ML20128N599
| ML20128N599 | |
| Person / Time | |
|---|---|
| Issue date: | 10/12/1983 |
| From: | NRC COMMISSION (OCM) |
| To: | |
| Shared Package | |
| ML20127B155 | List: |
| References | |
| FOIA-85-110 NUDOCS 8507130126 | |
| Download: ML20128N599 (258) | |
Text
{{#Wiki_filter:, . _ _ - _ _ _ _ _ - _ UNITED STATES OF AMERICA 9, NUCLEAR REGULATORY COMMISSION In the matter of: PEER REVIEW OF THE BCL REPORT ON RADIONUCLIDE RELEASE UNDER SPECIFIC Docket No. LWR ACCIDENT CONDITIONS - VOLUME IV, V i and VI l O.
'i s
i Location: WashirJton, D. C. Pages: 1 - 257 Date: Wednesday, October 12, 1983 8507130126 PDR 850425 l FOIA ) ALVAREZ85-110 PDR TAYLOE ASSOCIATES Court Reporters 162$ I Street, N.W. Suite I004 Washington,D.C. 20006 (202) 293-3950
. 1
r_..
.meh/r&t 11 UNITED STATES OF AMERICA
. ,f es - .
NUCLEAR REGULATORY COMMISSION 2-- BMI-2104 REPORT (DRAFT) 3 RADIONUCLIDE . RELEASE UNDER LWR SPECIFIC ACCIDENT CONDITIONS 4'
.5
.6 Gallery Room Georgetown Hotel
.7 2121 P Street, N.W.
-Washington, D.C. 20037 8.
Wednesday,' October 12, 1983 10 The Peer. Review: Meeting commenced at 8:40 a.m., 11 pursuant to notice, Mr..Melvin Silberberg, Peer Review ut Chairman, presiding. t 13 ' COMMITTEE MEMBERS PRESENT: 14 M. SILBERBERG, Chairman M. JANKOWSKI 15 ' D. ROE W. CASTLEMAN 16 W. KASTENBERG A. REYNOLDS _ 17 R. RITZMAN
.L. ZUMWALT S. LEVY D. WALKER 19 -
PARTICIPANTS: 20 - J. A. GIESEKE
, 21 P. CYBULSKIS-M. KUHLMAN 2 K.. LEE H.'CHEN.
El J. KELLY. F. RAHN: 24 R. BARI R. LIPINSKI
.2 D. POWERS
. ,., . - : n.- . ~. -
g; , -A.
~
., e -
2-s c1; PARTICIPANTS:- (Continued)
..~. ,
l ' T.,KRESS-
~2-
' . .g .,
P. BIENIARZ~
~ 3' l-
~A. BUTLAND
- I..:DUNBAR
.4- [L. FULLER' .
- E.:WARMAN
- 5i G. GREENE:
- 6-lT ---
8
~
g.
/ 10
" 11 12 13
.14 l-;.1 15 17 --
18 ! 19
'_c-f22
; 23 -
25 g d+. i'
~4 .c
~ -
- 7
, s
- , C ~
3
~ ~
~
r
".1 P: RJO C EzE D I N G S 12 .MR. SILBERBERG: The meeting will come to order,
_ t3: pt,,,,,
^
,4 -Good morning,fla' dies'and' gentlemen, welcome~to T ~ 5-
?our Fourth Peer . Review for ' the .NRC reassessment of soure.e
.6I '
? term' reviewing-Lthe Battelle-Columbus Analysis BMI-2104.
-7: I was.~also going to say.welcome'to downtown sunny 18 .' Washington, but.I cani t[do that today.
~
(Laughter) L10 -i
~MR.-SILBERBERG: A;special'welcome to our guests
' 11 from overseas.who are-with us today. We'have.Dr. Zaffiro 12- 'from Italy, Dr. Butland_from the-United Kingdom Atomic
-13 l Energy Authority, and Dr. Dunbar.from"the United. Kingdom 14 Atomic Energy ^ Authority who are attending their first meeting.
~
15 - We -have . regrets from' Dr.- Pretrangeli- from Italy 16 ' -whc was not able- to be . here . and from Dr. Torgeson from - AECL
- 17' Canada,- and from Doug Cooper. 'You may or may not know but 2 18 after the:last-LPeer: Review meeting Doug informed us that he-19 ' twas accepting a position.with IBM Research in New York and 20 - that'he would try to help as much:as-he could but it would be 21 - ' difficult: for him to attend: this ' meeting and participate.
-22~ 1So, I"know._you:will all join me in wishing him L 231 . welliin his new' venture,...new. endeavors, and we certainly
' 24l alliappreciated the contributions that-he has made to these
.Resiews'.over the'last eight or.nine1 months..
25 - -4 A ' . . _ ,. I T a
4 1- The reason why we are here and not over at 17th 2 . Street is that the regular tennants of that' room, the ACRS, 3 are'using-it today. So, we are here. I think you will find 4 it comfortable and cozy and I am sure that it in no way will 5' affect-the business at-hand that we~have to get done in the 6 next two days.
'7 Also regrets from Dick Vogel from EPRI. Frank Ron 8 will be here today from EPRI covering for Dick.
'9 Bob Bernero was to be here this morning and he
' 10 sends.his regrets. He is over attending the meeting, the
' ll ACRS meeting, in the other facility. So, he wishes'you U , success for a productive meeting. It is possible he might 13 join us tomorrow, I am not sure.
14 We would like to apologize for the late receipt of 15 material for the meeting.- This being our last meeting it 16 was actually very special and very difficult getting the numbe::
-17 of items together that Battelle had to get together. They 18 ' .have brought with them Chapter VI and Chapter VII for all 19 'three reports. We will.have copies-made and hopefully we 20 .w ill be -able to provide them by the end of the day. Is that 21 right,-Mike?
E MR. JANKONSKI: Yes.
'M MR."SILBERBERG: Because of the. late arrival of 24 material,.in' order to provide'a saving grace Battelle will N .be the sponsor of~our coffee breaks during the meeting.
- .. =.- -
5 *# '
, 1
. ? , ,
. S s'.,, . -n -
1
.,e,. < 3; be J(Laughter) o -.-1 ~
-y ,
- c. c _
J2
, _ _ -MR..SILBERBERG: 4 WeLthank Battelle for their
_- _ .* 13 . ho'spitality.
, - .4 J According to the agenda, I am supposed to- provide hm .
