gives, in order of
-appearance of.the terms on the righthand side, the mass source rate to state m, the mass transport rate to state m from other states in volume i, the mass transport rate from state m to other states in volume i and the mass transport rates to and from state m'due to flow in and out of volume 1.
If m signifies a surface state, "8f,. represents a mass release rate,
.. P . 'At the present stage of TRAP-MELT, all these terms are set to zero. If m signifies a volume state, "8f, can be written as "8 m V d (2) where vd is deposition velocity of a given mechanism and Ag is the appropriate deposition surface area. V9 is the volume of the control-volume in question, i The bar indicates the average over particle mass distribution (if m signifies a particle state) and surface areas.
Each contml volume is assumed homogeneously mixed. Mass transport due to flow between volumes can therefore be expressed by d Fg = d 'n r g/o39 Vj (3) where d Steam mass flow rate from volume i to volume j (input to mg = TRAP-MELT) o39 = Density of steam (and hydrogen) in volume .i. t-Deposition Velocities - L- .
..(1) Particle settling due to gravity '
2 odC T. '= h = particle response time (4) t' , vd " '9 (/ b3 - , ,, , ,_ . , , . _ , , , , ,, ,..m ,, , , 3
V \ where ; o = Particle density p
'd = Particle diameter C = Cunningham slip correction factor u = Dynamic viscosity of carrier gas g = gravitational acceleration.
TRAP-MELT distinguishes between settling across and against steam flow. If , settling is against the flow, then . vd"Vd -u u<v d ()
=0 u>y d , .where u = steam flow velocity.
(2) . Particle deposition due to diffusion from turbulent flow (Davies *,
- theoretical expression)
'(#)
Sc-2/3 v, = j (6) tan724-1 +xg ] 14.5[g in (j,,)2Z 1+ 4. -1 j_ l o = Sc /3/2.9 u, = hl/2 u f = 0.0014 + 0.125 Re -0.32 vd " V+u . Here Sc = Schmidt number = v/D v = Kinematic viscosity of steam D = Diffusivity of particle in steam Re = Steam Reynolds number in the volume of interest
, , , f = Fanning friction factor-u, = Steam friction velocity.
- Davies, C. N., Aerosol Science, Academic Press (1966).
i I n., .,. yw'. * *'
A (3) Particle deposition due to impaction from turbulent flow (Liu and Agarwal, modified by Lee *). An empirical correlation of Liu and Agarwal, extended to small particles by Lee, gives: v;=6x10~4 g +2 2 x 10-8 Re t < 0.1 (7)
= 0.1 i > 0.1 2
t, = t ujy vd " V+u,. 4 (4) Particle deposition due to diffusion from laminar flow (Gormley and Ke nne dy**) . Laminar flow contradicts the general assumption of homogeneously mixed control volumes that is fundamental to TRAP-MELT. In order to, neverthe-i less account for deposition under such conditions, a fictitious deposition
! velocity is introduced that, when used in TRAP-MELT, gives the same rate of deposition as would be calculated by a differential treatment' of plug flow.
For_ pipe flow, it can be shown that this deposition velocity is: M y ) u d = (1 - (8) where R = Pipe radius L = Pipe length Mj = Particle mass concentration entering pipe M, = Particle mass concentration leaving pipe, e
- Gieseke, J. A.~, et al, NUREG/CR-1264, 'BMI-2041 (1979).
(f **Gomley, P. G. and Kennedy, M., Proc. R. Ir. Acad. 52A,163 (1949).
A
! ., i
' According to the theoretical analysis (substantiated by numerous experimental investigations) of Gamley and Kennedy:
M g) [i = 0.8191e-7.31h + 0.0975e-44.6h + 0.0325e-114h . h > 0.0156. [M = 1 - 4.07h2/3 + 2.4h + 0.446h4/3 i . h < 0.0156 h = LD/2 uR
= Pe'I L/R where Pe = Peclet number = Sc x Re.
(5) Particle deposition due to themophoresis (Brock *). Brock's theoretical treatment of particle deposition in a temperature gradient, vT, at a wall surface gives: vd =- v$Cf where j kg /k +C t K" i
*"(1+3C,M +2k p+2CU t /
.C , = Steam momentum. slip coefficient Ct = Temperature jump ccefficient Kn = Particle Knudsen number kg= Themal conductivity of gas (steam + hydrogen) kp= Themal conductivity of particle.
;.
- Brock, J. R. , J. Colloid Sci . , l_7,, 768 (196?).
, , , - - - , - n
- Note that for large particles, Kn + 0 and kq /k,
*
- I~+ 2 kg/kp yielding an order of magnitude variability in v , ddepending on the choice of_kp. In TRAP-MELT, the necessary temperature gradient in Equation (10) is derived from the' simple pipe flow heat transfer correlation:
0 Nu = 0.021 Re .8 , ())) using the identity hai = kvT. (12) aT = T,,jj - T gas is derived from input data. Nu is the Nusselt number. (6) Vapor sorption on wall surfaces p (*% , -- Molecular iodine from steam to stainless steel surfaces (Genco*) v -8,8100/kBT (cm/sec) (13) d = 9.0 x 10 kB= Boltzmann's constant. Molecular tellurium on stainless steel 304 (SANDIA**) vd= 1.0 (cm/sec) (14)
-- - Cesium iodide v = 0.
d (No data available) (15)
-- Cesium hydroxide (SANDIA**) ,,, vd= 0.01 (cm/sec). (16)
- Genco, J. M., et al, BMI-1863 (1969).
** Elrick,- R. M. and Sallach, R. A., "High Temperature Fission Product Chemistry and Transport in Steam", Proc. of the Internat'l Meeting on Thermal Nuclear ,
. Reactor Safety. August 29-September 2,1982, Chicago, Illinois.
,n
,- . - - - - - + y -
E-- T !
'e L ;
f ,, , Y Species Phase Change In each control volume, each chemical species is permitted to con-dense on (or evaporate from) particles and wall surfaces according to the mass
- transport rate equations:
~dC,. pP s dt
" ~ A,kw V
( s - C,s) -(A y k ) @3 -Cp) dM dt
=
A,k,(C, - C,s) (17) dM
- 5 dt AkP p (C s -Cp) t-9
_- s where , j e M' s s C, = 7 = concentration of the nuclide vapor in steam M, = Total . mass of ...e nuclide vapor in steam 'f
, V = Volume vf the control volume y M, = Total mass of nuclide vapor condensed on$ alls i M p = Total mass of nuclide vapor condensed on? aerosol particles lCgs = Equilibrium vapor concentration of the$uclide at the
, temperature of the wall surfaces (asstaned independent of
%: pressure)
C 5~ kdciuf. librium vapor concentration ofIthe nuclide at the p temperature of the steam (assumed 0 independent of pressure and';iaiticle surface curvatuVe)i: ~ A, = Area of Mii9s,Upfadds$ A = Surface area ~of ' aerosol particle - p k" = Mass steam and transfer coefficient forinterface wall surfaces-steam nuclide transfer between k '= Mass transfer coefficient for nuclide transfer between - P steam and particle surface-steam interface
=v . s A . . - p . , e o .w w. ~ we - - .mg,,,,2+-...v- + ,
- y. m .
m. ( kwis.taken from the Sherwood number (Sh) correlation for turbulent pipe flow (Dittus Boelter): 0 0 Sh'= 0.023 Re .83 Sc .33 (18) Sc = Schmidt number kg= D/ r with r a particle radius. (A kpp) is the average value of App k over the particle size distribution in the control volume of interest. Equations .(17) are solved analytically on the assumption that (Ap kp changes little over a master time step. This is borne out in practice. The effect of condensation / evaporation on the particle size distri-bution is taken into account by noting the total mass (summed over all chemical species considered) transferred to/from the particle state according to Equations (17) over a master time step. This quantity is distributed over the discretized particle size distribution such that each size class is aug-O mented/ diminished in proportion to its associated mass transfer rate. s Required vapor pressure data (C,s, C p ) for I , CsI, 2 ~Cs0H, and Te are presently incorporated in the code, particle Agglomeration The aerosol component of the radionuclides tracked by TRAP-MELT is distributed among 20 size classes. Agglomeration among particles in these size classes.is treated by a method developed in the QUICK aerosol behavior code
- and since validated against numerous experiments. The coupling of this treatment to the flow equations of the TRAP-MELT code is described in the body of this report. Here we exhibit the agglomeration mechanisms considered.
- Jordan, H., et al, " QUICK Users ' Manual", NUREG/CR-2105, BMI-2082 (1981).
w - 4 ye-- N 9F T
e
.t '
Brownian Agglomeration Defining the agglomeration kernel, Kj ), by (19) Rj ) = Kj )N jN) such that Rj ) is the rate of agglomeration of the Nj particles per unit volume in size class 1 with the N) particles per unit volume in size class j, the
. kernel for Brownian coagulation can be written:
~
K jj = 4wkB T(Bj + B))(rj + r)) (20) where C 0 1 " 6wurg and rj is a characteristic particle radius for size class 1. Gravitational and Turbulent Coagulation. Following Saffman and Turner *, the combined kernel for gravitational, turbulent shear, and turbulent inertial agglomeration can be written 2 N K jj = 2 42 (rj + ry)[c j(tj ,,j)2
+
c)[rj-r))22+ g (rj + r))2 3 1/2 where the as yet undefined quantities are: (Fi,ry) cj ) = 1.5
# * "I" r r'
, r' = max (rg r)),
; *Saffman, P. G. and Tiirner, J. S. , J. Fluid Mechanics, 1, 16 (1956).
-.%-% 4%ww. +- , .
4 . . _ . . _ _ . . ._ 4 1
$\
; (/
.the collision efficiency for hydrodynamic interactions and E the turbulent energy density dissipation rate. TRAP-MELT. uses Laufer's expression *:
3 3 E = 0.03146 u /(D Re /3).' t i 4 i i
\
t (. l I
\
\,
t, 1 i + 9
. (. .
(.;' ;
*Taken from Delichatsios, M. A. 'and Probstein, R. F., MIT Fluid Mechanics i Lab Publication #74-5 (1974).
l
. . . , , . . . ~ . . ~ . . . . - . . . - . . .
- r. .
.;*=,m e
PEER REVIEW CO MENTS ON- - BMI-2104 - VOL. 12(SURRY) O . Christopher P. Ryder Accident Source Tem Program Office Office of Nuclear Regulatory Research April 14 -1983 4
l ACCIDENT SEQUENCES AND SYSTEM BEHAVIOR ; ISSUE REVIEWER AFFILIATION From a given sequence, alternate release S. Levy Levy. Inc. pathways are not considered. Such pathways may result from breaks at a variety of locations in the primary system. For sequences that lead through the J. L. Kelly Univ. of Virginia auxiliary building, retention by the E. P. Rahe Westinghouse - auxiliary butiding is not considered. B. R. Sehgal EPRI E. A. Warman Stone and Webster H. Kouts BNL For the TMLB'~ sequence, the cause of the E. P. Rahe Westinghouse 89 psi pressure spike is not identified. If the cause is a hydrogen burn, the occurrence of such a spike is questionable because steam will likely attenuate the burn. ( ~ ? e (.).,
. . . . . _ . . . . _ . . t. . _ . _
v - , v
FISSION PRODUCT RELEASE FROM FUEL
)
~
ISSUE REVIEWER' AFFILIATION In addition' to' cesium iodide and tellurium, J. L. Kelly Univ. of Virginia . - the release of other radionuclides should E. A. Warman Stone and Webster ! be investigated and discussed. . Apparently D. O. Campbell ORNL ! aerosols of uranium, plutonium. and silver .
, are not considered.
The codes have cesium iodide and tellurium E. A. Warman Stone and Webster being released when the core melts. Indications are that these radionuclides are released before the core melts and that the release is completed when the core melts. Also, aerosol production may : begin before the core slumps. 4 1 The maximum core temperature should be A. W. Castleman Penn. State Univ. i - determined because the production of l aerosols is temperature dependent. ,
' O, ;O An estimate of the size of aerosol particles D. W. Cooper Harvard University ~
4 is based on experiments with Nacl solutions. ; From these experiments, it is predicted that sparged particles from the melt- ; concrete interaction would be lum. Because particle size is dependent on the square j root of surface tension and the surface < i tension of the melt is much greater than the surface tension of aqueous NaC1, the lum particle diameter seems small. ! More study is needed on. scaling up the A. W. Castleman Penn. State Univ. results of experimental fission product releases. )~ The aerosol particle sizes are assumed. S. Levy Levy. Inc. It is not clear why a particular size distribution is chosen. - It is erroneous to assume that control rod C. E. Johnson Argonne i material will vaporize simultaneously and at the same rate. The composition of the
- vapors from the control rods will likely be
/, 1 different than the composition of the control D rods. This is supported by data from TMI-2.
~
, .a., _ - , . _ . - . _ , . . , _ . _ - _ , _ , . _ . _ _ , _ . . .
4
,. THERMAL HYDRAULICS ISSUE REVIEWER AFFILIATION High pressure ejection of molten fuel from A. B. Reynolds the vessel is inadequately considered.
Molten fuel striking water in the reactor cavity may fragment and capidly transfer heat to the water; a steam spike would result. Hardware surfaces would be warmed by A. W. Castleman Penn. State Univ. decay heat from radionuclides. This J. L. Kelly Univ. of Virginia heat influences radionuclide deposition. S. Levy Levy. Inc. The codes do not consider the decay heat- E. P. Rahe Westinghouse deposition phenomenon. A. B. Reynolds . D. S. Rowe Rowe, Assoc. Flow in the upper plenum is inadequately J. L. Kelly Univ, of Virginia modelled. Upper plenum flow influences B. R. Sehgal EPRI the rentention of fission products. 0: e Mass and energy balance are not modelled. H. Kouts BNL A. B. Reynolds D. S. Rowe Rowe, Assoc. The modelling of the boundary layer in A. B. Reynolds relation to thermophoresis is not discussed. i 0 f ('. L.- l
, . - - . y , -, - - ~ - ,m. - , - --
t ! CHEMISTRY ISSUE REVIEWER AFFILIATION l The codes .have cesium iodide undissolved in E. A. Warman Stone and Webster : the primary system. This is unrealistic
- i -in light of the copious amounts of. water
- .present.
I Reaction cesium hydroxide on tellurium B. R. Sehgal EPRI and stainless steel have been observed in E. A. Warman Stone and Webster
! the laboratory. These reactions are not l modelled. The reactions are important -
because it is a way that fission products - I can be retained in the upper plenum. l Chemisorption must be better understood A. W. Castleman Penn. State Univ. and considered in the models. D. O. Campbell ORNL The elemental form of iodine is in the A. B. Reynolds codes. This form of iodine is doubtful. , High radiation fields produce ions. The D. W. Co'oper Harvard University differences in ion mobility result in a net charge on an aerosol. The charge , differences lead to enhanced coagulating. 4 The net charge leads to enhanced scattering. i The computer codes do not account for ! ionization. ) j It is unclear whether deposition is a D. O. Campbell ORNL i - reversible physical deposition or an ;
- irreversible chemisorption. l Before fission product chemistry can D. O. Campbell ORNL be modelled, it must be better understood.