- 5-
'a brieflupdat'e on where we ; stand,on1the reassessment -of the
- 8)
'^
sourceEterm,cand I'think Uin'the interest-of timeLI won't 7L 7 [ pres ~ent anything. formally;-buti;let me go through-my list of 8: zwhere we are. v: 19? ' If you recall,;.we.have an Element 4 in the' reassess-(10 1 ' ment :which .was 1all: the- othar work in~ : support of what will
. c-l11- Tultimately<be called NUREG-09-56, which' includes the special-12~ Lstudyof.~co'ntainment' failure modes,'a review of the IDCOR
' ~
y - 13- ' work and ANSLwork,fandt of-course ~ associated with that: element-14 - but.really!what~we-call our Peer Review---Element No.43-was 15 :
- anLindependent review byta' scientific' organization, the
'16 American ~ Physical Society. -
~
17 " Briefly,-let me'go;through the status of those { 18
.threelitems which; if you will, yare goingj on in' parallel with
~
[ 19 all the other activities in, Elements 11':and 2. l. 3 E20f
- Now,. containment failure probability, as you know,
- we have'a< containment 1behaviorTgroup internally within NRC, 214 1
t 22 ; made Jup of representatives : from NRR: and Research. We have 323 .?supportingi:us Ja~ group sof expertsion' containment loads that '
; 24I 'are providing1 a: basis for. the : five plants on what you do as 26 5 : .besties'timate, if.you will, or distribution of the-best' ,
! i*
4 i 4
'I :-
r
.g
't I g
, . ._ _ . .m.: ,w-. w-_, - . - - - - -- . , - + - - - - - - - - - - +- -
c:- 6 1 estimate of containment loading. p
- 2. We have another activity going on that NRR is 3 pursuing'on containment capacity, that is the capacity, 4 pressure and temperatureLof' penetration, seals, hatches and 15 thingsilike that. If you recall, I mentioned that at the 6- ~last meeting.
l7 To date, on the containment loading side, we'have
's had a PWR standard problem which was completed by the working 9 group of experts.' This was for a Zion-like plant. -We will 10 follow that fup with a Surry plant and 'Sequoyah on the PWR' 11;- side._ We now have a standard' problem that is just being 12 ' circulated for the MARK.I and. MARK II BWRs_and we will be 13 circulating one for the -MARK III BWR.
14 For the MARK I and MARK II we will be focusing on 15 the overtemperature problem related-to loading, primarily 16 - temperature loading, of seals on - the drywall. 17 For the MARK III, the focus will be-on' hydrogen 18 burning. Ig ' That work is moving along, it is tight and we have
-m a' lot.to do. We will keen you. apprised of the results as we 21 move on.
22 Hopefully, as you know, when these containment 23 - failure-probability studies areLdone, we would use that to 24 perhaps go back and refine some of.the Battelle BMI-2104 25 analyses as well.asiuse it' to'put the whole question of source
- p: .. t , x . {.;
+: b _ d;; } /si , _
7
.g s ,
au g.g ; -
. a
[ L' - 21f tArm 'in' perspechive irelative': to : containment failure.. 2 5:i > l 2 -. ;In 'regarditofthe IDCOR-review, we can only-review 2 3 :- materia 1 which we have' received and:to date we have just
* (4;
, s
. received the1first. package of " material 1 that relates to -- the 5: primary thrust:willibe-towards containment loads. ~
.So',.that s ; kind of' fits in.very: nicely with our;on-going work on containme lt-
;s -
- - 7 . failure prob'abih.ity. That; work 1is underway.
's The remainder of the IDCOR. work, we will put 'it into
- P -
(32 review?as::it;is' received. -It.-could.be.that to the extent'that to ' we can; review; that work, the- depth of the- review of that work .
-11 -could well'put the: things'that:we need~to-do:in the next .
- s. 12 ~several' months.andubeyond, may we11-put that work on the
-13 critical path. So, we. willi-have.to: watch it:very closely.
14 On'the American Physical Society. review, the grant
- 15 .
was' awarded in September. -The American Physical Society had their;first organizational:and: orientation meeting the end of is - 17 September and they ' are now preparing to get into the problem - J 1s ~ in :some depth. 19 - The. schedule :for- the APS study is basically as
,a follows,-as I understand it, the principal study'will be
~ 21. drafted as-they move through the summer, next summer, in. terms L 22 ' of putting it all.together. My understanding'is that the.
fm- first public presentation of their.. study would be at a' meeting-se r of- the - Society -in November sometime, November 1984.
'M m of = course',1following that, the' APS would then publish e
1 , 4'
~
e- -- 8
~ 1- the -fullJ report in:the Reviews of Modern Physics, as they have 2 'done'in the'past. ,
3 'I.think that brings us up to date on the parallel 4 Jactivities - that are. in progress -ort the source term reassessment . 5 Without further~ ado, I will turn the meeting over to Battelle "6: Columbus, ' JimGieseke, who - just with a brief correction of 7- the . agenda -- will be' covering primarily the comments 5 rom 8 Volume IV.from~the.last meeting, from the July meeting, as 9 well as some on-going issues from some of the previous-10 meetings.. Jim? , 11 MR. GIESEKE: Thanks, Mel. I don't'know if it is 12 'necessary to have a microphone up here or not. It is? In 13 that-case, we will have to1 rearrange this. 14 Can 'everyone see the screen okay? No problems. 15 I will wait a second while-people play musical chairs. I 16 think we are set.' 17 I.would like to add my welcome to that of Mel, it la
'Is good to see all of you'again.
1st What we are talking about today,Dat least as far 20 as the Battelle presentations,- there is a handout that you 21 will be'given that has four elements in~it that are noted
- n. here, beginning with the review of comments.-
' - 23 But before.we get into the comments, I thought it 24 might be useful to go back over just a little bit of the 25 introductory: material to say 'what we are doing and why. There
a o _ m r
~
9
- 1' are a few new people here. So, I~ thought I would'just review
;~ l 2 the-objective. I' am sure -to some of you this is the fourth 3 time you.h' ave'seen.me read this' slide or.show this slide. !
14 OurJfirst: objective-_is to. develop updated release J ?5' - from plant fission productDsource terms for four types of
~
.nucl ear power-pl ant s and-for accident sequences, given a 6-27; . range of_ conditions.
~
8 That'i's important,_-as I.have pointed out in-the past . 8' ' We-are' interested in range of conditions and we are not
-10 ; - necessarily.trying to-direct our: attention to an'y particular 11 . sequences in . terms .of . risk, although we would like to look -
s12 at' some of the sequences that- have in the .past been considered,
. 13 more significant in terns of risk. We are looking for a 14- .- range of. conditions. ,
15 Am I in front of you people? It-is difficult to is . get back far enough. 17 The estimated source terms are'to.be based on is . analyses of. fission product l release from-the fuel transport 18 Land deposition by using improved computational tools in a
# consistent step-by-step manner.
21 _our second objective is'to determine the effects of fission product releases associated with major differences 22
. 23 in input parameters, particularly.those associated with plant 24 - design and accident' sequences.
25 The third objective is to' provide in-plant. time
~~
1 -, _
< w; ,--
/;
, :f: W 1
p-4, , 10
~
itj Land llocationJdependent. distributions of fission product-mass. e- ' rk - J2 ~ f :N'ow I.would'like to}go-over what we are looking for v: . . iin' terms?ofJreports and:such,-in case y'outare confused.
, 3; s>
^
" ~
> ~4 We i are looking-initially.at-really a-seven-volume ,
~
. s' : set.- The first, .all of BMI-2104. The first volume deals
. -6 .twith the PWR' Surry plant -- notL " Slurry" as itds' always on !
L Q 7i this'. slide '-- using a MARCH 11-1 analysis as' you can see and. - J. 8- - what we:will be talking about ?later today as - Volume V, which 9' .is the:Surry plant, using:the MARCH 2 code. j
~ The secorid . volumeC has -to - do ;with -MARK I design BWR,
~
10 - , 11: Volume III with theLMARK III' design BWR. Volume IV,is
~
12 .'Sequoyah, an' ice condenser. containment design'. As I mentioned ,
~
.13 : Lv11s-Ea Surry revisited, using different analytical techniques 14i tand particularly.the MARCH code. ;
'15 There arejalso'some differences-in the detailed i 16 .information Lwe - had available on the upper plenum - structure 17 - fthat we factoredlin here. -I will go over.some of those 18 : items'.in more detail..
< > 9 a . m ~. . .
lei The last one_on thisolist-is the Zion plant which : sof 2
.we'will talk'about today. Volume-VII has--to'do with ;
- 21 1 ' compilation of.all:the comments::that.you people have made and~
'our. response;to those comments.
Just.to go'over the: status,-VolumeyI,'we.-have i 23 - , 24 : . published theifinal~ draft; Volume II-and III, I'think you have j
. 25 L 'seen'.our first draf t and we 'have :had peer -review and talked l 's s t
j
e-
;y - ; 4[ [-
11 e
"L ' "
'about;those.
~ w? t; The -final'. work for publication, or distribution
,; +. <
i
, 2; anyway, fis '.in , progress.-
- 3. The .. Volume IV. Sequoyah, . we have a partial first draft
- 4- out; ThisEis:what you have received parts of and will be s-5 getting tot.y' additional parts of,'and we.are'having our peer 6' review, of course.-
(7 - "V61ume Vfhas the same status, and Volume VI has the 3: sameLstatus. .; Volume VII we are just beginning,andLworking.
's- with' compiling;the comments 1that you people.have made, so that to we-can answer them.
- 11 ; To be a'^little.. bit more..' specific, we do not have 12 all the calculations,done that we would like.to have done. So-I thought it would be onlyl fair tot ell t you what we have and
~
g -- 13 14 - what we'do not have'at the-moment'regarding these last three is plants. ' The "Xs" means. it'~is done :and the "Os" mean - that it
. is t is2not done or in progress. .
Some of these on the thermal hydraulics, I believe 18 ; these are done the first time through. This one, also. We
; 3, still have a--little bit of work on-those-getting the thermal-20 hydraulic information~in: shape to be used in the subsequent 21 calculation.. ~
22 ~ tSo,Ethe bottom line really is, the-information that: 23 you willisee in your , reports reab.ly- has the thermal hy'draulics 24- for- all' of those ' sequences land ~f'o r ' those marked by "Xs" here 26' for Surry.. and for Zion, ~ what. we have checked here, there are
12
'15 someftransports in;the primary system pieces. The brackets 2 indicate that we do not have to_do-that twice but only once.
3 So,-there are a few-cases'here that are misssing. 4- 'I might mention that some of the strange notations 5 'here,'th'e betas of m, has to do with the multiple compart-6- mentalization of the containment. 17 sequence 1 and 2 has
- 7- to do with .the compartmentalization for the primary sys~ tem 8- .alongithe' flowI path. I1think that is all that needs clafi-9 fication at the moment.
- 10 But the bottom line is what we have for the 11 containment is missing,-the denoted-results here, so you u know what is there and what will not be there by checking on 13 that' graph.
14 - Now I would like to talk a little bit about some 15 of the peer review comments,.that comments that you people
~
-16 had or the questions you raised at the last meeting -- In 17 July, I believe, it was.
18- There was a question about the purpose of the work 19 which is kind of an overall question, and I thought I would
' 20 bring _ that up again. Although we have a stated objective, I 21 guess'there have-been questions and it largely relates to how 22 many sequences'we are going to be analyzing for each of these M plants.
24 I would like to reiterate that the understanding 25 amongst those'of us who are doing the work, anyway, is~that
-i i 13
-t . the major. thrust of what we ~are 'doing is a demonstration of p
, 2 .afway of analyzing the' source term from plants. It is more Y~
3 a l demonstration of the methodology- than it is a compliation of L
.4~~ -a lotlof results for a lot of different plants.
5 We have'tried to cover several plants and'a number t
.6 ,of sequences in order to get.a . feel for the sensitivities
'and-the major variables, as were pointed out in the objectives, 7-L .s- the differences among plants and among some of the sequences s that'have afrange of condidions.
10 But unless someone wants to change what we are t 11 : doing, the purpose is not to compile a horrendously long list. i- 12 : of sequences and variations on sequences, that is more like a sensitivity or uncertainty analysis which is being done
, 13 14 elsewhere.
15 There are a couple'of comments that I think Peter i 16 Cybulskis can address. There were questions on hydrogen
- l. 17 . burning and the possibility for by-pass of the~ ice compartment is which, I think, bothJrelate to Sequoyah that we are discussing l_ 19 today, and also some questions about water layer over the 20 melt,'especially over.the, core concrete as,they interact.
121 So, I will let Peter Cybulskis talk about those 22 items here 'for.just a minute. 23 MR. CYBULSKIS: At the last meeting, we presented 1 24 a number of calculations in which we' indicated that for the 25 Sequoyah containment the hydrogen burning, even with ignitors, 4 L t..
. 14 1
may file the containment and it caused quite a few comments 2 and questions. 3 As I indicated in the last meeting, we were not 4 completely.done with the calculations and I will present some 5 of the results of the latest set later on. 6 I think perhaps the key point to be made at this 7 point, without going into detail, is that we have in the 8 Sequoyah calculations looked at least two sets of situations 9 for the TML and the TMIS prime sequences, one in which the 10 burning fails the containment and the other in which the 11 burning either does not.take place or does not fail the 12 containment. 13 At this point, we are not making any representations 14 as to the probability of the failure of the containment for 15 these sequences. As has been pointed out by Mel earlier, there 16 are a number of working groups that are addressing the 17 questions of containment loads and containment capacities 18
.which_are needed to define the probability.
19 What we are saying in looking at containment 20 failure as a result of hydrogen burning is that based on 21 the results, as we see them, it looks like the containment 3 may fail, therefore it is prudent for us to look at 'the 23 possibilities. and bed the question of probability until the 24 working groups come back with their reports. If it turns out 25 that these probabilities are sufficiently low, then some of
7 15; 1- these sequences _or"subsequences that we have looked-at may 2- die a normal de'ath. If not', they will remain in the. picture. 3 On thel1ssue'of11ce compartment by-pass, I guess
'4 - since'the ice compartment-is such a, crucial thing in the 5 -Sequoyah. containment in terms of holding down the pressure, s possibly in terms of removing fission products, I think the.
7- obvious question that has been_ raised by the reviewers'is,
-sl there.has to be a way to by-pass the ice compartment.