- Significan't problems in the mo:!alling of C. E. Johnson Argonne 1
tellurium exist. For example, compounds i of tellurium are not identified. i s t
, , , ,w -.,. . , ,. , . - - , , , - . , , . . - - , _ - -,,y, . - , . , , _ , _ , , _ , , . + - - - _ , , . , , , , . . - , - - , . - . . - - - , -
=
-PRIMARY SYSTEM-TRANSPORT, DEPOSITION, AND REENTRAINMENT
, [ OF AEROSOLS ISSUE. REVIEWER AFFILIATION The effect that deposition (in pipes) has on D. S. Rowe Rowe, Assoc. flow is not modelled. When aerosols pass through pipes at high D. W. Cooper Harvard University speeds, turbulent deposition occurs. Turbulent deposition is not explicitly considered. The deposition of aerosols in primary piping E. A. Warman Stone and Webster and on vessel intervals is inadequately modelled. The code fails to include retention by quench E. P. Rahe Westinghouse tank and the relief line downstream of the PORY. j
/'
wh
~ ~ -- - - -
,,yg- y- - , , , , - - , , -
9
,o CONTAINMENT-TRANSPORT, DEPOSITION, AND REENTRAINMENT 0F AEROSOLS ISSUE REVIEWER AFFILIATION The model predicts an aerosol concentration D. W. Cooper Harvard University of 103/1m3 to to persist for thousands of seconds. This high concentration is unstable and should not persist long.
Convection currents in the primary containment D. W. Cooper Harvard University should keep the atmosphere well mixed; diffusion of vapors from hot to cold surfaces . occur. This is not modelled. When substantial volume fraction of D. W. Cooper Harvard University atmosphere containing aerosols are condensing on walls, diffusiophoresis should occur. This is not modelled. O " It is stated that particle size must be about D. W. Cooper Harvard University (" 0.6 um for condensation on the particles to occur. Typically, condensation occurs on particles as small as 0.01 um. Also, particles that contain soluble species will grow from condensation, even in an < unsaturated atmosphere. Battelle inferred that spray droplet size A. B. Reynolds does not have an appreciable effect on ; containment purging. It is expected that droplet size will have significant influence j on purging. 1 The model assumes that no convection flow T. Ginsberg BNL is in the primary containment. This i seems unrealistic. 1 i
. ., ~ . . .. ..n -~,.-.n.~. - - -
e- 4,---.,.-m. , ,-,-- n, - - , ,, ----,-- -m w-mr.y -- .
l k.. - CONTAINMENT LOADS AND FAILURES ISSUE REVIEWER AFFILIATION A mechanistic model that has failure mode, J. L. Kelly Univ. of Virginia failure location, and failure time, is E. A. Wannan Stone and Webster needed. The model of the containment is a single E. A. Warman Stone and Webster zone. This is unrealistic and does not D. O. Campbell ORNL represent a "best estimate." Data from TMI-2 indicates that the atmosphere in the containment is heterogeneous. In the model, the containment leak rate is E. P. Rahe Westinghouse 1%/ day. The design basis for large dry PWR containments.is 0.1%/ day.
~
l l U
~ '
.a ,
-( CODE VALIDITY _AND SENSITIVITY -
ISSUE REVIEWER 57FiLIATION is oversimplified. J. L. Kelly Univ. of Virginia Some Many instead of theofmodelling few co 'ntrol volumes might A. B. Reynolds have been used to model long systems. R. L. Ritzman SAI Reactor internals, as they influence aerosol' D. S. Rowe Rose, Assoc. . behavior, are inadequately modelled. Cesium B. R. Sehgal EPRI .c. - - iodide and tellurium should have been ,, E. A. Warman Stone and Webster modelled so that they enter the vessel internals, piping, and pressurizer. The.m - 4 empirical release model, from NUREG-0772, g is used in the CORSOR code but has questionable . validity. Hydrogen burn would likely be attenuated by steam; this is not modelled in , r ' the MARCH 1.1 code. / ; -
) i .,
The codes out perform the current knowle"dge A. L.' Cahn ~ Sechtel , r of the physical processes. Some pbetions T' D. 07 Campbell ORNL of the codes are extremely sophisticated -A. W. Castleman Penn. State Univ. while other portions are extremely simplistic. J. L. . Kelly Univ. of Virginia Thus, the output is only as good as the input .H. Kouts BNL _.i and the portions of.the codes. S. Levy Levy, Inc. ~ G. Thompson UCS ; Uncertainties are inadequately treated. T. Ginsberg . BNL , What they are and how they are to be resolved W. D. Harrington Boston Edison should be discussed. H. Kouts BNL R. L. Ritz
. SAI E. P. Rahl Westinghouse G. Thompson USC The results of CORRAL-2, NAUA, and industry - A. W. Castleman Penn. State Univ.
codes are in disagreement. This should be D. W. Cooper Harvard University discussed. A. B. Reynolds ' A sensitivity analysis is needed for tha . . A. W. Castleman penn. stati niv. analysis of aerosol deposition and resuspension. H. Kouts e BNL The methodology is inconsistent. The results' ' D. O. Campbell ORNL .. are thought of as "best estkate" but some W. D. Harrington Boston EUlson calculations are conservative ano some B. R. Sehgal EPRb sequences are worse case. ' () ' ' s r / _ - , , . ~ . . . . , . . . . .. .- ~ r .=
*v---- -,--r -- --
wy , ,, , , - --
*r ,,_ -.
4 CODE VALIDITY AND SENSITIVITY (Continued) ISSUE REV.! EWER AFFILIATION l The assumption that the ccre instantaneously E. P. Rahe Westinghouse melts through the vessel after slumping is questionable. The quoted volume of the auxiliary building E. P. Rahe Westinghouse seems small. - The melting of the control rod is simulated D. O. Campbell ORNL in an experiment using tin and steel. However, tin and steel melt at temperatures that are several hundred degrees hotter than melting temperatures of control rod materials. The validity of this simulation is questionable.
., The codes produce results that are conflicting. W. O. Harrington Boston Edison
] For some sequences, the results from updated R. R. Hobbins EG8G
- x. models are similar to results from the WASH- H. Kouts BNL 1400 study. Other codes that produce differing results should be producing similar results. These discrepancies must be rectified.
The results of the code should be verified A. L. Cahn Bechtel 4 against TMI-2 data. D. O. Campbell ORNL T. Ginsberg BNL A. B. Reynolds D. S. Rowe Rowe, Assoc. B. R. Sehgal EPRI G. Thompson UCS A " dry" sequence is assumed. This is unlikely. H. Kouts BNL B. R. Sehgal EPRI [ s al i \
+se._ -~m.
g- -
-w., -- -rg-, --
3 ,
[ REPORT STRUCTURE AND ORGANIZATION
- ISSUE REVIEWER AFFILIATION The objective and purpose of the report D. S. Rowe Rowe, Assoc.
should be more explicit. Equations 7.2 and B-9 should be expressed D. W. Cooper Harvard University more conventionally. C. E. Johnson Argonne o An updated review of the data is needed.. *A. B. Reynolds The results should be highlighted. D. S. Rowe Rowe, Assoc. The methodology and the equations should be R. L. Ritzman SAI ; discussed. D. S. Rowe Rowe, Assoc. 3
., The report should be published with caution. D. O. Campbell ORNL Lj' Many qualifying statements are needed. J. L. Kelly Univ. of Virginia A position on the highly controversial topic A. B. Reynolds of containment failure should be taken. At best, the report should be considered on '
an interim measure. The basis for the gap release needs to be A. W. Castleman Penn. State Univ. , stated. Temperature is expressed in both the Fahrenheit E. P.-Rahe Westinghouse scale and the' Celsius scale. One should be
-chosen.-
The Surry plant is not representative of the E. P. Rahe Westinghouse current large dry PWR containments. The fission product release is dependent on. E. P. Rahe Westinghouse
-the-location of a break in the primary
. system.. Though this is in the research, this
, is not stated in the conclusions.
i
.I '
REPORT STRUCTURE AND ORGANIZATION (Continued) L ISSUE- REVIEWER AFFILIATION The calculations and the assumptions are E. P. Rahe Westinghouse not placed in an adequate perspective. The terms " deposition mechanism" and D. O. Campbell ORNL
' "depositon term" are poorly defined (p5-16).
i d 9
('; : O .') 4 SUPPRESSION POOL. SCRUBBING SPARC CODE A.K. POSTMA P.C. OWCZARSKI . W.K. WIN EG AR DNER BATTELLE PACIFIC NORTHWEST LABORATORY MAY 24,1983
c -: c7) OVERVIEW OF SPARC DEVELOPMENT ! e PART OF PNL PRO. GRAM ON PERFORMANCE OF ENGINEERED l SAFETY SYSTEMS UNDER SEVERE ~ ACCIDENT CONDITIONS ' POOLS, SPRAYS, ICE, FILTERS, CO'O LER S' e HIGH PRIORITY, LIMITED TIME ' e PRELIMINARY - SUBJECT TO l IMPROVEMENT i
o . 9 , GOALS OF MODELLING EFFORT i e REALISTIC PREDICTION OF
- PARTICLE SCRUBBING IN
. SUPPRESSION POOLS l e ACCOUNT FOR MAJOR PHENOMENA
-PARTICLE SIZE DISTRIBUTION
-POOL TEMPER ATU R E i -POOL DEPTH ,
! -STEAM CONDENSATION / l ~ EVAPORATION l e COMPATIBLE WITH TRAP-MELT AND MARCH
1 o ; g g , MODELLING APPROACHES l CONSIDERED i e BUBBLE SCRUBBING ; I e ENERGY DISSIPATED / UNIT ~ l VdLUME OF GAS I o APPLICATION OF EXPERIMENTAL l DATA . i l i
g :. O .O~ APPLICATION OF EXPERIMENTAL DATA e DATA BASE IS LIMITED ( i -SATURATED POOLS , L -STEAM / GAS RATIO - ( :;
-PARTICLE SIZE EFFECT .
3 l -SIZE SCALE OF TESTS e TECHNICAL JUSTIFICATION OF l SELECTED DF WOULD BE DIFFICULT
I CUT DIAMETER VERSUS ENERGY fl GAS-PHASE PRESSURE DROP (cm, H2O) 0.5 1.0 2 3 4 5 10 20 30
- 5. .. . . . . . . . . ..... . .
4 - 3 - E , , la ,
.5 lb 4
m -
% 1.0 -
$ 3*
~
5 4 o 3b - 0.5 - r
> 0.4 - /
3"
~b 0.3 -
REGION FOR SGS - 0.2 - - b' g,, 1.0 2 3 4 5 10 20 30 40 50 100 GAS-PHASE PRESSURE DROP (in., H2O) la SIE\fE-PLATE COLUMN,0.2-in. HOLE DIAMETER 1b SIEVE-PLATE COLUMN,0.125-in. HOLE DIAMETER 2 PACKED COLUMN,1-in. RINGS OR SADDLES 3a FlBROUS PACKED BED,0.012-in. DIAMETER FIBERS 3b FIBROUS PACKED BED,0.004-in. DIAMETER FIBERS 3c FIBROUS PACKED BED,0.002-in. DIAMETER FIBERS . 4 GAS-ATOMlZED SPRAY ' - 5 MO' BILE BED,1 TO 3 STAGES, HOLLOW SPHERES >
~ ~~~ ~
' ~'
L M ETHOD e SIMPLE, REQUIRES LITTLE INPUT ^ [ e SUPPORTED BY DATA OBTAINED FOR SUBMERGED ~ GRAVEL l SCRUBBER . l -MECHANISTIC MODELS GR.EATLY l U,NDERPREDICTED EFFICIENCY i e WOULD NOT ACCOUNT FOR STEAM l CO N DENS ATION, G EOM ETRY, POOL ! TEMPERATURE, PARTICLE GROWTH l : e ENERGY METHOD SHOULD BE l RETAINED AS A BASIS FOR COMPARISON 1
E n ,( BUBBLE SCRUBBING MODEL 1 l' 1 ll e FUCHS DESCRIBES BUBBLING MODEL L L -SEDIM ENTATION
-INERTIAL DEPOSITION e -DIFFUSIONAL DEPOSITION L
e DOES NOT EXPLICITLY MODEL REMOVAL IN l BUBBLE FORMATION REGION - J l I o BUBB:LE SIZE IS CRITICAL INPUT ! PARAMETER i e BUBBLE CIRCULATION IS A CRITICAL l t FACTOR e IMPORTANT DEPLETION MECHANISMS CAN l BE MODELLED l
- s.
- C O l DEPLETION MECHANISMS IN SPARC .
e CONDENSATION OF STEAM ! e SEDIMENTATION j . e CENTRIFUGAL DEPOSITION l e DIFFUSION l e STEAM EVAPORATION INTO RISING l BUBBLE i l e GROWTH OF SOLUBLE PARTICLES ! e ENTRAINMENT OF POOL LIQUID
*3 .
s- . o DECONTAMINATION FACTOR ( DF = AEROSOL MASS ENTERING AEROSOL MASS LEAVING 1 = 1 DF = PENETRATION 1-FRACTIONAL EFFICIENCY
- FOR.'A SINGLE PARTICLE SIZE Po = Pi P2 . . . Pn AND DFo = DFi DF2. . . DFn 1
l i FOR' AN AEROSOL POPULATION i=n i=n P=Ii=1 FiPi =I Fi / DFi = 1/DF 4 i=1 i i
n g e> : (D r3 e y . 4 DEPOSITION DUE TO STEAM CONDENSATION f o Xi ASSUMPTIONS: [l" . jQ' ; BUBBLES ATTAIN THERMAL e lihI; ',i .. j3; f EQUILIBRIUM NEAR OUTLET v ,,w: -
, Z NON-CONDENSIBLE GAS INVENTORY
. ws " ', - ~$j[{gy l 1 yO IS CONSTANT .
4 Xi FRACTION OF GAS CONDENSED = 1 O i ! ASSUMING AEROSOLS GO WITH GAS, DF = X ,DFB1
- Xi 1
l Xo = 1 - =1-PT P + ph
- l j 'X = MOLE FRACTION OF NON-CONDENSIBLES
! P.= VAPOR PRESSURE OF WATER IN POOL L P = ATMOSPHERE PRESSURE l h = SUBMERGENCE , p = WATER DENSITY 4 i
(5 3 .: .
. , a,,
}
o BUBBLE AND SWARM RISE VELOCITIES l' i g- -_- j ~ < ^ s c; < ;
. 2 _ . -m o .
^
- 4,,-,m..-.
BUBBLE VELOCITY . ,e - .
= mVbh.,. .. u z j * "
- 4 O. . x' . < i .E ...
?, b . *Es s?si.
., . ,SWAR M !.
VELOCITY = UL 1
- (*.
1
. % ..y. ~ g At.= h/U s .,
t.. j a k { .' l '
^ ' '- 4' '# '
, ; g-_ y
' -:S.- ._ f. ;_ .~J m
" SHATTERING
. ~ .