9 We-have not explicitly. analyzed any sequences that 10 by-pass the ice compartment.
- 11. One reason why.we have not analyzed any sequences 12 is that, again, based on what we know it does not look likely 13 : that there. will be major by-passes of the ice' compartment 14 unless you invoke such phenomena as hydrogen detonation that 15 might breach the barrier between the upper and the lower is compartment and thus by-pass.
-Quite some time ago, back when we were doing the 17 is RSSMAP studies, we had done some sensitivity studies on is what. efficiency -- if I can use that term -- must the ice a condenser have to fail the containment.
21 Atr that time, we - concluded that for things like 22 small break LOCAs, intrancients, you could actually by-pass 23 'an awful lot of the ice. As long as you had the sprays 24 operating, you' could still hold the pressure down. For large s LOCAs, obviously you have to have the ice or the containment
16
~
L1 won't survive.- Of. course, the whole concept of the ice 2 condenserfis designed-for the large LOCA and has been
- 3 ~ demonstrated experimentally.
4 There are a couple of other fission product paths 5 that might-by-pass the containment, I-think, that are worth 6 mentioning. One of them is ;the air return fans and I think -- 7 -I.do not want to' misquote Dee Walker but I believe -- Dee 8 indicated at the last meeting that there are-check valves in 9 the air return fan lines which would prevent backflow of 10 gases'and vapors from the lower compartment-to the upper 11 compartment. 12 The other, perhaps obvious, path that does not go 13 through.the ice are the drain-lines for the water from the 14 sprays going back down to the sump. Those areas are'relatively is large -- relatively small, I am sorry --- in comparison with us the f flow path through the ice condenser that.we do not really .17 feel that-there is special need to consider them.. 18 So, I think I will stop at that point on the ice 19 ' condenser by-pass comments. MF The' comments regarding the water layer over the 21 . melt, I believe, had to do with, how do we treat - largely n- . addressed how do we treat the fission products as the core 23 debris attacks the concrete. Let me just say a couple of 24 comments about that, and maybe Dana Powers may want to add Mi something to it, maybe not.
11 17 g-In'the MARCH calculations,.we do-keep track of
.2q Lwhere the _ water is, -and if water is predicted to be in the ~
3: cavity, then,the calculations of the core concrete interaction 1 4 .. do take into account.the-heat transfer to the' water. Basically, 5- . heat transfer to the water will be controlled by a combination 6 of radiation and film boiling while at the same time-the core
;7 T debris may be attacking the concrete.
s' So, the water heat sink is included in'the mass and g- anergy balances and from discussions with Dana I understand 10 - that when he does his CORCON calculations which he uses as L?. gg input to VANESA, he also assumes that' there is a low. temperatur e 12 radiation heat sink for'the VANESA calculations. So, we 13 believe,that is done fairly consistently if not exactly. I think. the next -couple of comments, I think I will
~
14 15 turn them back to Jim. 16 - MR. GIESEKE: Thanks, Pete. 17 A question came up regarding the gas access to the is ice in the baskets in the ice compartment. The thermal gg hydraulics -- Peter Kirk-told you about this -- but the 20 thermal hydraulics is based on' experiments, large-scale
.21 experiments, at Westinghouse -and the gas contact with the ice
. zt is a correlation of. experimental results. That is for the 25 thermal hydraulics' part of it.
24 The fission part, fission. product deposition 25 portion of it, is a little bit different. The main phenomenon i' i 4 .----r-- 3 y . -
*t- -
+ -"'P'
_ . . , . -. . _ . , . . . . m . . - . . ..s .. _ . __ 3 - ; x ll% ,,, ,
~ ~_
.E ,~
V , : -
< 18.
, 4 ,
' - e 4 ., .
112 tthat(is important to fission product. deposition in the ice
'.o" ' l2: . compartment islreallyithe condensation-of the: steam, which t -'
, ^3' -isThandled or comes from the' thermal hydraulics.- So, under ,
~
- 4 .those conditions whatyyou need to. knew-is the. condensation
- 1. 5.. rate ~and(it is: not: sensitive to how you . divide up- the l
18' l surface.of the ice. IIt_followsithe flow of the steam as it
- 7i : conden'ses .
ef - So,.it becomes.an1 irrelevant question in a sense 8 'if!the~ correlation of the~ experimental results are good. 10 There are. otherfmechanisms that come into play, and , 11 ?for.the_other. deposition: mechanisms such as defusion, there
- 12- is an2 assumption that~there.i's, gas a'ccess_within'the ice
~
i 5 13 - . itself, that :it has some structure and _ the gas can get in ; 14 ~ there. f 18 : However, there'is_-no sedimentation assumed in the 16 ice. So,--it-is-sort of aimix.of things but really, the most _ h - 17 ' important. aspect is the' steam condensation anyway which is, , 18 as I. mentioned, . the fresult of the experiments ' or based on . L1s . experiments. [ 20 The.,last comment that we wanted ~to address had ;
- 21 to.do with.the' data-base for'CORSOR, some questions whether as ' the release is<too high,'and this relates to the fact- j j
c 28 e tha't George Parker's prediction:of,the source is somewhat-24 ' ' lower 1 than'whattweHare: predicting ' .If-you add up all'the man s .l' 25 that-~ we .have, being evolved : from the core , that it comes up -+
. ?
i-
* .. h i ? ,
s f + t- 9". ,,y c- 9 * - y g g.-m-e ,wem e # g - . - - - - n- , -m--r .+-- 3 . y .ww a ,
( _ _ 19 1-to somewhat higher amounts of material than George Parker
-2 estimates from his experiment. ;
3 I think maybe Mike Kuhlman might want to talk about 4 _it, but we are constantly getting new input coefficients for 5 the CORSOR ? from Oakridge, from Dick Lorenz, who is going back
.6 over the material or the data that he has available to him.
7 This is being readjusted as we go'through these reports'. It 8 is one of the elements that may not be consistent through the
'8-whole series of volumes because.the coefficients have changed.
10 - As you_ recall, we changed how'we: handle the tellurium and we
' ll ' 'have now' changed how we handle the control rod materials.
12 So, we have added that in.
'13 I' think Dick is working on another model that has 14 to do with . release of other materials in thd core. We do not, 15 I do not think, know-yet for sure how those all compare with.
16 the previous calculations. It may well be that the total rate 17: will drop as we use his more recent information. 18 That finishes those comments I wanted to go over 18 -briefly. I showed this in the next slide before because 20 they deal with what we did in Volume I and Volume V, and I 21 think they are pertinent today_because there areca number of 2 things as a result of comments that were thought'to be importar.t 23 and were in some cases addressed in Volume I as.it was publishe,d, 24 but in some cases were not. We have put them into our 25 calculations of Surry in . Volume V. I would like to just run A .
, ~3 1
~ g' '
~
_~ 20 y . ,. A ji%. ' <-
-Q f' 1; through quickly.- The'se are what we expect to have; included 2' ~
}when;we: finally get through with Vo'lume V.