;,g 0 gt + 1, '
LEVEL GAS IN > ;--
- ~
m. sq<
;. , ^ . .
m s v s i 4 h( , _ /
..j POOly.
m_ , s , l i i 1 I 1 1 5 -
mui
^
O .D STEAM EVAPORATION VELOCITY e FUCHS CITES TESTS OF REMY TO ILLUSTRATE SIGNIFICANT IMPEDING EFFECT l OF EVAPORATION INTO BUBBLES ' e INWARD FLUX EXPECTED FOR RIS.ING ' BUBBLES e WATER VAPOR FLUX COMPUTED l ROM PENETRATION THEORY AT ACROSS LIQUID FILM AC ACROSS GAS FILM FLUX DIRECTED NORMAL TO SURFACE FLUX INCREASES WITH HEIGHT IN POOL AND WITH POOL TEMPERATURE
b
~
i ' (~' : 0 POOL TEMPERATU R E
- . e EVAPORATIVE COOLING COOLS POOL 1 BELOW NORMAL BOILING POINT I FOR 1 ATM., ASSUMING SATURATION OF L EXITING GASES:
Ti, C Xi Te, C -
- y
! 1000 1.0 77.4 l ? 500 1.0 68.0 t 1000 0.1 98.3 i e RE LISTIC POOLTEMP NEEDED FOR STEAM I- EVAP. VELOCITY e SPARC ALLOWS EQUIL. TEMP OR USER INPUT i i
~
l t
, b) .
! PARTICLE GROWTH BY WATER I SORPTION e SOLUBLE PARTICLES ABSORB WATER AND
- j. GROW IF RELATIVE HUMIDITY IS GREATER l THAN CRITICAL VALUE I
e AT EQUILIBRIUM VAPOR PRESSURE IN SOL'N DROP EQUALS WATER VAPOR PRESSURE IN . i ATMOSPH ER E ' ! e USING FLETCHER'S FORMULA:
~
DRY PARTICLE EQUIL. DROP RADIUM, m RADIUS, m H=0.9 H=0.99 0.01 0.0195 0.0295 j 0.1 0.195 0.425 1.0 1.95 4.45 e SPARC ALLOWS FOR WATER UPTAKE BY - SOLUBLE MATERIAL l i
O O O BUBBLE SHAPES a/b IS AN INPUTTO SPARC
- n i
b
! 1r
! < a >
; SPHERE OBLATE SPHEROID
~
i i m --~ - - - - - - _ - - - __
f O O
~
.'i :
SEDIMENTATION L 4 V.= Vs- p Sin O l'
------r dA = 2nR2 Cos O Sin O d6 V* O/ -
f VdA = wR2 (V -2/3 v) ~ g ' u , ! DF = exp [ 3/2 Vs- F ]At ,DFM1 .
. D l Vs = SETTLING VELOCITY I F = STEAM EVAPORATION VELOCITY D = BUBBLE DIAMETER . At = TIME FOR SWARM RISE i
l i i l
- .- _ _. 1 _ _ _ _ . _ _ _ . . - _ _ - - _ .
{ C ': O
^
- 3 SPARC PREDICTIONS .
1 03 . NO SOLUBLE PARTICLES Z _ O _ POOL AT 20 C ' F l - 4 . h = 12 ft ! Z l 2 1 02 _ T = 20 C t
< u. _
! FO :
- Z - '
l OE - i 00 - f }f wF OU ' - AXIS RATIO = 4 3 2 1 w< l 4 m' 10 -
- E ,
- A 3 %A m
! m - ! D b W - 1 ' li ' '- ' l'i ' 'il' i 10- 10-' 10- 10' !, AERODYNAMIC PARTICLE DIAMETER, pm j
o
~
j, O ji - 1i l KEY INPUTS TO SPARC e PARTICLE SIZE DISTRIBUTION e POOLTEMPERATURE COR INLET GAS ! TEMPERATURE > e PRESSURE ABOVE POOL ! e EFFECTIVE SCRUBBING HEIGHT i. i l e AVERAGE BUBBLE DIAMETER e STEAM / GAS RATIO OF INLET GAS j e INLET GAS COMPOSITION (H2, H20, CO, CO2) I l 1 i t
l KEY IMPROVEMENTS AND l EVALUATIONS . { e INTERNAL PREDICTION OF BUBBLE SIZE DISTRIBUTION - j e ENTRANCE IMPINGEMENT DEPLETION [BCL ADDITION) , l e COMPARISON WITH OTHER CODES, MODELS i l EPRI, GE, ENERGY - CUT DIAMETER i e COMPARISON WITH EXPERIMENTS e ENTRAINMENT MODEL ) e BUBBLE CIRCULATION l
p -.
,g g -
1 i
SUMMARY
AND CONCLUSIONS l' l e SPARC CODE DEVELOPED IN SHORT TIME
! FRAME 1
l e DOMINANT DEPLETION PROCESSES HAVE BEEN INCLUDED
~
~
! e PREDICTED DF IS VERY DEPENDENT ON PARTlCLE SIZE i e OVERALL POOL DF WILL DEPEND CRITICALLY ON AhRODYNAMIC PARTICLE DIAMETER e IMPROVEMENTS AND EVALUATIONS ARE l U N DERWAY
! ~
)~ ~
[
~
l 1 l l RADIONUCLIDE RELEASE UNDER l SPECIFIC LWR ACCIDENT CONDITIONS -- t
. VOLUME lil,. BWR, MARK 111 DESIGN i
! PEER REVIEW MEETING !
- l t
i O ;
- U.S. Nuclear Reg'ulatory Commission ,
Washington, D.C. l MAY 24 & 25, 1983 i l l l Presentation Notes ! l i i< o OBattelle
- c ,-e., a e... .,,,,
y . 1
[.[
}~
PRESENTATION TOPICS : e INTRODUCTION (J.A. GIESEKE) e SEQUENCE DESCRIPTIONS AND IHERMAL HYDRAULICS (R.W. DENNING) C e RELEASE FROM FUEL AND IRANSPORT IN REACTOR COOLANT SYSTEM (M.R. KUHLMAN) e
SUMMARY
(J.A. G!ESEKE) ~ OBaffelle ce,.e., u e.,......, t_ y
l O O 0 l
, DESCRIPTION OF ACCIENT SEQUENCES AND THERML-HYDRAULIC RESULTS BWR -- k nx Ill DESIGN ~
l 4 PLANT SELECTION j GRAND GULF'WAS USED IN RSSMAP ANALYSES.
~'
~~~
SEQUENCE SELECTION RISK DOMINANT SEQUENCES -- TC, TPI (TW), TOUV
.ii, i:, , n o! i 1l8 ! *.
c 4
, $ *t i i
,, . ,e-_ . . . - . -
(,. k < I
. r -
CONTAINMENT FAILURE MODE 1
!: FAILURE PRESSURE -- 72 PSIA LOCATION -- JUNCTION.0F CYLINDRICAL WALL AND DOME i . s I 'j '
t -
; FOLLOWING PRIMARY CONTAINMENT FAILURE, ENCLOSURE
< : . c, .
BUILDING WOULD NOT PREVENT MAJOR OUTLEAKAGE - w s
^' , s
- i - .
f I y ~ % ..*
.. v
~
m - s
' f p "'
%-s
, '\ " + 1 x- *
}
g g . ,, ,
*4 "
[ ( ,
,<$6
- \ / ~' '
, r:
g .
. . 4
- \c ,' i
,.4 . 3 , ..l.I e
- a -
a -~r. - .- - - .,n. ..
e
\s -
e.. ..
. OUTE R CONTAINMENT q .
- 6 a
a , - *
,.1 .
W ..
? .
! l.' !J l F 7* ,; :,
=
,- 6
. .s. , .. .
i- - >< .; , 2* .<
.s .
,., t_ . .
l ; *
; l. ,
C.. .:.. . ;} *' *. )- f**= * ' :* . : .:.*. :
~*- ..
. e *
'g *.
- : ... REACTOR ; * . . . ./. .
*l*
'.'. */. VESSEL .
1
. .u - ::- . .
ORvwEu.
>g l :,
' *% . s ., .
[ G, . .'
,y g<. . '
4 l t*..., 4 [. 'y
. i..., .
- f. e :-: + :.
--:e.
.. _ _-, _ - .: "r. , D..
. _ 3:_x-y - < .< -
._ -_-_ - : - r_
- - SUPPRES$10N
__1 :
- - -- - - :Ow -
- >... :. , ~ +::m. :=>. POOL
--_-c s*: . .a ... --
... w ....'"l ,=- .. -
^-^--t.
- p. -:+. :. ..1 g . . . t'.,:.:*. . . . . . .**'. . . . .. . . ... . ...
...... . . ...........m:..
. , . , ..~ . . .
. . .....:. ...-.......... ... ... . .4.- ..........
. y ..g...... ..-
.e ...... . .. : . :.-...
I ! FIGURE 4.1 BWR MARK 111 CONTAINMENT DESIGN l
\r
_ .#., .,. w. ..e.. + . w p * -s =-
,m.,'...
.., .. + . . . . . , . . . . . - , . , , . _ . . , . . , . , , , . - , . , . . . _ . - ,_ _ _ _ . , . , .
j p-Q. A'; L ( SEQUENCE IC -- IRANSIENT~WITH FAILURE TO. SCRAM 1 i TIME (MIN) CONTAINMENT FAILURE 80 4 BEGIN CORE MELT 118
- PRESSURE VESSEL FAILURE 196
-CONTAINMENT FAILURE MODE - 7' 4
[
- p1 ;
i 4 l t e 1 f
. . , y . -, , , . , .- c - , ,
i ( l l i
,._s
- i. ,
i i +
~ # " !
Steem Line esm Dryers ,/ ! N l i L
* - l
~
, km l I
- Seperstors , l l
~~
r
- , ?
l
- p Core i' 1 -
FIGURE 6.4 FLOWPATHS FOR FISSION PRODUCT TRANSPORT IN RCS - SE0VENCES TC, TQUV, AND TPI t
..e._..,,_. -., ... _,, -.-. ..-, , y _.,,_c..._.-m.- ,_,
O a UPPER STEAM DRYERS OUTER , , , , , _ . . . , ANNULUS S 6 LOWER PIPES / SEPARATORS OUTER STEAM LINE ANNULUS 8 4 7 RELIEF LINES 9 SHROUD HEAD 3
' TOP GUIDE 2
POOL CORE l () FIGURE 6.5 SCHEMATIC OF CONTROL VOLUMES FOR THE GRAND GULF SEQUENCES
l l j f 2800 2600 - 2400 - 2200 - 2000 t b 4. L 1800 - e B s CJ
' E
- " =
N g 1400, e 1200
+
1000 +'*~+
,j +'* + f{
800 -
- , 1 2 : : -
60 .,. ,, ;__ =- -- ;; _ , 400 I .I I I I I I 0 600 1200 1800 2400- 3000 3600 4200 4800 Time (seconds) qi FIGURE 6.6 GAS TEMPERATURES IN RCS VOLUMES - SEQUENCE TC
t
\' 2400 -
2200 - h 2000 -- _. 1800 - _ 1600' o'
$1400 -
t h
+
s "u 1200 - l --
+
4 8 y o +N . U 1000 - 2 < M + 800 - 6 0__ g ,,' O ? ,-
. _ _ r_==r
. u t _"+'
(( -
/
-6 400 -
200 0 600 1200 1800 2400 3000 3600 4200 4800 Time (seconds) f- FIGURE 6.7 STRUCTURE TEMPERATURES IN RCS VOLUMES - SEQUENCE TC (.,' P-
i r iJ
. e...
CUTER CCt TAlt:P.tfriT
/.
- E i:
p. (. 3 T .
,' .),
.=
l .* 7.
-,s I.
- l*
E
,' f __
. , . ~,' i<
d,
. . .:4e'; <
.r3 _1.,., ., f .
*f. ;..... ...
g .
} "". " * ...
0:: .
.l. f " ^ -. g
~'?l:../ $ C. :
- .-)
.: f .RCACTGR
........... [ '.. L .. . .
.;.. ~ < *
~
*/.;4 VCS;EL 5...
.. . u. - .
- v i. : >.
.Y...
,. m .::
. .. :.. .?? *1< *
. . .J.,
*.,*I, :. ,
...a: * .
.%.n.. *
- f r.
;,. . n vwrtt
- s}
.; * + n L.':; l e .". P. * .e.'.
.<
- L :. . . . . .; - ky.- * ~ '
_ _ .v: , .f- -
- .- c ,-- b
- :3 /
.;;.Q 1
. :] v T.:*pu L,.,,,. .4.;.
,. w.\.t.d ,.
. q ~~ )
; g :..;;, <
I
.;. 1:._n : : ~: =.... P*.
- 2. - m
.<~_ j~ '/.; *, SUPPGCC5 TON 4 .-
} .td'~;C:~.2
.~ ~~c'-.- <. *.
< ML~
=gm .v,T. .". **. :, .k. . ..W.'.
p.OO L
=~-q
.~
. . t ; * *. ... . ..av . .u,.. .: :.:.- t t ;. .C . . .~ <. .
w..i + w .-
. / .
w *. :4.
...f.*......1.;...........
... *. ..,.,.,.................-...s.......,...,....**..*..m.
.. .... .... ... .....,...f.
,. ., u.
. E
.a f
FIGURE 6.8 FLOWPATHS FROM THE VESSEL AND REACTOR CAVITY (;/ TO THE ENVIRONMENT
(] g ^} Y SEQUENCE TPI -- TRANSIENT WITH STUCK OPEN RELIEF VALVE AND LOSS OF DECAY llEAT REMOVAL TIME (MIN). CONTAINMENT FAILURE 1323
!_ BEGIN CORE MELT 1635 i . PRESSURE VESSEL FAILURE 1953 t
FAILURE MODE - 7' i t ,g, , it ( 4 0
- # l'
n.. ... . .. ...u., .... --...,.,...m-.u..u....... .
) C
- l GRAND GULF TPI- -
80.0 4 DRYWELL
........... WETWELL i i .. ,..i i.. i n. iso il '?0.0-
- l lI 60.0-s in i k 50.0-
. Erl l 40.0-J M i 4 30.0 - ['s 20.0-
%_ L 10.0 , , , i i i , i i ' '
- 0.0 200.0 400.0 600.0 800.0 1000.0 1200.011400.0 1600.0 18b.0 2000.0 22k)0.0 2400.0 2600.0 TIME - (MINUTE)
J FIGURE 6.10 PRESSURES IN CONTAINHENT VOLUMES - SEQUENCE TPI
;n. \ . .-
4500 _ TR0(R,L) where R = Radial zone number L = Distance above core bottom, ft 4000 - - ji TR0(1,6.5) -
- 2000 3500- - TR0(5,6. 5)- '
3000 - 1500
- u. o 2500 -
E - U R 1 5 E ( 15y 2000 W
--TR0 (1,6. 5 ) b 1000 #
1500 -
-TR0(5,6.5)
TR0(9,12) 1000 -
'TR0(1,1) - 500
~~'
TR0(9,12) [ 500 k j -
.O I I I I I I I I O
O 12 24 36 48 60 72 84 96 108 Time, seconds FIGURE 6.11 TEMPERATURES IN SELECTED FUEL REGIONS AS A FUNCTION
/ OF TIME - SEQUENCE TPI
(/ w-h+
4 ( , 2800 - ~ 2600 - 2400 - - 2200 - ~ 2000 - C 0 +
- 1800 -
L' k' 3 s, b1600 - j $ 2 ( s 4 1400 e 1200
~
t&
++p+ + /* __
1000
\
* ~
i.
. / . .
600, , 2 m- 1 v 400 ! ! P O 1000 2000 3000 4000 5000 6000 7000 8000 9000 Time (seconds) I
/
FIGURE 6.12 GAS TEMPERATURES IN RCS VOLUMES - SEQUENCE TPI n
.y. m mm e -mm^ r.N . "' **
O
,e'*-
- . . . we-** ***
2000 - 1800 1600 - 1400 C E 1200 - 5 f ,+ .i.
, e C f1000< j \
~ '
j ,,, _ Q NN_ .-- ._ .4._ . _., e g .- i . -_L... 600'1BM M E x _ ======== 5W- " 400 - _---
~
200 - I I I I i - l 0 0 3000 6000 9000 12000 15000 18000 21000 l Time (seconds)
./ FIGURE 6.13 STRUCTURE TEMPERATURES IN RCS VOLUMES -
SEQUENCE TPI
(') Qs. r) a 15 SEQUENCE TOUV -- TRANSIENT WITH LOSS OF ALL COOLANT MAKEUP-
- i d
- l TIME (MIN)
CONTAINMENT FAILURE 96 BEGIN CORE MELT-- 83 ? PRESSURE VESSEL FAILURE . 2fl0 4 . _ _ ._ 4 FAILURE MODE - 7' (HYDROGEN BURN) e i o I 3l $ 1 t 1 i 2 ll It'
..-, , s . . _ . .
C'i O O o i GRAND . GULF- TQUV , 2500.0 DRYWELL
.......... WETWELL l
- it < t
- e nt ts - f 1 , .t s 2000.0- .
d g t . , .M 1500.0-
*a q , b k
1000.0 - , Y
- 500.0-- [l 1
2 l l l l l l 5 l 0.0 100.0 200.0 300.0 400.0 500.0 800.0 700.0 800.0 900.0 TIME - (MINUTE) FIGURE 6.14 GAS TEMPERATURES IN CONTAINNENT VOLUMES - SEQUENCE TQUV s. i h n
.,.4,, . - . - __ A c- o..- a, GRAXD GULF TQUV .-
80.0 . DRYWELL
.......... WETWELL 70.0- ,,,,.,,. ., ,, , , ,
80.0- 4 <3 4 m N k 50.0-m g -i b' m 40.0-N M (1 30.0-l 'II - 20.0 - [ ,1_ ^ 10.0 , i e i i i i i 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 TIME - (MINUTE) FIGURE 6.15 PRESSURES IN CONTAINMENT VOLUMES - SEQUENCE TQUV 1 s
s,j
-- - - 4500-- -- -- . . . . - . . .
TR0(R,L) where R = Radial zone number ' _ L = Distance above core , bottom,-ft -- 4000 - _ 2000 3500 - m
"^
C . E. O 3000 - 8 m -
~
i D, 1500 Ex ~2500 -- h _
?
) # # J s a w &
g 2000- - g W - 5 TR0(1,6.5) 1000 1500 ( R0(5,6.5)
~
1000-500 TR0(1,1): J
-500 TR0(v,ia.oj. = - - .._
- 0. I I l' I I l- I I 0
0 1000 2000 3000 4000. 5000 :6000 7000 8000 9000
-(7_ Time, seconds Qi FIGURE 6.16 .TEMPERATI..RES OF. SELECTED FUEL REGIONS AS A FUNCTION
-- 0F. TIME - SEQUENCE TQUV
,n, e
~ ~ ~~
. . . . . . . . . _ i f
3360 - 3120 - 2880 . 2640 L 2400 - v, 2160 O
~
1920
\ :- ^ *qM a 7X E X X X~X X a X 2
- -^ ^ ^ ^ ^ ^ ^ ^ ^
1680 X 1440 - 2 - - - - - - - - = ; ; ;; j$^g O O ._ 1200,, f-960 -
)
720 I I I I I I I I 480 (~ L/ 0 1000 2000 3000 4000 5000 Time (seconds) 6000 7000 8000 9000
- FIGURE 6.18 STRUCTURE TEMPERATURES IN RCS VOLUMES - SEQUENCE TQUV
l l RCS TRANSPORT AND DEPOSITION FOR THE-GRAND GULF SEQUENCES (TC, TQUV, TPI) 1 e RCS THERMAL-HYDRAULIC CHARACTERISTICS
.e RELEASE FROM CORE :; '
- O e RETENTION IN/ RELEASE FROM RCS e EMITTED PARTICLE SIZE' 1
r OBallelle v( c , _ e . o se,...,,,, y
~
- - - - ._ _ . = . . - . _-
( 0 )
, / )
RCS EVENT
SUMMARY
FOR GRAND GULF SEQUENCES CORE MELT VESS[L VESSEL COVERY START DRY UUT FAILURE TC 88 118 190 196 FLUSH OF RCS AT 168 (PARTIAL) TOUV 46 83 123* 240 llIGH FLOWS TILL 126 TPI 1525 1635 1792* 1953 FLUSH OF RCS AT 1752 S h J OBallelle i ( , ,..
#,_s.,
,, g
.se,.. .,
y.
Time (s) e 1000 2000 3000- 4000 5000 i 1._ 1000 - ,.c--
... , ,/.
g. i
. i , . .v ,
s . ., , c100 -
~~
s4 ea4 / t - t e- e r .f : r.*: : . s .: , ,r; i : ( - ia4 4 i 5 , e
. r.r i j [. :. 4. .
; a.]
4
- t
.',.L ,= i,aiavi r. t.:. :. J :. , l : j .r ..J : : . . W.i L '
- .l ; a $ : .. 3.a .
t1 L f .- .: =.;/i .h:i .:: r a t . it.4 ! -- I .. r. .:n.; .:::.: : 1.. i; :
;::.:..:: : .r ;. .
. := 5. i: :-:. : .: --T 2..
t 's - 4:. e _- i 2 r I,,,,j I i u .. . ..
. < .. . . . u. .. , .,
. --=r=- , . . ,. - , ,, . . . . . . , . . -
.L T., .y _m. ~ e t
.a r a_ a . 4 < . , J ; a .. ,a.a .ia r _a -? .f r i d4 i s.. s1 et
^
%r-
- - - - - 7:= .h_ls. T .QU. V.G ---
* :-t- ~==V i a M-t + + =-
m.rs=1 m n ' . w. . . . ___ D _=* w =:."
=
'"..,_*7.*. ... = __...__
. _ r1I C. _ .
.g.
_._..**.C4..**,_ "_
* * " * " * * ~
-*'_l*7..*.'****.
. . _ _1"._=o." .".Y.~._"*."_"._.O**.*5.-
. . .. _. _r I_=Z___
_-__'**I*..d.*.=C_..*.*..
.=. .._ .v.c= - . __., _ .-
. .. -_-h.___.,_.-. i .
. _ . . = . ._.__. _._._
r.- --!
,f..,m'>
l _ i i
-100- r .,m.c . . , _. .,_,,,. , , .d 1 - ,w J..: ,. . 10.0 -
---- t .m.. n, -_ = . , ,. ._,.u,,,,, .g.,,
w# _y.f. w.o
-"J :. Jil - s_i . _z_ i- -L -
3;i.r z ; a. 6 4
-if ['.. .i s ti- v.-i jz ? y? t _at M Q-?
i t - t . :- :- id
+5^rm. = =k= - . - = == =r.;-h
. .,-+v I--:--"3-- =-F== L nEm 4 c==-n ~~' ~ '"
=__:-, = _:_- _ =_ _ =. .___:. i :=__ - TC A =.==-m=_-- --r._~.:__-"
. . -^ --
T~-_
--==1====M.=. -
r, n . 1,,.,. . -- rc_ . . . g"*r. .e-w . .4 + a-2. m , : . ..,.m
.n . c n>
c_ m.m .. .u r.. ,. . ,. . - . .- , ,, -w_m m w +c - - g r_;ps+-=- :=19 rrs == = 3 , ., w-av-a_I=_1r_8 +.-w
= w -m u.. ,.m i + _=: -o ===~: w :a-:.1-=+m
- = ,_,_ c_ .= r _ =5
= = = , _ - - - _ .-. d- a_=. r _=. . .=-=r- :_:-. :.= . = ~
=:..--- -
- ~ =_=.==.=-L.,. . _ =- - : :1: _
a:. =.=-_=:_ =.a
- .===-
z..\_ . . . - -- ,,__-
.:. =.
.. . _ _ _ =__. . _. .-
-n.
- m_
- : - o y
f
] _ _ _ .
._.I w,.'.. --_ > >
m 10=0 u., . m., -w---. 1.0 cr
~m -
r.. __...w., n .nmn-w a g .y. i a . , . ,, -9.u . . n <_ .m..,.u_._ .s
. . ur. . . n . .! . .- m ,,-.
s._ w .- m u:-:s.v ..u . s w
==5f- , f.i - c= _.=. - i ._ i . p. r_ r.2._.2 n. 4_ w r 2.mr .+ g
' .t .:_.w;_ g_l :=_: a e = r m r.,3+-? r-g _
EEE "-=.1W-- F=
.'2 3- _--- ll ii iJ:"_--^e' 1 i.
% !~-K= in ' # "-{+ =~~;ifi:U.M=-WM .
_._ -r _t _.:. .- 7 : : u \ u: = =.. :
,.:-.:,=.=__,..
n:._ . .. _ - s - _- : =, =_ . _:_y .=_ __=
= . ;
._ .. _ _n _ _=--
_a5 y t t . - ---
- a. '.._r. v .- m : , .m.. z. n .: : . .-(._
n i+- m je -# _., -m ~.Mm - --=- w.3 .a-=
-q. :m .m. x. .,2.r .m .. ; e i .a. ; . , _: _ m z.,_ p_.y. s (..:. m.s.
&-=- * :.^'_.'Jh.#itts +==nV?*:9= .as- -
u_--?-%3z 'E . . h h-^E&E =-: - _-s& == #+-~- a - &2X _ 25
=-r N :: =.W tc-EP=d =. P\=:fvC=i -F==:= _
=':~r :-fC - . =-i:'~- ~. -' .'.=-' =_= :M '- _- .:
==%;f '~~1- __.. ~~ #3::f-5-0 .p=--T = % -t. t---Ti:-- --E- = :-- --- =:=*-M ": F- _ ~ =~
t =t.:- ---+ . _ - _.\ u r::= __. - - ::::.=: .:r. = - - ---
- \.=
_--- \ _ . _. . _
.l-- . . . ._ ._. _
- m_ \ .
-t r- L i
+
- l. -
4
; i 1.0 . w . . g - .- - - .,r
.- w.- % g .
.u' ,.. - - , .= u ; e.s.m.-c- , _ - - ,..a . # O.1 F,s_i,_,_. _.\ .
, r 4 e i . .r,4-e i. e .:. 1.- 4
- 4. 4 . r ta. e.
. p ..- 1 - : - 9 . r. , =
. liu. . -i. _ g % a -s-sE :. r.t X i.g _t 4 e 4._ m a (.u.-l y \ h c.- u.r s _ . -: ,_s> 4 . ..- h;_h . t - . < . _
=f. a NN I-i - *- 5-* -- M . *.N
- WI. b ; 8.I .. _\_ -- C* d 'A I b!i I*bi 4:5 " ~ N 5*h.N : d*-
. : FEi =$
._= :;;E:p:E.H . - '-
- r. ~"+= :-g.t.i - .
T.x " =. q. . ~ _ : i:.! .:-i C.~:._ i - i -[5.;.; :se.ES-E)
'x1.I
~ , , . 6 . , , .. :
. s.i .1 ,
] .s-i e >.*s s
{- . t t
.\ . a t 4 -
- r$ a t t t .- i .t t a . .*z z i
-U .8 '5 t J- .5.?- --d bt* 4M . t- E -
3- \ -I -M i i - d .4 4. . b r ..*... _.4 5.7- A i- 4 e 1 4-.- _r-4 r--- - ...
+ :. \ n r 2 = i4 =- s- --
i r $ fe rd i a .4 a F.:.i r_:.: -12.: m== =o-.r. : i . =i =i
=g:p'Fj.:..E R .m .Pa==._y :. . .
. F i:E. N %- :. .q ;-i w .#f-y?-@:liL .: q : i_
"==
..r______.
. . q.s _..:_. . . :... :. .;.:. ..::
.r-
.. ..-s _.=: c : .. :.1.
- 3. --
r _--. . . .. .-., .. w _. .,-
. . . .___ .._ s __. . . - . _ . .
--0.1 , ,
c , 0.01 0 2000 4000 6000 8000 10000 RCS FLOW RATES FOR GRAND GULF SEQUENCES TC AND TQUV l
, , . . _ . ,,,c - . . , - -~
T~ CORSOR PREDICTIO!S OF MASSES OF SPECIES RELEASED FROM THE CORE (TOTAL) AND TRAP-MELT PREDICTIONS OF MASSES RETAINED IN THE RCS (RET) DURING THE TOUV SEQUENCE FOR THE. GRAND GULF PLANT Csi Cs0H TE AER0s0L , TIME- RET IOTAL RET TOTAL RET IOTAL RET TOTAL (S) (KG) (KG) (KG) (KG) (KG) (KG) (KG) (KG) 500- 0.7 8.2 5.8 56.5 0.4 .5 4.3 57.1 1000 2.0 19.6 48.4 .115 1.8 1.8 38.2 .219 1500 1.4 23.3 53.8 149 4.1 4.2. 55.8 341 C 2000 1.4 23.6 54.1 150 6.1 6.2 57.3 385 3000 0.8 23.6 50.8 151 6.4 6.4 58.4 398 3990 0.2 23.7 47.5 151 6.4 6.4 58.8 399 4980 - 24.3 49.1 155 6.5 6.5 61.2 403 5980 0.0 26.5 53.5 169 6.7 7.0 72.4 420 6980 - 31.4 62.9 198 7.0 7.9 121 482 7970 0.1 35.6 76.0 222 7.6 9.7 284 660 8970 0.2 36.3 86.6 227 8.3 11.8 530 906 9470 0.2 36.3 90.4 227 8.7 13.0 655 1030 c- OBattelle ce m . , u se,.. ,,,, y
73 L-TABLE . - CORSOR PREDICTIONS OF PERCENT OF INVENTORY EMITTED BY CORE PRIOR TO VESSEL DRYOUT FOR THE THREE ACCI-DENT SEQUENCES FOR GRAND GULF
.. Inventory Species- TQUV TC :TPI (kg)
Xe 0.66 1.00- 0.97 412 Kr 0.66 1.00 0.97 27 I 0.66 1.00 0.97 18 Cs 0.67 1.00- 0. 97 220 , Te 0.17 0.27 0.40 37-m
~
! 'I Sr 0.03 - 0.10 0.16 67 Ba 0.08 0.23 0.34 112 Ru -- 0.01- 'O.02 -183 Mo 0.04 0.12 0.19 252 Zr(FP) -- -- -- 284 UO I) -- -- --
.147000 2
SnI *) 0.24 0.58 0.66 1190 I Zr")C1ad - -- --
'78200 Fe(a) 0.01 0.02- 0.04 9410
.(a) Nonfission product species.
i ( b, - m
~.