{, , There are some
^
3 Litems;that Jhave not_yet been included inthe' calculations "4
.just because some of the -sequences are not done, as I pointed
;5 (out-earlier. '.-So,KI thought I would go those quickly. *
~4-
~
z This first item used. MARCH ~2 for all sequences, has
'7 been'done. The second item, the revised upper plenum, or
;8-
-an. improve'd. representation of.the upper plenum'has been done.-
8 We:have: changed, as I have just mentioned, the release rates 10 from th'o fuel as we get more' information on the coefficients. 11' The inventory distribution,. originally we' assumed
- 12 that it was uniform throughout the core. However, I think
- 18 since.the'first volume'we'have always had a. distribution of 14 i 1
fission _ products ~ within- the core based :en the power: distri- ;
. 15
.bution.- So, this has been".-included really~in all the volumes f
- 14 since number one and it is again included in Volume V.
17 The effects of decay heating'of deposits, we have
' Is
, tried to-make an attempt at that~which,was not complete, and 18 we talked about that in the previous. meeting. We are still in i 20 .
the process of doing that for the Volume V. That is not' 21 -. complete b'ut that will be included. 22 - V-sequence,=that is one of the sequences that<we 88 ' are still working.on. It will.be included but it is not there , 88 - yet today. i L 28 - There were some problems in Volume I on the geometry I
21 1 we used. I think we have corrected those. 2 Water condensation onto the wall, that was included 3 in the final version of Volume I for some of the sequences 4 and it has always been included in all the volumes since 5 Volume I in the calculations and will be in Volume V, Surry. 6 There were some cpsstions about the flow rates. I 7 think we corrected those after the first Peer Review meeting 8 for Volume I and will again use the corrected values in 9 this latest volume. 10 We did some sensitivity to spray removal, drop-11 size sensitivity. That was included in our Volume I. This 12 is again a question for the B-sequence and this is one of 13 those sequences that we are still working on. 14 The compartmentalized containment is a sequence that 15 is not yet done, also, but it is in the works. Of course, we 16 did include some description of the Trap Melt Code. earlier 17 That is just to summarize the points that were made 18 early on and to let you know where we stand on those 19 calculations. 20 , MR. KASTENBERG: Question. 21 MR. GIESEKE: Yes. El MR. KASTENBERG: Will you be able to tell as, as 23 you go through the presentation, what the effect of moving 24 to the MARCH 2 is with respect to that second column? 26 MR. GIESEKE: We will try to do that. We have not
p- ,
.g s -
.f'
~
3: +
w J. _.
~
,,c ,
c- .
) .
,g 3= "
. ::a Q .
.> ' )
, , _ . y :11 (hadlaflot=of.-time to-sit down'and look at~all the numbers. 1 yJ , (. (
.2 You may havi had more time"than ' we i to 2 cok < atr t.he . numbers .
But
~
3 e a. c t3 we will try.to'do as much of that as.,we c..n'as we'go'thSupph, 1,,
+ '
i4 not on,1y~on the temperatures land the:thermaY hy'raulics e but A n \ 4 5 also,7 there'are some differences in what those:mean to the. , ( 18. fission productj eransport Jand deposition. +.
,We will try to point yl
-3 thosa out_also,('a's we go'through. ~
~
.)1
.7 '
4 ]' *
.8
.p,4 j,}(.
r,i
...i.'-
m N As-I\said, y the etraluation and. th'e' h6mpa'i-icipn , we
. ig < 1 ;( y qq se have..not \had:a lot of time to look at those r
us things / so we willebe' sort of off the cuff. i But there';;are some elements 10 tr. , 5 - '/ S ;. . g that have shown'up,?I think. f v 3' ,
.d . 12 One of the things, though, is that'it is difficult (
\ '(
- 13 to make:a sequency-by-sequence compa ison without spending 9:
14 : p
, sometimeonit[hecause,,1fyou'lookat,t i like Me bottom line s
15 release number, you can say,. well, that is theT- difference betwee n cy 16 - the two codes, the MARCH l.2, MARCH'2. But tha't'does not tell' 17 - you much because what' it may mean is that the containment it failure time is different. 19 It is thos.e sortsj of elements that you need to go
, 20 through and really pick out. That takes a while to go through ~
A. r. 21 and" sort it out', what is/ causing the difference -- which is l really wh,at we would like'toi see. 22 28 f .y T,hatlis all on the Peer Review comments . - Shall we a, ,
-i 4
. 24 ~ go on'with Sequoyah?.- '
1 ' 25 MR. REYNOLDS: I had a bestion, Reynolds g om u ~
,< o w
';- .< ; ' s f
- _: - -- - l A -- A- - I W ^ ^ - - - - - ^ - ~ - - - - - ~ ^ ~ ^ - ^ - - - ^ ^
) r e ; 23 Lil Virginia'.'
2
~
ILwantedito.ask,'why is not the V-sequence included
- 3- Lin'the Zion plant? "V* . is one of the - high risk ones in 4
.WASHc1400' forSthe Surry plant, .and it is still- one of the high Y 5' .releaselones for your analysis for.Surry.
6
.I am wondering why you did'not include the "V" in 7 Zion,-therefore.
8 MR. GIESEKE: Do._you:want to answer that, Peter? 9 MR.LREYNOLDS: I might add in fact that in just 10 casual conversations with the . IDCOR people I had heard that 11 the V-sequence was not very important. Then it seemed"to be 12 - important: for Surry and it was important in WASH 1400. 13 So, that is why I am-trying t 'to understand why we are 14 not concentrating on "V".more. 15 - - MR. CYBULSKIS : . The;V-sequence was . very important in 16 the case of Surry for two reasons: one is~the very obvious 17 one that if you by-pass the containment directly, you will 18 get very large releases. lit The other reason, of course, that you have to keep 20 in mind.is the_ probability aspect. In WASH 1400 was, I believe, 21 really the first time when the significance of the V-sequence zt was. identified. It was unknown before that, and it turned out-12 thatL for the- Surry plant because of the arrangement of the 24 check valve and, I believe, the lack of monitoring between 26 the check _ valves, et cetera, the probability of a V-sequence
24 1
-occurring ~was ,substantially higher than in other plants.
2 Once'the problem was identified, I believe the 3 situation was corrected in the case of the Surry plant and 4 in a number of other plants. For some of the analyses that 8 I'am familiar with, in particular the RSSMAP analysis, the V-sequence was identified in all cases, I believe, as a 7
'possible sequence.
8 But I do not believe that any of the cases in RSSMAP, for example, came in anywhere near as high in probability as I' they did in Surry. 11 For the Zion analysis, again I believe the V-12 sequence was identified as a possible sequence, but its 13 probability was. considerably lower than for the WASH 1400 I4 analysis. 15 The reason why you are not looking at it is basically 18 the probability argument. I may'make one further comment, 17 in terms of consequences I am not sure that it is really
- 18 necessary to look at the V-sequence in Zion and the V-sequence is in Surry, and the V-sequence in_some other plants. I would
" . expect that the consequences would be very similar. Dee?