F'- 13 TABLE .. - CORSOR PREDICTIONS OF PERCENT OF INVENTORY EMITTED BY CORE PRIOR TO VESSEL FAILURE FOR THE THREE ACCI-DENT SEQUENCES FOR GRAND GULF
. . . ...=___ .; - iei . _ . . . . . . . .
Core Inventory
- Speci es_ ' TQUV TC TPI (kg)
Xe 1.00 1.00: '1.00 412 Kr 1.00 -1.00 1.00 27.3 I 1.00 1.00 1.00 18 Cs- 1.00 1.00 1.00 220 _-- Te 'O.34 0.28 0.53 37 Or (/ Sr 0.08 0.10- 0.20 67 Ba 0.23 0.24 0.46 .112 Ru 0.01 0. 01 0.02 183 Mo 0.16 0.13 0.31 252 Zr(FP) -- -- --
'284 UO 2
() -- -- -- 147000 SnI ") 0.59 0.59 0.82 1190 Zr(*) Clad -- -- -- 78200 , Fe(a) 0.02. 0.02 0.05 9410
. (a) Nonffssion product species.
,.+w~w.. ,w=*,s=++a. . + p-h -h- e**M=sr e g+*EN,'
T ' ' ' * " " * * " ' * " " '
4
)
MASSES (xa) 0F SPECIES RELEASED AT
. VESSEL DRY OUT.AND VESSEL FAILURE : .
FOR GRAND GULF SEQUENCES : . TC TQUV TPI r Cs 186/220 . 148/220 213/220 ' 4 I 14.9/17.7 11.7/17.7 17.1/17.7
-- TE 8.0/10.5 6.4/12.8 14.7/19.5 ; . _. @ r F.P. - 46.8/69. 19.1/73.6 100/148 AEa'o h Non F.P. 776/1010 379/966 1286/1665 AERO o
i OBattelle C( c . ., u e.,...,,,, y . g e-a%+ ... y y. M WWL he-O**i6 e .gepA% is- , V
- e 1,% 4 - -*w-
, , . ., . . . . . - ,,, ,, ..... _ .c . + - , - - . . . - - - ,
CORSOR PREDICTIONS OF MASSES OF SPECIES RELEASED FROM THE CORE AND TRAP-MELT PREDICTIONS OF MASSES RETAINED IN THE RCS DURING THE TC SEQUENCE FOR THE GRAND GULF PLANT (TIMES: MEASURED FROM START OF CORE MELTING) CsI Cs0H TE __ AEROSOL ._
. TIME RET IOTAL RET TOTAL RET IOTAL --RET ~ TOTAL
^
(S) (KG) (KG) (KG) (KG) (KG) (KG) (KG) (KG) 200 - 2.9 - 22.9 - 0.1 - 12.2 610 0.2 7.1 1.7 51.0 0.1 0.5 19.9 58.0 1020 .1.1 11.5 8.2 76.8 0.2 1.2 88.1 .37 1420 2.1 15.4 15.1 102 0.3 2.1 187 241 . .- g ~1830 3.1 18.8 21.7 121 0.5 3.3 307 363 2240 3.8 21.4 26.6 139 0.6 4.6 437 498 2640 4.5 23.9 30.9 254 0.8 6.1 571 633 2850 5.8 25.5 39.8 163 3.5 6.8 630 704 3750 10 32.4 111 205 8.2 8.3 741 858 3660 13.5 35.4 114 223 9.3 9. 4' 755 984 4060 13.5 35.9 113 226 9.8 9.9 761' 1040 4 OBaHelle ce, e., u se,,,e,,,, y
.. ,, ... 1 ,,
a O.
- s. . .. <v*,, .p y. i . ,
(*) TRAP-MELT PREDICTIONS OF PRIMARY SYSTEM RETENTION FACTORS-(RF)'AND VOLUME . SPECIFICRETENTIONFAQTORSASFUNCTIONSOFTIMEFORTHE.TCSEQUENCEFOR- .,
.,, . , -THE GRAND GULF PLANT , .
c. , (TIMES NEASURED FROM START OF CORE MELTING) 4 CsI Cs0H TE- AEROSOL TIME STEAM- STEAM STEAM STEAM STEAM STEAM
, (s) RF SEP DRYERS RF SEP DRYERS RF- SEP RF- . CORE ~ SEP I 200 -- -- -- -- -- --
.08 .08 -- -- --
610 .03 .03 --
.03 .03- --
.12 - .12 .34 .32 .02
{ 1020 .10 .09 .01 - .11 .10 .01 .17 .17 .64 .60 .04 1420 .14 .12 .01 .15 .13 .01 _ _ _ 14_ _- . 14- .78 .73 .04 1830 .16 .14 . : 02 - - .18 ~~. 16 .02 .15 .15 .85 .80- .04 i
- 2240 .18 .16 .02., ,, 19 .17 ,03 ,.,, .13 .J3 ,.,,88,,,, . 84 .03 t 2640 .19 .17 .02 . .20 .18 .02 rit .
.13 .13r i 990. '. : .87 .03 2850 .23 .21 .02 ~ ". 24 - .22 .02 .51 .50- .89 .86 .03 j,
- 3250
.31 0.0 .24 .54 .22 .
.25"' .99 .96 .86 .71 .11 !
3660 .38 --
.28 .51 .09 .33- .99 .96 .77 .62 .10 4060 .38 --
.27 .50 .09- .32 .99 .94 .73 .59 .09 l ii. ,.,i .
.n
[ e . . . ,
. i ::
O O. ' '~ G u- t :s . TRAP-MELT PREDICTIONS OF PRIMARY SYSTEM RETENTION FACTORS (RF) AND. VOLUME ,n SPECIFIC RETENTION FACTORS AS FUNCTIONS OF TIME FOR THE TOUV SEQUENCE FOR THE 6 RAND GULF RLANT. (TIMES MEASURED FROM START OF CORE MELTING) 4 -
- s .
CSI CSOH TE AEROSOL TIME STEAM STEAM STEAM STEAM n~~ (s) RF DRYERS RF SEP DRYERS RF SEP RF- CORE
. .500 .09 .01 .10 .05 .04 .82 .80 .08 .02 1000 .10 .06 - .42 .08 .22 .98 .S0 .17 .08 1500 .06 .06 .36 .07 .23 .93 .80- .16- .10 2000 .06- .06 .36 .07 .22 .93 .81 .15- .09 3000 .03 .03 .34 .07- .23 .99 .81 .15 .09 4000 .01 .01' ! . 31 .07- 1 t23* .99 't.81 m'I' .il5 m .09 d
6000 -- --
.32 .08 .52 .96 '" . 79' ' '"' "' '17 " .12
' ,1 8000 -- --
.34 .14 .19 .78 .66 .43- .39 9470 .01 --
.40 .19 ' ' ' 20'
. .67 .58 .64 .60 4
k
.i i. ..
'a l'l . l : . lI -
1 h -
'h 1
-. , , r,,- - - , -
. . - 1
.1000 -
.. 1
- p. _
x-- : a
' l
' Aerosol i
. I 4 m i
i n
~ '
Cs0H -
' ' e' '
---- l'
- - -100- ,
r
~
/ 1
,, t a
MD k , e a J f \
.. CsU .
SO ) en 84 n 10 - .. s
/. ?
mm Y
- I e
- > l A
h f'# 5 R a (
) l
,-= 6 '( /
0 1 2 3 4 5 61 ' 7 8 , 9 10 {" 3 t(10 s) i *
/J" . .
MASSES INJECTED INTO SUPPRESSION POOL 'FOR GRAND GULF TdVV 'g.
.( '
- ,r
- , , . +-,--,,.c. -, . , - , , - ~ - -.,w-- e-, - - - g-- --y n .s- y n , - --
- .-. +a > ~ . - -~ -
I 1060-- - - - - . f 't m 4 Aerosol - M .j Cs0H 100- r- -- m == M M _f~~'*- CsI . Q -
=
10- _ k mee m BAE t (103 s) MASSES INJECTED INTO SUPPRESSION POOL FOR GRAND GULF TOUV l
, - - , .- ,- - ,m.
B m E I B
.=
1000 - _ . - f I_e [ d . -I - - ~ _ 2 1 - ILM . I LA-- - 1.J v*I - i . : m'- 8 m, . 3 . + : = . t_ a . . , w; % . ;n m. < w H-.rd.. , _ e~ r Ts. ---* * *-
. .- n - ? . _ . -1rt y T- =ry 8 - -
i a,a
. i :-i-5 f r 1 4 t 3 [ t iI - t ,
--, ii g , 1 1 v u
- r. . 1
- _. i 4q 4 ,, g . 3 g, l , r, ,- _- y i _ [ ; ; 3_ -
y,3 h Ni~--r - mii m. nrt .2_m.~ ~m_ t.2-2 t -m s e i b a -6L. L 1E- 'Nk -r t*4'mi- * ~~m -*.1 Y s-mu1m am a. esm s
.mw m #+ nre m'6 m i ea.r:=mu = cx_z- m _ w: wa- wm .s+,-u,; e , msm v
< we T, 1 tmm L_ .=__--._._
_ = _ _ _ _._. w- y __ _ _. _ __ _ _ _ . _ _____-w.- g-.. __ _ _ _ _ _ . __ . __w f_ n - = _ . - =.= ._ .- .
+- v
=__ _ _ _ _ _ _ _ __ _l- - -- ., - . . .
.}-..-m--. p. - . . . _ + . c.---}------- . f' . -~
. TOO < .. - -m - - m, -_ - - - .. m ,---- , 1 ,, , -m -om 10.0 - - - ' U I' ' .N,4E' A f 4__- 2 - ' + 1-- _ n_l . _ ? ' MI *f - * 'I l E I '- F-
.- I -I' S s I '* 'f 6*ECr .[ Up f -- p p f E-++.-_mg-
- __1 Z g
.j wru.p -+-f7 ).5 7 s T ) j MTi -W-11Q
-"JE E F
- f pC b 1 n J _
d '-'
'l 'eit {_C
_E P'- ".' - TM.-VOEi-tEF F-Thra'b ; - - -~in'* .'s&Jfh=S ?'M '- N WWP O ii.IT"E FEf g '." E T T '- U _I--- Eh
.._- N EEA _N.._'- - . .$ N~1~.M- . I _ ~ - E- b c '
g r- m mmr .c - _..'Al_$ r m s --u , -- r cx . _:-N [.S ^ ,u< t r -- = - = 1 v --
~ rut- .
.2%m e a__<1 > ; "Y wax _u;, -m., m. mc us.;- u + : .iw a. - -: u*'n 4 I; IJ i La s +** l'I'm u
'h bE* NL fN}
--MM_.' I 'C J._mW uT d d-~[_M.--M..UN hI b.I' E pM M Ml 1N f"U -
N -f M aa - M { I[d 2N,,. _h^ k k [. j W M_r .W Sz _. jig _El_-'= ++^-P..T- _? f 0
- WL =G E
_ .. ' _SSifF#2-5FM. N=i Lt.-*4 W3W5_.I.'
'fl E
,a, = =__ .u..-titF
__ . =_ - u . . = c F.- E c 10.0 __ . _nn_- m m. -,m_.v 1.0
. . ,,m
~
a
..m..
.. . ,._ - ..1 < - ~m
,-= ,
.wm= -I**4LJ M- & '+-5 3? MM:-Fllr-M M =a '-4, N Ei- ' PMrh . P #5_ u e +d -* @I~s mm 4 MrT 'mm m . 93 *
'5--
-* '.-t 1
__, _ _ m m- _-4 M d %_i T t _: m , _ 7
.. E'
- w.
-E'- _7_ _' _ _ ' [._ _
$ Y-- . _hN 2 "IM .[_
_ i . 1~. 1 n~w s a w_ e . a ww a ,uu , a r <r , r: , _r n,<w_, s.u c <- E ( C w
~ OG_nM
.~! __,=- ~_ r' Z .w.
=
- -_"-~~Z~Z _ .
=- = 1%WJOGi,% J^=-=~. =': n
- _~.--_: ' .
. .~
J_ ! ? ; v _ . MMRWn 1.=~=J.~=~^1-~ 31MiElw_- " _- S6 x-em'T - ~
.-' : '._1.-^'__'.; -~_ .-
- = _
=-=m,-__--- = J :'
-=
= z, -_.n ._ __. _ _ _ . - , __ _ _ = g - z , Cs1 _ u w a - m n - c 1.0
~
, - r- -
= , - - - - - -<4m--- --m 0.1 E
E
- h MmW GWiMWilh4_OiMR
- M'M M - 1mm q#9 q
.-~ _ s..__s e
*_ z_ ._- a._ ,____._n .
--=-.r._.-_ w m
,p.~.__-. _
.. _; .-m..,. N s=v__ 6 e_ + i-t m+=. = _w
[ g _ .
- -.n, ,- - - - - _ _-r. _ _m = e, m , ____.=r,
--1 _ . ,,u ,_ - > 1 w u- a e _
t s. n g_; s + _. ><u1 . _._ ._rt , m m--m.- uw _
;12 w2 .
_ ; . .x m m v_m_ em1 _._1 v 11.; u I_+. -4_.4 h*m+ ' WrE--Jw U- E4NMMC . a5 riE d 'id L.r^= -i pwJ M-i 4i
- UMfNE'T~-- " ~~ FM 1r
- m-E sr*s ~tN
==a=aa ; --_wwnr.-.ww -m ~=rmu=
m ==
. - - = -+ mm W = == --
=-
m rw=w - ~ - =
== : =.=-m wx--
c- - _= g E
.. Aerosol
= 0.1 ' ' - - ~,- n ~ _- ~_ m :__., = nn . _ s~ ,- - - -
- n,r . . , _, 0. 01
- n .._ . . . ._ . ~
q --. M4-2-^-iW izMV T C=^ w wmwsm=====er
-_;n -
L _.h - % *_ :i l. 3Fw_a=& W W4--L m
. =n"- :: :
e w rw' %%w a
*t. "? t -m. s !i i M a =, u .~ - -; .- -
at ~iIf 5-?--N ^
.= w = n -=n" =I 25E r :'4
--w -.=;
a
& :u= - .r=. ;, u_.=.n5-uy9 u e g -= r > 1 - au .n euiir _ _ . =. w . u; w;2ye;__ - _
-t ,____:
. . * ,r _. r : ..1. .
._r , m : _i
- m. _.i,.
~mm .
.__: . _, _m ___t _
*E .
' *' d-* ""
Y I
=
T d} +- as' . [ N i N. @__ N- Y N"Y N.i i I I
-+ e(
5 -
- b ES b '.D [U' IIbN N E N[ -EY [L S N I Ni~k E
r g .: -
=, a
=;_m a= .: =-
=m. m* &-e =sm
._..,M.- __ _ m= --~rs _= _
. _ = _ = _ - - _ = _ __ ._ _ _ - , w : -- ===.
EE"- E . 4 E ' * - - - - - - g 0.01 i z - 0.001 ( 1000 200u 3000 4000
) Time (s)
E TRAP-MELT PREDICTIONS OF MASS INJECTED INTO THE CONTAINMENT SY SPECIES AS A FUNCTION OF TIME FOR THE TC SEQUENCE FOR THE GRAND GULF PLANT E
m I 3.0
; i
._ _. ,6 i ! ,
_ l ; i
+
. ; f I ! .
i I i !.__' f i r i i-l l j i
'i . l l l
. _ . . _ -l * **
- l i, .
i i i i ! l
. - _ __7__ ,: . __ ; ; n I !
i - i p-7 ,
. I . . I i
! ! I i ! '
l l i i ! l i 8 i ;
! l i i ! i i i :
. ! ; i
. !i.I i ! I : l
! l l i i l .
2.0 - - - i ' i l 1 l I
.i '-l i
l
- i. :
l i i l i. i. 1 l , I i i r I l i ;
- s. ,
e : l ; . ..._~*-. .. 1 l . -
= ..
E- . . . . . ..e .. t _ i j i l l ! s lzf - l : J E- RELIEF LINE i i l ! l l l i I c E
^
so o ! l .
! l .ig I I
- s. _.. =_
i .i m u , . . fg l . : ! i 1 i i l l 1 . vi . I ; [ i
. . f i .
g z . ; i ! 1.0 -- .- , . , . .
~'
i . I l I i ! i '
! !I l' l i i i '
. i ! l l l . !
! I. I .
li
. i
' l
.. ~
CORE -+
- i. j .
r.
. _ a _._i _
I i ! i ! i ! l ! ! i i . >
. i
...___.:.__._3_.....j_.;_.!.
, i :
..._..i ..._ _. ;.
f f f a t
. i .f , . . .a ,3 ,3 1000 .2000 3000 4000 5000 6000 7000 8000 9000 l
Time (s) , TRAP MELT PREDICTI0flS OF MASS MEDIAN DIAMETER OF SUSPENDED PARTICLES AS A FUNCTION OF TIME FOR THE CORE Atl0 RELIEF TIl1E FOR THE TQUV SEQUENCE FOR THE GRAND GULF PLANT
- I:
1 l l 7, m . t- _......_ t. _ _ _ . . . . . . . . . . . . _ _ . _..-...L.__.....,._j._.
. . , . i
.._._.i_ . . . . . . . _ . . . _ _ _ . __ _ _ . . _. _ .
.i l, ;* i j .
1
..a.; _ _ . . ..
f _. . 3,g .. ._ ' _ _ . . _ . .- . _ . .
.. . . ...C.. ORE _ .. r - --- !
t I. i
._ __ = ._ .
i l l _._ . . . _ - - . . - .
. i , i .._ l _ .
i.
.= . __ . ..._ _ .
i .
. i
- s. 2.0 ~ ~- - - - -
l--t--- -- W u ._ __ _ _ __ t_ . . .
- m. .
-s, I. - _
Eg RELIEF LINE".. ...... .i o . . _ ._
.C % U * .
ae __ . _L. 2_ . . . . _. e- , I x . _ _ . . _ . .. v, ._ . _
= .
sc . i ! .
. . _ . _ _ . . _ _.. . _ . .l <
i
. I 1 i e i
1.0 - -- - -- -
, T - --- -
. . .. .. _. _. . . . . . .._ .___.I...!
s
. . - . . . . . _ . ;L.... ., .
l i ,
. i
. . i i i 1 .
. , I
. . 4 .. . . . 1 .. . r _. t .
. - . _ . t ' =
10b0 20b0 3000 40C 0 Time (s)
' TRAP MELT PREDICTIONS OF MASS MEDIAN DIAMETER OF SUSPENDED. PARTICLES IN THE CORE. AND AS' A FUNCTION OF TIItE FOR THE TC SE0UENCE FOR THE GRAND GUI F PI ANT
4*" -
" " ~~' .' '
, .- #lgt
*+--,--*N.-L. -
l , i
- 4 .'
.'; j . i
- RETENTION t1ECHANISMS FOR SPECIES-IN GRAND GULF SEQUENCES (Fractionalidistribution of species.'at vessel failure) .
{.
-__--.n.--.----,.-~___ -
'TC TQUV-4 Cond ~ Chem- Aero Susp Cond Chem Aero- Susp Cs!' .18 --
.19- .01 0. --
.01 . 27.
Cs0H '- .17 . 12 .21 .01 0. .40 0. . 14 Te .O. .99 O. .01 0. .67 'O. .33
- . Aerosol -- --
.73 .02 -- --' .64- .03
---_a___,~.a_.----- ---.- .........- _ _ __
e e 4
.' l b
i e 4 6 e
C RADIONUCLIDE RELEASE UNDER SPECIFIC LWR ACCIDENT CONDITIONS -- VOLUME 11, BWR, MARK I DESIGN l l PEER REV'EW MEETING g U.S. Nuclear Regulatory Commission l l Washington , D.C. MAY 24 & 25 ,1983 l Presentation Notes fT-OBattelle Columbus Laboratories
'%.i
- . - . - - . . ,. . . . - . .~.,-. . .
. _ . - - , . - = . . - - . ~ - , - - , , , , -- - -.-r ----r- ----w,- -
--c--e--- w-v-w - ew e v e -w--+
ew g cm 3 he ACKNOWLEDGEMENTS 8 GENERAL ELECTRIC 8 EPRI e ORNL 8 SANDIA
, m i
OBaffelle cei e. ,<.e.,...,4., _G.(. N
a
.b
(' SELECTION OF TYPES 0F PLANT 5
,e SELECTION OF
$PECIFIC PLANTS
,e SELECTION OF' ACCIDENT SECUDCES
,e
$PECIFICATICN OF PLANT INVENTCRY GEOMETRY AND ACCIOU T --------
SE%'ENCE PMDOMENA QRIGEN
,r OVERALL THERML HYDRALLICS MARCH v ,r ,r s
PRIMARY SYSTEM RELEASE FROM FL'EL THERML HYDRAULICS ------------ COR$0R _ _ _ . MERGE
,e
*P .
CORE-CONCRETE PRIMARY SYSTEM TRANSPORT M ERACTION TRAP-MELT CORCOM CORE CONCRETE RELEASE
- P0(X, 3CRU88ING VANE 5A . . .
~~~~ ~~
NAUA.4 AND SPARC ; I REtusE To av!R0mMT.. c OBaffelle Columbus Laboratories i A.L . . . . _ - _
l l ov.. . PRESENTATION TOPICS 4 INTRODUCTION (J. A. GIESEKE) e SEQUENCE DESCRIPTIONS AND THERMAL HYDRAULICS (R.S.. DENNING) e RELEASE FROM FUEL AND TRANSPORT.IN REACTOR COOLANT SYSTEM (M.R. KUHLMAN) 0 0 IRANSPORT IN CONTAINMENT AND ATTENUATION IN SUPPRESSION POOLS (K.W. LEE) e
SUMMARY
(J.A. GIESEKE)
\
l 1 l OBallelle Columbus Laboratories t
. Ji
, , ,, , ,.. n -. .~- . . + . - - - - - - . - -~ .-. *- ~
a .-
i i Ci . O E])L DESCRIPTION OF ACCIDENT SEQUENCES-AND THERMAL-HYDRAULIC RESULTS BWR -- MARK.I DESIGN
, PLANT SELECTION PEACH BOTTOM 2 WAS WASH-If001 REACTOR.
SEQUENCE SELECTION RISK DOMINANT SEQUENCES -- IC, IW LOCA SEQUENCE -- AE
.Q D (')
DESCRIPTION OF MARCH 2-n PARTICIPANTS. BCL, SNL, 0RNL, BNL, TVA L
.! CHANGES fROM MARCH 1.1-MODELS, CODELSTRUCTURE, LANGUAGE, CORRECTIONS l
1 -. 4 IMPROVED MODELS-ANS DECAY HEAT-4 .WATER AND STEAM PROPERTIES HEAT TRANSFER CORRELATIONS IN CORE' ,
- DEBRIS'C00LANT INTERACTIONS . _ , _ _ _ , . _
~
_ -ZIRCALOY-STEAM fiE Ci!ON
. STEEL-STEAM REACTION j HYDROGEN BURNING BWR SHROUD MODEL WAS NOT USED IN THESE ANALYSES.
C'
.Q
- k. ,/
f'i
'j 1
CONTAINMENT AND REACTOR BUILDING FAILURE MODES ~ WASH-11100 i FAILURE PRESSURE -- 175 1 25 PSIA I LOCATION -- WETWELL , AMES STUDY FAILURE PRESSURE -- 132 PSIA ~ ! LOCATION -- DRYWELL - DESCRIPTION OF REACTOR BUILDING
- - . . - - - . . . . _. . _ ~ . - . . -.. .
' (5
- i
~
r v (hK' Mh) - v
]
I M y l
- 1 F
- 1 i.
5 7 REACTOR - l l (
,O 3
,k 7 i PREDICTED LOCATION
,OF DRYWELL FAILURE i ,
SulLDING [ J DRY WELL 9 7 , r - I I LL it _ c
/ %
,,, WETWELL 1
g PREDICTED LOCATION 0F WETWELL FAILURE
-k Sun ION a o _ u u FIGURE 4.1. BWR MARK I CONTAINMENT DESIGN
/
, , - . ~ . , . , , . , - , . - - - .
,-r., ..,-,.-,<n. -- -,
e -----,,-w-,n--. , _ - - . . , - - , - - - . - _,,n.,v.,--,-- . - , , - ,.n-.,
i CE -O .i) SEQUENCE AE -- LARGE LOCA WITH FAILURE OF ECC l TIME (MIN) . BEGIN CORE MELT 12 CONTAINMENT FAILURE 34 REACTOR VESSEL FAILURE 126 CONTAINMENT FAILURE MODE -- 7' e 2 j
- I l
4-l l / !
/
.- 7 ,
\
Steam Line Steam Dryers / l
- c .
, Steam ,
- Separators ,
)
, N. s I -
- Core -
/D #
Recirculation C ' I [ *---Jet Pump _ Pump - l l ("j FIGURE 6.2 FLOWPATH FOR FISSION PRODUCT TRANSPORT IN THE RCS - l SEQUENCE AE ! I m - -,---,---,-m-.--- ,.,
.--.m. m . . , - _ -
PIPES / SEPARATORS 4 LOWER OUTER ANN US SHROUD HEAD 3 v I N TOP GUIDE \ 2 CONTAltiME?iT CORE
\ms/
1 FIGURE 6.3 SCHEMATIC OF CONTROL VOLUMES FOR THE ,~
, PEACH BOTTOM AE SE0VENCE
'g. m ,,e-- , - + - - - e- y - 4
f l - l-Cora Exit. Plate Head 3500 - - - - Separators i Lower Annulus -
.3000 s
2500 w f \~,, 2000 - ' 5 f' (^2) ' o
, ol 1500 -
I 1000 -
/' /,, //i/ W s s
y # \
- ), ill
- ~ C - - JJ 500 _ ______________s l- 1 I I I I I I O 200 400 600 800 1000 1200 1400 1600 1800 Time, Seconds FIGURE 6.4 GAS TEMPERATURES IN RCS VOLUMES - SEQUENCE AE Li 1
l l l
l l
,-~
~ ~ Plate
--- Hea d 3500 -
---- Separators Lower Annulus 3000 -
t 2500 - \
/ \
/
L 2000 - l \ ' C N e 1500 -
/ / 'A-E
. e I j 1000 - g- ----
] ,,
s00 7-K = C _ AA f e 0 .200 400 600 800 1000 1200 1400 1600 1800 1
- Time. Seconds I
1 ' FIGURE 6.5 STRUCTURE TEMPERATURES IN RCS VOLUMES - SEQUENCE AE r, N,l p - - - + ,-
l \ l
' amma; l '
lmuzzl l l Blowout Blowout suma Panels Panels ~ l Refueling Bay 1 l l i t=l ) l l
- j
.. m= = , .
- d ' .' c L' ,
[x -
.g. . l i o 7 -- : 5
=~__u-. _u _,-
g 3 " l ' l P '
... l d4
^ ' ^
~
'Is a l lp % k'0 J ,jW,~e' N Reactor l f f-- '. l l2.*9
, A.
. 7.
- 0
. Vessel l C. }c.
'l. '*60*c*W *
- CW * ' ,
3 l
. g.s jN ,,$ ]l c .
)
h.
, - - [.
- s. .. . ,
s , .
- z. . . . . o .
, c
- i
;. u= * ~%. :..., -
l , ( .* ~ro.,.: w_ b ;o g g , y c ..
) .
- b. j: .do .
- p
. ee$
)
y g y f 1,9.
- (f
- g. '%.
4: n.;
?on a
, .0,
$$hb . Q%%W?'c it o === o
>? < .
<c M
. .. . a;g., .:
. .. .. yc..;*c
,'. .v. . . .- ,. *o.9 . .
9a .
. .a . ,. :
3c.<
&. m _
.w m m
- c. ...
n = ..r -- --
'oc at ,
?
'.~- 1? s' . ;
o*aoq
-.- p.' \
l b)&.%. 0.T Lo &p Q V; i 1,. i.th m.$ 2 % } l ~
.. . ;l .9
.. , .. t "41 om.6 MM%fe'. h, ,
u FIGURE 6.6 FLOWPATH FOR FISSION PRODUCT TRANSPORT BEFORE CONTAINMENT FAILURE - SEQUENCE AE 1 l l
r, em L^ ): PEACH BOTTOM AE 2500.0 !. DRYWELL
........... WETWM1 i
4 2000.0-
- t. .
i
! 5 w
. M 1500.0-h 4 i i A . y 1000.0- , X i iN i I. 500.0- ) 1 : i wii ,.....***"*"
, *g ........................... ..................... ........................................... .....................
4 0.0 , , , , , , ,
- 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 TIME - (MINUTE) s FIGURE 6.7 TEMPERATURES IN CONTAINt1ENT VOLUMES - SEQUENCE AE i
]
w- . + 4,.. m - .Ae e .4 A_m u. .h,_.. d. # . - yew.**'f
+'4e -
~ . .
l PEACH. BOTTOM AE f 140.0 DRYWELL WETWELL t i i 120.0 -
, i -
1 . 100.0 - i A
~
4 N i k 80.0-N m 60.0- ' x A
- 40.0-I L
20.0-I I I 5 5 5 5 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0
! TIME - (MINUTE)
FIGURE 6.8 PRESSURES IN CONTAINHENT VOLUMES - SEQUENCE AE
1 O L y l mens l l smuzl Blowout Blowout 8"" Panels Panel
; Rdueling Bay
' 1 i
m . w = ,- , i k D* h
! $ :. N P.2 l I
L- h !