21 MR.~ WALKER:- Nes, let me comment on the V-sequence. 22 From what you have said, the probabilitj of V-sequences
' 2s was indeed down after the Surry analysis. But when'we did 24 Zion, in the Zion-PRA we_ identified a.V-sequence which was, 25 of all probability, was still quite dominant with respect to e *
[ ..
m - 3 25
-1 consequences, and that was failure of the section site gate 2 ' valves.
.3 So, I think Al's comment is appropriate based on 4 the design PRA. However, I think you will also find design 5 differences between Sequoyah and Zion that you probably 6' adequately covered.
7 MR. CYBULSKIS: Do you happen to recall what the 8 ' order of magnitude probability for "V" in Zion is? I 9 don't.
~10 ' MR. WALKER: To the magnitude of ten to the minus l
11 seven. By the time the other sequences are done it it 12 quite important with respect to the risk. 13 MR. - CYBULSKIS : I think that.is a point well 14 taken because in the case of the Zion PRA all the probabili-15 ties were down relative, say, to Surry or something like is that. So, it may still have some significance. 17 MR. SILBERBERG: I would just add that Sandia is 18 just now concluding,a. major study updating the accident 19 - sequence evaluation for the Accident Sequence Evaluation 20 . Program. It is a document, I.think it has either just come 21 out or is coming out which I think will put the entire
- 22. questien of accident sequence probabilities and various EI' Levents like that into perspective, which I think will be IM useful.
25 MR. FULLER: Ed Fuller from the IDCOR Program.
i
-]I' 26 1
I would like to add a little bit to what was said 2 by both Dee and Pete. -In the IDCOR Program we are looking at 3 the V-sequence.. We are looking at it with the Sequoyah 4 plant with the idea, though, that it would be equally
- s. . applicable to Zion.
e We are not particularly interested in the actual 7 probabilities but again from the consequence rtondpoint.
~'
'. 3 MR.
BIENJARZ: My name is- Peter Bien'jarz and I 3 would like to add that in the NUREG-19-89, October ' 82 10 . the : analysis ~ was done for the V-sequence in Zion which 11: specified that to'be equivalent to.a small break, one and-a-u half inch. Could you comment on that?. Are you aware of this? 13
'MR. SILBERBERG: = Excuse me, the report was NUREG?
14 MR. BIENYARS: 19-89, and it is entitled, " Analysis is of Hypothetical Severe Core Damage Accidents for the Zion 16 Pressurized Water Reactor," and .for the V-sequence it is the 17 suction line sequence. The break was -- it says that the is break was considered ~ to be an inch and-a-half diameter break instead of a large break.
- 13 20 Tnat,makes a whole world of difference as .far as
- 21 the-V-sequence consideration.
.22 : MR. CYBULSKIS : -I' guess to answer the direct 23 question, I personally was not aware that the. break size for 24 that. case was.that small. The V-sequences that we have
-s analyzed in this particular study have been characterized as a
- gi L = a .
e - 4 51: break in the-six-inch diameter-line. There conId be some
- 2: dif ference.
p 3 ' MR. GIESKE: Okay, Lwith that we will move on to the 4' next topic.which happens:to be'the Sequoyah calculations, s- ~ volume 'IV . and a' summary of : those. . I really don't'have 6 much to say in the way'ofoint'roduction in-this case, I think. I apologize to th'ose people who were -not here
~
17 ( 8; ibefore, but we have not planned to go back through a lot L
- 9. .of the introductory material again on Sequoyah since we
! 10' :wentLthrough that,.when.was it, in' July? We went through 11 the"first.go-around on Sequoyah, particularly the selection l' l 12 ' of sequences .and the thermal hydraulics in a L description of (' t
~ 13 'the sequences.
l 14 - . So, weiwould.like'to move on to the other aspects 15 - . of JLti 'right away. The people that are going to be. talking-i i i 16. -about the calculations-are Peter Cybulskis who will be dis-y 17 ~ -cussing?the sequence description and the thermal hydraulics-.
' 18 Mike'Kuhlman will talk'about;the-release from the
~
19 -- ' fuel and primary system transporth and Ken Lee will talk
!E. -about containment transport.
1 21 : With - that , let's go ?on ' with Peter on' the thermal L ' Br : hydraulics and -the (sequence description. 23 ' . MR. . CYBULSKIS: I am' going to; digress-a-little bit 1M- Tfrom the agenda ' item :and go overisome. perhaps rudimentary . V . 26 . points,;but I'think it'might be worthwhile. . t
- 4 t * +
o -. y g l 4 n[%,, -,%. $.- A-- <4 , ~e---r -
28
-1
- There have been a number of questions at the previous 2: ~ Peer Review meetingsfas to why we analyzed a particular 3 combination o'f events in a sequence as opposed to another 4 combination of it. That is, why is sis subsequence more 5 Limportant?. Why did you look at this? Why,did you look at 6- that?.
7 It seems kind of late in the game.but I think it 8 might:be worthwhile to run through all the things than can
.9 -happen--given a core melt accident or at least try to get an 10 ' idea ofLall the things'that can-happen and why we look at 11 some things as opposed to others.
- '12 Some of this is quite old and many of-you have 2 heard-me.say this before, I-apologize for that. But, bear 14 with me. ' Hopefully,.it'will add a.little more perspective 15 to some of ' the things that we' have done and why we have done 16 them.. I will try not to take.too long in this but hopefully 17 it will. add something to the understanding.
. 18 This is just a very simple representation of the 19 interactions that one is. concerned about when you analyze 20 - a core of meltdown accident. The thing that you are always 21 concerned about as the whole topic of this exercise, this
. !H program, is the fission product release to the environment.
M The fissien product . release to the environment 24 . comes about -through a number of interrelated processes. The 25 - meltdown thermal hydraulic,.what happens to the containment, J'_
m 29 El - 'how the containment fails -- and there are various ways of 2 failing an'd they have direct and indirect effects on the 3' sultimate f fission . product source term. 4 LJust to belabor the point.a little bit, if I take t 5 .the sameidiagram and elaborate on it a little bit -- and I 6 won't dwell on it -- but just to emphasize that the core 7_ . melts,-it' releases. fission products. It also releases gas
.a and vapors.which may.do'something to the containment. You e 'may have-wha't is called steam explosions here, but more to ' generally energetic interactions between hot debris and 11 water which may fail containment directly.
12 You may generate' hydrogen, which may fail the 13 ' containment. You may generate overpressure. All these 14- -things have am effect on' the fission product release and-15 - for.any sequence the magnitude, the timing of the fission 16 product- release is very depen' dent on which particular path
*7 .through_this maze you assume.
18 The-way'that people have chosen to_get through this 13 maze'is the.so-called cont'ainment event ~ tree. Going back 20- in history, this is the containment event tree that was
'21' . used -in the WASE 1400 analysis.- It is very simple, '
22 -deceptively simple. 23 In WASH 1400>we said, "Let's look at the containment
- u. failurefmodes. I'f something'doesn't fail containment, we l won $t' worry _about it, it has no effect," it is a. convenient
~
26 7 '7 T W T w
cm , 7 , :q -- m
. 30
~
^
'1 (assumption,..it-can.also'be a dangerous one.