,@.' I..
,4 ' l i
\s.r_
v)e ,[. m . _ = . .a e, l j" ; . ll l1 c! i
', - - -~ y mgw.'i lo 'a, a ! F.C .
l g
- R.ac1 , l g o .
.: i
;, : ; y ,,,,, :; .:mm ww.
g,-- 3 < s y .
,: ; L i fPk ..(ns b i b-;
9 ;i .z t a j i. p \ : .i,6.;w,ey..9 '
?
- .)%- mr.;.u; h n:. h
}Q[o $ E, e ho g, " c fg;6; '
- ::: 4'u . .*g
, . p 4 .e h ..:$.g..:: ~'bMQ:e .o '.o:'nfq,;f * , ;n..,
**-< - T -
. . c.
4 n** - - - - - - 'Ctc 5:. . _ _-; y
. _=-
?
e g: 6.. . . 6 : 4
. >3.q i
" z agI,0k.,e' V
f ;.gf?kf f Qjkj.Wo*fo*l0,*.h
' b' s h u
i i I (' FIGURE 6.9 FLOWPATHS FOR FISSI0f4 PRODUCT TRANSPORT AFTER CONTAINMENT FAILURE - SEQUENCE AE w --- , , - , - . - . , , , - e ,- -- - - , .
o
~
! eo o 4 1 i SEQUENCE IC -- TRANSIENT'WITH FAILURE'To SCRAM 1 ! TIME (MIN) CONTAINMENT FAILURE 58 BEGIN CORE MELT 9 11 - , PRESSURE VESSEL FAILURE 217 9 CONTAINMENT FAILURE MODES - 7', 7 i _a -
O ( , l I l l Steam Line
- esm Dryers .n
/
- <.a
- r
~
l Steam l
- Seperators ,
) I s' l
! l Core
- - l
- , 1
- - )
# - l I
l l l l i FIGURE 6.11 FLOWPATHS FOR FISSION PRODUCT TRANSPORT IN RCS - SEQUENCES TC AND TW i
, =
g
(~% l
'v UPPER STEAft DRYERS OUTER ANNULUS i l
5 6 i i LOWER I STEAf1 LINE PIPES / SEPARATORS OUTER ANNULUS l 8 l 4 7 RELIEF LINES 9..J
, SHROUD HEAD 3
TOP GUIDE 2
/
CORE 1 5 <' FIGURE 6.12 SCHEMATIC OF C0fiTROL VOLUMES FOR THE FEACH 00TT0ft TC Afl0 TW ACCICEf4T SECUENCES
.O N..,
k i Core Exit
. Plate 4500 --- Head
- ---- Separators . .
---- Dryers _.
- - - Upper Annulus - - .
4000 -
- - - - - - Lower Annulus
- - Steam Line .
4 - Relief Line 3500 - ; 3000 - g -
- 2500 -
, I 2
= .
2000 ! %' /-... f 7j
/. -
r %...- -.- - ! 1500 -
/ ,
,, \ .
/ /
,/
./. , ;$l' j. -..,~~---------
1000 ::.- - f/ [
,.,., n c ~ =- ,
500 - 1 I I I I I I I I I f 0 400 800 1200 1600 2000 2400 2800 3200 3600 N Time, Seconds , i ! FIGURE 6.13 GAS TEMPERATURES IN RCS VOLUMES - SEQUENCE TC f 4 s-
- s. 4 -m -
. - - - ~ - , -.-.r----,-_,---..-em.-.c+. -. ---y. .- ,.y
1 l
..m
-- Plate 4500 - - - Head
---- Separators - .
----- Dryers -
4000 -
-.- Upper Annulus
------ Lower Annul us -
-..- Steam Line 3500 - -.-- Relief Line 3000 -
- /\
,.O
. 2500 -
t
= ^A
(_s -\ u B
/
q.> m [ , s 2000 -
/ % . ,* % ---
i N
/ , ~ % e,.= =: .- - .s:
1500 - / ', / " """"* " - "* h *
-*m
/ :
/ l
.d.
/ -
1 1000 #s ./
~t * /*
~ ~ ~~ / ~____-_--------
500 L
=-~~~~~~ -
l; ' , I l 1 1 I l l l 0 400 800 1200 1600 2000 2400 2800 3200 3600 Time, Seconds
,_ FIGURE 6.14 STRUCTURE TEMPERATURES IN RCS VOLUMES - SEQUENCE TC f
Qd,
. t
( ' e: l t t I a l C x' 1
) i a
l , lmmmm ! . l l Blowout Bicwout Panels Refueling Bay . Panels
\ \
7> ; .> i' '
'mzza l -
l Y ., ,. - = = . .
,i #
a 6+ '1 1
!y r
- p r .
I d e - - A h !
- L. ,.
I,. - f
' \V)g _ . . .m . _ u - . - _z, s ,
;: :, y r.a
[K'N:o
. Reactor : { ? " ~* l.
- 4 s
4 d 2,- . ve.,ei :: 'i n v m e , a g) p-
,s
*a
~
8 i i h-- De u [-- - @ys;@# .. ): *
)< .' . . . i.d.ws;gm,gs,!
- st.= Line o3p.R J
y 3 e
< M.eky/ , B - i e w. y < m ,
; t:: 4ma , g i +-
c,"- . . =e o b. .
?g 1
w a . .r , + safety 21 ) T,4- a, 0k.'. s Reljf
- = M.0 -
m. f f'S# p t a ===a a. aimac
&f.;. g :: .o.q*b:
e m-1 *.?Q: os ~ .c $f ~ * < 'i,97,' n.
*N m -
N = 3n- __ =- s'Oc, a.=1 . e.. q~~A_ ~=U;g
~
;{,
fff ..,e
- k oo 1.x.
g"'*.^ ' ~ '# ' l*ii.??- ^l 05%5*doYN' 2 4W J \
$ >l$'Gk!f$$f,I '
h I FIGURE 6.15 FLOWPATH FOR FISSIC'i PRODUCT TRANSPORT BEFORE c k# VESSEL FAILURE - SEQUENCES TC AND TW 't 1
- - - - , - - - - , - - - . , - ---- m_
(~~ h D y- qu . (';, PEACH BOTTOM TC 1800.0 DRYWML
.......... WETWM1 1400.0- .
1 1200.0-5 N 1000.0-D s
-4 800.0-G A
Na-
' 600.0-l H
400.0-i i r
,- ** l ,........
4. l 200.0- / - . f 4 i t a0 , , , , , , , , O.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 l TIME - (MINUTE) i
- FIGURE 6.16 GAS TEMPERATURES IN CONTAINMENT VOLUMES - SEQUENCE TC i
f ' l
= .-- - - _u - . - - - _ _
0 O l 1, PEACH BOTTOM TC .; 140.0 !! DRYWELL 3 .......... WETWELL
- i 120.0 -
. 100.0 - ,! / ~4 CO
~ 80.0-ca N b l 4 M 80.0-g A 40.0-
-~
0.0 , ; , , , , , , 0.0 100.0 200.0 300.0 400.0 500 0 600.0 700.0 800.0 900.0 TIME - (MINUTE) FIGURE 6.17 PRESSURES IN CONTAINMENT VOLUMES - SEQUENCE TC L____ __ ___ _ _ . _
O u-l l l l Blowout Blowout ' Panels Refueling Bay Panels ' l l l fo l F 4 i
., , . , . . . . - - - c: v ; - - -
t,
,i .M <
y y- ;
, r ?.c ,
5 0 $w o. h a k m.< i,,.. . m._
..m. .=_z, r.
i c.;.
. k
- e % :c
, "' ~~~@,y.-19ja,l, d a
Reactor s : f-- j d i 1 , g v. m vesses :: n, ':ww + uw n.
) g jp.- .,k .'
]
. . s ,
s . , , .
, y. n '
< - , ,1 ..
j .e.-oi,.g. gg ,gg. ;
~
(~~ 'To'.* Mg'.f, .. l , , l 7e m i k7 4:- v.- I n g,- [/
=u,c o
P;S;. E! I f f
* .. :s Q:
'D-
, $f,A@c'N .h. SWE?F,Th.
.. 0/ n '$ v pf - --
sj 4 )4 f;.. ): Im'*?a?:tbdf.,'~-
% *. Q Y,,I
' c .o.. v . .o. ot 3 ..,
, y y j _5 -i f,
*d
..#6 h{ -
.:= 2=- __
Sf
$?
c
. / -
3 2 f. c< p 9.' o.. 3.,14 .
,.,i.?E7M*1@Y5N 0;isfico*j!**o- .
p,p ,;. o h,22 ** >
= c '
- 1 g',e,*o, u
f'. , QJ FIGURE 6.18 FLOWPATH FOR FISSION PRODUCT TRANSPORT AFTER l VESSEL FAILURE - SEQUENCES TC AND TW _9 -
) (T 0 :?~)
i l SEQUENCE IW -- TRANSIENT WITH LOSS OF DECAY HEAT REMOVAL
-1 l TIME'(MIN) i
- , CONTAINMENT FAILURE 1756 -
i BEGIN CORE MELT 27118 REACTOR VESSEL FAILURE 3055 CONTAINMENT FAILURE MODE -- 7' 9
? L m__ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . .
=- - -- - + . - -
i O O 'D i PEACH BOTTOM TW 600.0 DRYWELL j WETWELL 4 500.0-
, I e
Iz.e id M 400.0-b s< ' f __, ....... y-a 00.0 - s f' ~ n.....- rc ............ ...
.... <w r 200.0- ,
4 200.0 , , , , , , 0.0 500.0 1000.0 1500.0 2000.0 2500.0 3000.0 3500.0 TIME - (MINUTE) FIGURE 6.23 GAS TEMPERATURES IN CONTAINMENT V0t.UNES - SEQUENCE TW
~ .
4500
- - - - - Plate-
- -.-.- Head 4000 -
"-"-~ Separators Dryers
-~ Upper Annulus _, , .
3500 _
* " - - Lower Annulus Steam Line
- -- - Relief Line 3000 -
0 2500 .- i n
/ I b
/ 1 i
e r
# 3
?.
b ,' s f::} Q ,^ >2 2000 - 1 I
=
a* a
.\
i 4f- #,8 - t ,'~,.s -
./ i.l s.%,
1500 - A 5
- go * *
ss t
*,e l
,. .# s -
'/s,.-
i
* '\.
"*.. ,/ ,4'f g 1000 ,.' .
( ,,,. s t .
** ,, / *
.,,.,e**=""*****"'**,,,,....-..-..
# .-- ,..#...
- g S.'
~ jt 50 -
- - . - ~ ~ ~ . ~ ~ ~ ~ ~ ~ ~ ~_ -g ~ ~,;,JS
~ *[p." ~. .~I*g. "
g% sua raen.w" # * -............. ......... T ."'".. m .=:=. T ."""".-- . { l I O I 1 i , , 7000 '8000 9000 10,000 11,000 12,000 13,000 Time. Seconds FIGURE 6.24. STRUCTURE TEMPERATURES IN P.CS VOLUMES - SEQUENCE TW i
- i 0
0 0 5 3
%)
(- 0 . 0 0 0 3 . L4 LI g EE H WWT T _ E MR DW 0 C N 0 E
,0 0 W
T 5 2 0 E S S E
) M M 0.
0 T E U L O O ,00 U V T 2NI T N E T M H N O ( I A T B 0 - % 0
- s. 0 4
,0 E u
H 5 1 M l' E C I T R U S A S E R E P 0 P ,0 0 5 2 0 1 6 E _ R U G I F 0 0
,0 5
0 r 0 0 0 0 0 0 0 0 0 _ 0 0 0 0 0 0 0 0 4 2 1 0 1 8 8 4 2 1 g
~- -
(>. t
< gOCA Nb gP4 g
' ! t >j!I , -' ,
l ,;;. j !i ; !; iII : l i! ;l l 4
m _ u-% 4
- J O
~s s
's I
t_ O l .
=
1 M T
, w p h O M I r b g
>= I O M z O l -
8 p w E- 2;
~
- s E- 2 E v
O . v , c O l O- i lm E x r 1*
-n-a e
\ W b l w
\. >=
. M A \ s.
a 8 e, e. e r a
. w I
. w -
j
- i 1:1 .-.
9,
... J e I R R R R 8
# l 12Hn1VH3dR11 1 i l
l
. . . e-- .- - , + - . .
s'-p u---r--y wN'er- y -fws e w- g -wg- vr y @wW'*'F+4F '
-*'MT""hWT 'M*-N *'M'9 v=W'='-**"**eP - * " =*N-= T*"'"*4"-"'-- *'TN"'*-'*'4- - * - - '"
s e f A - w G-- m ~s=-- KEY MODELING UNCERTAINTIES-1-
-IIME OF CONTAINMENT FAILURE.
- l-LOCATION OF CONTAINMENT FAILURE.
1 RCS THERMAL-HYDRAULICS. - RESPONSE.0F REACTOR BUILDING. 1
(, RCS TRANSPORT AND DEPOSITION FOR THE PEACH BOTTOM SEQUENCES: AE, TC, TW e RCS-THERMAL-HYDRAULIC CHARACTERISTICS e RELEASE FROM CORE { e RETENTION IN/ RELEASE FROM RCS e EMITTED PARTICLE SIZE l r l OBattelle ( c. .. e.,<.e ,...,,,, y
,3
- t w:
PEACH BOTTOM RCS THERMAL-HYDRAULIC CHARACTERISTICS e SIMILAR THERMAL RESPONSE TO THAT EXHIBITED IN SURRY ANALYSES e GAS FLOW RATES OF SIMILAR MAGNITUDES, BUT SIGNIFICANTLY DIFFERENT TIMING C e DIFFERENT CORE MELTING MODEL LIADS TO DIFFERENT CORE THERMAL HISTORY
~
/~- OBaHelle Columbus Laboratories
. - . - = . - . - - . - . . . - .. ..
p o o O T 1 ( ) . RCS EVENT
SUMMARY
FOR PEACH BOTTOM SEQUENCES s
~
CORE MELT VESSEL' VESSEL i UNCOVERY START DRY OuT- FAILURE l AE 1.5 12 .40 118 FLUSH OF RCS AT 37 TC 73 95 156 217 FLUSH OF RCS AT 122 TW 2622 2751 2834 3060 FLUSH OF RCS AT 2826 , t i
- ~ ,
g 1 - 4 i ^ QBallelle
,( c .u.e,..., ,
j -
e 3 i 10 ... , m s ,. . u .+ r i n - . e.u ,_.wm m mi e .* r ,. > e.- w.m 4 se y- .- ?:r i J M 6 r&I m 5.E rd E .=-w ++;& -C= H 3-r 9 ML-+M'
! e 6 r: r
'-: :. . . .: .i. :4._.'. L. . t -E.r. ; i. :. L: [._ .=_.:. '
..i. -* :. ; -- 7 ' _:. .=_:. i ,_".._.hr. i_
. :: _h:=._-*2.
_ . .~ 4.--. ..- -i - -_. ih. .: r.: .l.-r. f ;.: ::.=.
.. _. - ..: :_ .: a .: : .e: -
= : ; __ e .
.= . . ;.: : . .
.;a . .a __ :
- ..t...__ _.
- t-
- - --._;;;_;. -f :.. .; .:
- _ . . : t---- r : T -*. . . . . _ _ . . - _ . . . -.: ~.:------ _
.L_ L 2 : =:-. , .:
.-- .t_.
r ;;. .:: -
-- r . ..._ ' _ -_.-. -.. _. e -- - ; i.a.:.;--'AE -
.- :- :_r.--
.. - - . j...- g . ..
a'. 5 89.-.O t.-i."1 %: rI . .r T.! L -i.H a 4 s 4Cs
.gl
- 4. L-+-y r ' & t. 178 "4. r- A w - a. :-
-!::=reE F== 3 '
d ~:~ : -N-5 *F 129 9 E.T*l "l@~E:l "L- .:.'--d. :5di-'dd . -
.F- _~_-$S~+"== V.T h V.2
.a :.:- :_g . :. .--
_...:_.-.__._r.;.__-=.:__- a:=
=- ==n_..-. ._= , :-- 4_.; - . : .=_. _=.:c , . : = ~_'i -r*------- ; ....._ :..
= .ea .y=._ .. . y...
- - = - - - - _ _
g;_,; , a .,, _
= __ ____. _I.______.. . _ . . u_ ._ .r.r--_.___
_- _ , , 2.,._-., t 1
, ^
- c. _ g Y4- ---
L* 1 i a ,..- i __I I i , . . . ., . . . . . . . ,. .- 2 1
-[ f~- ::"E #T -
a , [: i .*N [ -Q =- p3 d _.~~ _3 _D_ _ M,; '.' *
- - * . _ . _. ~ ~ ~ ' "D .Y ' .
- [. ' Z
_____= - - - i
,_-. =: .
- _ =.=
"__~~ j
,?_
y?.. m - r ? * * " . * " %* *
----__;=
.T : "_* -*.N
*'s " _=.=.:__11._"*"'":".Z"" "C..-
- . _ _ - _ . =
- . . . a--.
**'--**%_-*.*-.**.="Im"-*
w e _. -
. _ .__.____.:=.__a --___"-
A,. ._
- _____.__;=-. --
- 7 .
..___. _: C:= ,_;-,_;=. __ .
_.__=_=:_--__4 __ _ _ .
..m: .* t. : _ . . _ _ _ _ . _ _ . _ . _ . , _ . _ . . _
.__-m _.
n m .. 1 _,4-._..__== _._g a l - u..m,;.'. =_u. __ :- ._2Q.*'
- _ . . _- ::_==
- :..-- *' '; a rA-
- - 7 . . : =.r
.--_=- ---
c --__s. a. _ : _ _ a.
;--5_:t ._-;- ;_.I= ' =.+": _.; f m 23 :-=_t -y .:.=_ _-J. _. :=3 . ---39= --- .= :[. .-E ._ :.' II?
a,,_. ___ . O,- J ":""" *--*_! __1 ' ' * - _" " " - * .C, * ~ * ' " - " - - - ~ '* = " . " " " . * = .-
.=*."?.".L'__
_ y _.i _ _.7 g y .__ --
+* vm 4
' K g ....
- : ~a ; .
- b. 3 J
[. I 4 Im _ J f 1
.I i t i . i i 44
- I
} . ' ; 9 e t .
1 + . - . i 10 - (.. . , _ _ 3OJ _ L. _ . - . _ . .,g w f-- m.:-'.- _- g._- _,.,~ ._;;.. :. 3-- ._.;_ _ ._g__- .=L,_. ,,
.--z.:._.-,_
A _ g;; ; , _ _ . . . _ _ __. . . _-
\ - . = ~_ . _.: -_--.l--- . _ _:.- .---: = ._._.--;== '_._.
. L__._ _ ._: . _ _ _ . - = - - . . _ . _
-d 7 ;=;_:- ' T:. :._ ' . 4;* ." ' ' ^ *..* " ' - 2 .I. _ r .;: "- T-~~~* - '
.,C.
.7 . . _ . ._._T.."" . M"- 4.~==.2--.":.L_.J. ""
_ _ . , _ . _ . ' - - * . .~ - - - ' .. _.
. .\: r=.:. =;. . .: : .2_.:--- :; - ..:c;_- - -- '- - --
r.:_ : . __ -_. ,
..__.!)"l""*.*""'""*-*
.. . - . . . ._ 1
_ . p g _
._.[.
1--
. 6 2-b b '
.g.=""d
. h' . .
$e _ _
- *=
=,==.:a.1..._.._ ._==_t
.. .*'.m. .,2_.yn. .. -
_. y ..g - - - -_ _ ,.,_-.df.,, _-- g _ ;;;. .._ .
- = _..-,,,.._.,
,_ :._._=_.'h-
. :r.,_-_-.,
.*~
- -{ :-" ."""
-~'- ' !" . . . : :._.:.i.!
*.L . a:.I * * " - " ". : .;2: ..*. . ~L *C ._._~._- I,_.'-- T----""
.-'.f--*'%.**""_ "1."*.11"* ) - ' .* * *. : 2
- _ *. . . *' = * =. ." . ~ * * " " " , l _"-".:."*.[J_.:_.".T.'"..*_""-'--'
~ ~ - ~ ~ - - ~ ~ ~
' 2. 21=
.;. p.: . = = , _ _ p "".'" =.2='*---
;d"---;=;----
- h' 1.'"O 1 - .-- ._ . _ . . . ::: .:- . . __ 1 _.6
.r..._
. _ . . _ go _ .. , . . _ _...-
.4_. .> __
b _. A g -. _ .. l i s . . i i 4 . , , , .
-* i ae *.'8 .
8 ' .
. .4 ,i I eii 6 .64' .i t e .,l' ! , , s* . - 3 i
. e
!. { . lI e 6 i .
] Q _, i,. s 1 . . , . '
i r. <lep , s.
-10 0 1500 3000 4500 6000 7500 9000 RCS FLOW RATES FOR PEACH BOTTOM SEQUENCES AE AND TC
- A + s > g n .-- u .- -,- - -- - - , - ---- --
~
' RCS RESIDENCE TIMES FOR PEACH BOTTOM AE SEQUENCE l
6000 . .
% e % % '
I e i-I i e l e f
- t f
- 500 w
b s' s _
- 8. j 400 M
i W M . M - 1 y . 1 I E , e e I e i 200 ._. J 4 mmune IL' O C ! !
-0 500 1000 1500 2000 2500 l t (s)
..--w--,,n._ . ..- . - . - ~,. . . - -. ~
v.- ,r* ey- g --
--v@ ~*==swv
. e-e--$- -- -- + y-w e- ,--* ee-- e e w-- *--r------ v-~w-ww-,v --*----w-- -
a a 2 6- 1 l j' 2000
~ ~ -
r ! ,r. , I O 1500 - - e O N w 3 1000 A
.)
G3 E . 4/1
- O M
2 e f i 500 i l l c.
.. =.g. - - . .
0 1 2 3 "T (103 s) RCS RESIDENCE TIMES FOR PEACH BOTT0t1 TC SE00ENCE ,
, - w-w g -- , - - - - .--- -me , w ,-- --,w -,pp >- --'um-g W . _ , . - - - , g . - -
/'
.m. j MASSES (xo) 0F SPECIES RELEASED AT VESSEL DRY OUT AND VESSEL FAILURE FOR PEACH BOTTOM SEQUENCES AE TC TW
.Cs 165/207 156/164 196/207 I 13.1/16.6 12.4/13.1- 15.7/16.6 TE 3.5/7.7 5.8/7.0 14.4/17.0 C
! F.P. 15.8/51.7 23.2/33.1 49.6/76.1 AERO.
NON F.P. 337/839 488/609 998/1238 AERO. S OBattelle c.,_e.,m,...,..,y
Onmmn mmmr. .. . . . . . . . . . . . . . manmmnmnm.. < ~ N.
'l : '
. .anfission Products ission Products
; % o f To.
p t.a. . l....Ae. .r. .o. s o. l. % of Total
' Aeroso'l v 10
. 80 l l
70 J . 60 1 l 50 t t 40 . . 30 t 20 i 10 i I I i il l l 2 a r Ru Zr (Clad) PEACll BOTTOM TC AEROSOL COMPOSITION AT (T = 260)
1 i n annu mmmmmrmimmmm~ nj -- r= ,ninnn aie nun.in nin no
, Fission Products.
n i i i 7 i ' I Nonfission Products l I I ly l
, % gf. Total.,A rosol . % of Total __. . .
I i t-
; j 1 80 * *
[ ' 1, 70 l l 60 -1 50
}
i i 40 i ! I 1.. l 30 - 1 20 p- - p 1 7-- t g 7 r : i 10 dl - -- I-f- p l i ! ! l l
.- .; a, i l 4-i
' I I I ! l :ll ! I l ..I l I l
l I l l 1
; ll n, iih iill lili i I
... l !
11li iin iilt ill i ! ! Sn Fe Zr U02 Mo Ba Sr Ru Zr (Clad) TC GRAND GULF COMPOSITION OF AEROSOL EMITTED AT (TIME 196)
l I 4W - ---
,m i..
Cs0H(3) 30 I - A 11 i 11 Il i1
% 11 6 \ lI s = 20 t II I1 C/ ;
- g.
I I Cs0H(2) Tu ;' I I 5 I i
~
.5 :
I l
/ --E ,
I i
; I I -
i l 1 - I i
. I I I l I i .
I l 10 I g i i I l I i CsI(3) I l I I i 1 g sI(2) I \\ l I \1
\\
/
l
~ - '
~ I' \ f l\% _ _ s 1____1__l_L__
- 0 1000 2000 3000 4000 5000 6000 t(s) t MASS RETAINED IN VOLUMES 2 AND 3 FOR PEACH BOTTOM AE l 1
..._..m..*+---.m.------ - - - -
,,,m--- .- .,y- -- ,,. , 7-. . ,w--,,. e
- MASSES INJECTED INTO ORYWELL FOR PEACH BOTTOM AE SEQUENCE ~
.___ 100 1000_
p - . _
- q. --
- M e
40 _ _ _ 400 _ . / u-----A--- A _
. /
1 /
/ % IC Q .x f .
~T 10 . _ _
100 #
)-
- E, v
g 5 1 4 - _ 40 4 . I I l I .
,- . I I
O . . . 1_. l l l 10 4, . 0 1000 2000 3000 7000 t(s)
- r e w- -w - e wa e c--r+er*--e-e- y-+4 -w- -+- - -g.e- =my- n-e ,mW--g+---ew wF=*-w-9pr- -e***-e4- --tv-ww=MTwwaW-tv-q-ee*Fw g-
('M. u - 48 40 _ _ Cs0H (2) , 32 _ 3 a.m. .
$ 24 Cs 0H (3) by . .
p
/
_____ -= .-. ---- E / 16 _
/
./ .
f I ' 8 ./ ! /'
/
~/ Cs1(2\ '
- Cs! (3) 1 - - - - - - - - - - - - - _
t _ 0 E I I I I I l 0 1000 2000 3000 4000 5000 6000 7000 , t(s) .
/ MASSES OF CsI AND Cs0H RETAINED IN VOL 2 AND 3 FOR (s,
PEACH BOTTOM TC SEQUENCE
-----.w- e- w- +-i-+- e. - --v v -er----rr- ge-*7 ---*-v'-t- - - - eum**m-c-- p-
~
- W. 1000
- MASSES INJECTED INTO SUPPRESSION POOL FOR l _
l PEACH BOTTOM TC SEQUENCE l n.. m 40 _ . _ 400
/~ .- =- -
j I - 1 I I
\ .c 10 ,
_ 100 E f y
+
5' n j - + _
- . _ g
. u F
. . c 4 - - 40 4
<...L 10 k/ 0 1500 3000 4500 6000 7500
, t(s) r ~ . - 0
,, . . . . ,- - - . - - . . . , , . .,r. . ,. -
-...,_,._.-.__..n...,,, -,.- -,- ----. ,,,,.., --
._ - . - . -- - . =. .- -- - -_ -.
----1000 -- ~ --
fx -
;t .
AEROSOL d
/
//
/ / cs0s
/ f ,, ............
/ I
/
100 _ p l ll=*
/ : -
- / / ::t
/ :
i /
/ / :l l / 55
/ 4 ll t . >.. s .-....
a f
/
s = * *...:
/ / *** / .=....
c,I {#2y - J.** Y/ ...=,,,.....=****... ..
// *l, .
. / /
.! / ./ /
! 10 /,./ /
/
(/ l
- ,.. J( /
/
- / ~
- /
I I
.g-
[" , , l
- q .,L-- 1 -
~~'
0 1 2 3 4 5 6 7 t(103 s) i TOTAL SUSPENDED MASSES, AND RCS-ESCAPED MASS FOR PEACH BOTTOM TW SEQUENCE
-s ~..
q%. o RCS RETENTION MECHANISMS FOR PEACH BOTTOM SEQUENCES (FRACTIONAL DISTRIBUTION OF SPECIES AT-VESSEL FAILURE) AE TC TW* VAP. AER. SUSP. VAP. AER. SUSP. .VAP. . AER. SUSP. CSI .04 .14 .22 .01- .10 .06 0. .16 0. CS0H .06 .13 .14 .16- .10 .06 .02 .15 0. TE .14 0. .53 .80 0. .18 .97 0, 0. AERO. -
.69 .04 -
.51 .04 -
.65 0.
l . OBallelle ( cei_ , .m,.e,m, y
. . _ _ . - - .. ~ - . - _ . .
. 1 l
1 1 u. p . W i-o w
;. . t- - -
.: .p u; ni., .2-- '
!* 3 r-W;l
_ g .3i 4 ! j-j . __ f3 .$,,.p[.L g. . _ g g_l ppi' ..i fg.y . , l. q . _. .+ : i. _[.. 3: s.. .__ . , . ._. _4.. __ r.. . _ p. .t. a. .a. , . . . _, c
.,. .!, L _ w m .
e ._ i~ ,- . r _; i [, l._,~ > . . . . t
.. . . . . . . . o ~ _ - -
w . l . l 4 lm.. _ _ _ . . i
- g. m;,,1.. .el j a. ;p: ,. -l l .;q....,4- j,
.4. E
- w. - . i . J .,
ie._rl
._..i ,i .. _.u.:
[g; e ei. y j f f.
.} ;f g g E h- ' 1= l 3 j- .f .
f. s E i .ij- j[ 'di!.
. .. o . .. 1. . . l. .L . .1 p ii _ ,. iw J.:. ,
I ]I I e
- _ [i .. .
il .c ! - , , . . .
!* r, I'P 1:!l *
.! 1% 8 I . '3# - g
' 'i -
i.,.i.c .)i.. l l. l'.- ' p.l l2. jb[ Y' .li . l4 iii, f '}}