~
.y .
2 '. So, theLeontainment failure modes 'that were
~
30 iexplicitlyTconsidered;in: WASH 1400 were the containment
~L 4 rupture'due.to an in-vessel steam. explosion; containment
~
~
?3b -failurefdue to failure to isolate initially; failure due to
^
~ 81
, b'urning;-long . term overpressure .and melt-through.
- v. 7 - '
In.the1 WASH 1400 analysis, if these did not' happen, 8' [this$had to.'happ'n. e At :that time I; felt melt-through was 8- finevitable if nothing else happened.
' UF Notice'th'at.the. alphas', betas;: gammas, et cetera 1 11'-
Lare still L the nomenclature that I like' to use. It is not
'u. necessarily always applicable.
2. But the WASH 1400 failure modes - were. these as they 14 - are identified innhere, and11n fact therthing'that controlled
'18 riskimore often' than -not was thisilittle rascal here, over-is pressure ifailure-.
17 'For'the WASHi l400 boilder analysis, the' plant was 2 Ta little more! complicated. There were;two compartments;
~
19 e there1was . secondary containment,. and both:: the -course and
- 20 :
cons'equences: of -the. accident could change, depending;on whether 21 . Lth.e" failure was:in the.drywelliand the wetwell. I believe 15 EwefhaveLeove' red'thispointatsomelength.intheprevious 23 . Peer' Review me' stings.
~
.M? iThe secondary could have an effect. So, the 25' :
containment. event. tree was a little more elaborate but still 9 8 t
? O
.~ _._
31 1 relatively simple. Again, the things we were concerned about 2 at that time were steam explosions in the vessel, steam is explosions in the containment, over-pressure failure,
~
4 ' isolation failure., 1 5- But here, as.I pointed ~out, we concerned ourse.lves, 6 where was-the1 isolation failure, drywell, wetwell. We got a 7 little smarter and we said,'well,.if it is a small isolation s failure ~it means one thing; if it is a big isolation, it 9 is another. This is the distinction here. 10 If the the isolation failure is small, your 11 secondary containment may survive.' If it is large, it may 12 not. If the. secondary containment survives, the stand-by 13 gas treatment system'may.do you' good.
- 14. So, the tree got a little more' elaborate.
15 The questions that you are really-trying to 16 l answer in the development.of a. containment event tree ---and 17 this gets worse as you go along so I will try not to go to 18 fast -- but the things th'at you are really concerned about up are in terms of defining fission product release some of
=20 the processes or some of the questions that I have outlined 21 here. It is not an exhaustive list _as soon you will see.
22 But you are concerned about , containment failure, 23 ' steam explosions,'whether they fail'your containment'or not.
-u Steam explosions, or interactions in.the cavity, whether they '
25 fail your' containment-or not, burning, explosions, long-term
o -
~
I 32 1 _over-pressure, secondary containment, and last, containment 2 melt-through. 3 Now, there are many other questions that can be 4 interspersed in here, for example, whether or not you have
- 5. steam explosions may depend on how the core slumps. I have 6 not explicitly shown those questions in here.
7 .The Zion study which we have heard so much about _8 went a long;way in the-pointing out how complicated this whole 9' -process can.get if you-really try to examine the tree in 10 - detail. With due-apologies, what I.have done here is copied
-11L directly the containment event tree questions that were used 12 in the development of the Zion containment and event tree.
-13. I won't read them but just give -you an idea.
14 They looked at containment failure before core melt. 15 They looked at containment failure during the core melt. They
- 16 ' looked at whether the core failes coherently or inccherently, 17 'and they asked several questions, particularly questions about is hydrogen burning at several key points in the accident le sequence instead of just asking them once.
, 20 : They, in the Zion. event tree, came up with, I c1 believe if I counted correctly -- that is the-last-of them -- 12 .- there were about-19 headings. 23 : Now, . if you go through f the arithmetic, however way St4 you choose 11t, I think twoLto the 19th. power comes=out'to 1
.:s sorething:like.528,000. combinations.- Not all oof- them are
)
33
'1;
.necessarily feasible, but-just let me give you an idea what 2' kind of a. containment event tree.
3 Again, this is streight out of Zion, with due 4 apologies. This just considers events prior to core melt 5: which-is equivalent in size to the-entire WASH 1400 event 6 l tree.. 17 This is the next step.in the process which treats 8 only-the events related to the vessel phenomenon, to the in-9- . vessel phenomenon. I hope the point is being made that if 4 10 you try to look in great detail at these things, the 11 . process gets more and 'more. complicated. The possibilities , 12 as to what can happen get more and move involved. This 13 ' represent s the event tree that was used to treat the x-vessel 14 phenomenon. 15 These things will be in your handout in-case you
- 16 have not seen them,:and you can study them at great length.
17 But what the Zion containment event tree -- and I 18 - won't flash all of them on -- eventually wound up with and 1st I think Ifcounted them correctly. I am sure somebody will 20 correct me if I did not, I believe there are 1,047 distinct 21 branches cn1 the containment event tree in Zion, 'which is 4
. n- . combinations of: possibilities that were. considered to be I l
.2 reasonable, 'given a core meltdown.
24 Now, many of'these combinations led to benign 25 ' endings in the sense that they-led-to-~ successful termination l
; .7
- 34
'l of: the accident.
But many of them~ led to various combinations 2' of : failures or failure modes.
~3~ The point I am_trying to make is that if you were 4 going to _ exhaust an accident sequence in any sense of the =
5 ?- . word,' the' Zion containment . event tree might be a good starting 6' point. 7 We don't have-the time or the resources to talk t about 1,047. combinations.of event trees in this program,
~
81 9L : particula'lyfin r terms.of the details. So,-to illustrate what some of our. thinking was, I have shown a relatively 10 -
~ 11 simple ~ containment event tree, and I will be coming back to
~
Ut- 'this tree as I talk about=the specific sequences that we 13 d ~ looked at.
- 14 There are a couple of items that are.perhaps 15 notable about this tree. I have ignored in-vessel steam
-16 explosions and the likelihood of in-vessel ~ steam explosions
~
17 f causing-containment failure. If I had included them, the 18 tree would be four times.as big. But I_have: included some.of 19 the key processes that we did think about and some.of which-
~
20 we did explicitly include in our_ thinking, if not in our
- 21. analysis.
22 . The things'that we are concerned-about.are 23 ' ' containment- isolation failure. : CX'in my nomenclature here
- 24 areEcavity steam explosions or cavity interactions.just 25 occurring. Then the next column is cavity interactions
35
-1 'failing containment. The-next question is, does the 2 ' hydrogen burn? If_it burns, does it fail containment? If 3
it does not burn early, can it accumulate to where you get
-4 'a hydrogen. explosion? Long-term over-pressure, secondary L5 _ containment failure.
6 Now, I have used secondary containment very loosely in this sense,1not all' reactors have secondary containment,
~ - 7l 8- and.I also include _ auxiliary and safeguards buildings if they ~9 are appropriate.
10 - Then, eventually, the melt-through.or, more generally, 11- do you really get attack of the concrete by the hot debris, m I forget what the numbers were, I think there-were some forty 13 combinations"on this relatively simple tree. g 14 Again I point!out that~ not necessarily each.one 4 15 - of these. branches of the tree-~is . applicable to each sequence, 16 and I will be coming back to that point. 17 But again, if you are going to look at completeness 18 - in-any real sense of the word, there are many, many combi-19 . nations that'are possible. What I am_t,rying-to point out is 20 that-it is'not possible for ust to look at'them. We have 1. 21 pickedisome'that we thought were more meaningful than others 22 z and when I talk about the particular sequences I'will try to n- : point out why we' talked about some rather than others. 24 Any. questions at this point before I talk abcut the 26 Sequoyah1 sequences? Everybody must have. listened real well. w D
36'
~
1- . Going on:to the sequences that we have looked at
. '2. in the casefof the Sequoyah ice condenser plant, we talked i
~3 about .this at the Llast meeting so .I won't go back over what 4 the plant looks like, et cetera.
'5' The sequences'that we had identified as the ones 6L they L looked ' at are TMLB -prime , TML, and S2H. Again, TMLB
-7, prime :is loss of. electric power, no makeup to the primary 8! . system, no makeup'to the steam. generators,' no containment 9 safeguards.
10 The1 TML sequence- is -:the same as TMLB in terms of
.11 : no makeup to the primary system, no secondary makeup.to the U ' steam generators, but the containment safety ; features are 13 ' .indeed available.
. 14 The S2HF sequence is a common-cause failure of Hi - both~the recirculations', emergency core cooling recirculation 16 'and the containment-spray-recirculation, a sequence that was 17- identified.in RSSMAP asJhaving some, at that-time, fairly.
18 . sign'ificant probability. due to failure ' to open the drain
-.19 ' lines to.the sump and exhausting.the sump of water after 20 some period of operation.
21' Let me try- to go back to my event' tree and try to 22 - indicate'where?in-the great scheme of things some of these L23 sequences. fit.
~
24 If I go'.back:to my containment event tree and try
-to depict the TMLB prime sequence.
25
.The "Xs" here basically y
m
n: ', 37 1 mean -- I don't think they are based on the evaluations that 2 we have done at those particular branches, they are not 3 applicable to this_ sequence. 4 So, given.the sequence as_it is defined, it is l 5 .very unlikely that you will escape containment over-16- pressure failure. You may or may not have burnings. I 7 have wiped out these_over-pressures. s- .Since the core meltdown-as we have calculated in 9- the current analysis takes place withfthe significant amounts
~
10 of f ice'still in'the ice bed, it.is unlikely'that any debris 11 water interaction in the cavity that generates steam will
- n : fail containment. Soi I..have dropped out this set of
'13 - . branches.
14- Then there is a whole variety of isolation 15 failures-that are at least theoretically possible for this 16 sequence. 17 - But the-key -- incidentally, the check marks are is supposed to represent possible sequences, and then there
'19 are a few-in there that don't have any check marks that 20 perhaps are questionable or I did not think about.
~
- 21. But given the sequence-as it is defined, the key zt' containment: failure-modes that one is concerned about is
- 20 the' delta modes -- and'there are several combinations,
's actually -- which.is long-term'over-pressure.- If the 25 hydrogen does i not burn or if the hydrogen burns in such h
- - + -r-
a 38 11
'a way asJ to not' lead.to large pressure increases, in this 2 . sequence: eventually you will over-pressurize the containment.
- 3 The other ones. are -the gamma, and again there are
- several possib'ilities. If the hydrogen burns and burns is.- in large amounts at-once, you will probably have large 8- pressure. increases and you will: fail.the containment. There 7
~
.is-a variety of isolation failures.'
8 _. In the. work that we have done, we have looked at: 9' both delta and. gamma,1namely a.large-hydrogen-burn failing 10 - containment and a no-burn case where the over-pressurization 11 'is the case. 12 I-trust that everybody remembers what we said
'13 about;the flow paths in the containment.
Maybe just to
-14 refresh your memory, the time scale-for.when things happen, 15 we are-talking about the TMLB prime sequence in the ice condenser.
~
16: In our calculations we get the steam generators
~
17 drying out.in about an hour. Core uncovering, 98 minutes. 18 - ' Start to melt, two hours. The head failure,.158 minutes. 19 In this particular case we ' predict a _ large burn
- 20 .right after a head failure. ThisDis when the containment
' 21 ' fails..
in If you take'theJelternative sequence, there are
~
~
EM no power stressescin here. .You'say there are no admission
~
2 24 sources. The scenario is-identical up to-the point of head 25 ' failure, but instead of'a large burn you get slow over-m y ..m,<,
y - [W . m[
^ '
39 e.
.. ' l10 : pressurization'and your. containment failure takes place ge *y*
22 J considerably..laterfin: time. s
'3 -
Ilthink there w'ill be--a . number of- core node n s l43 -temperatureshis' tories andiprimary system. temperature
- 1. " ' ,
~5" ~ . his" tories . - IVam-not sure : thatiI should take the time to t8 4talkfabout them iniparticular, I" don't' think there is. anythinc r-
~
~
4 1 7 : re atively L st'artling. t I8 .Letlme saylaffewiwords about how we treated the c8: ' primary system a's;it. applies to the-fission product release. 1 s 1101 'This[is just a> schematic.. The' MARCH--calculation gives'us
.,o.
' 11$ 7 thelcore heatup and theigas. temperatures?at the exit.of the i
~ 12 ~ teore. L Of course ;: MARCH : puts (them -into ' the~ containment. -
2151 For purposes :of the XTRAP- MEI,T calculation, ~ we 114! wentfto;a little more-de' tail on;what:h'appens in the upper . is plenum.and structures, and:we'.take.thejoutput from the. MARCH, 16 - l feed : t! hem -.into MERGE . We. pass-themithrough-the upper' plenum.
' 617: In this'particular~ set'"of; calculations we treated 1the : upper' plenum as 'a- singleo v' lume but we put four different
~
q- - 218 . structures'in'it. TheEfour;different. structures represent, 20 : 'firstofall[the. upper' core;plateandthestructures-
~
( 21 Jimmediatiely.fabove : theic' ore, that' is the first structure, the NI_ ? , [ ' : 22 (second1 structure, Tare the' control rod guides--and support columns.
~
23 t 4 The thirdistructure:is:the so-called top-hat
,t a
- .i e (se} [structureabove1that,andithe;fourthLeolumn-isthecore T
R '
- 5. LizarrSl. -
1 + se - , eq ' s
.a 7 , ,.7.-,_
., , , . ,p., ,9 , , _ < , ,_,g, , ,,S..g.y, ,, . . , , , _ _ ,,,g y y. , ,.g,, ,y r, , , ,,
a <
=
~
# s J q, .4 5 l1 ' Then, afterithe. gaseous vapors, fission products,
. ?2i pass c hhrough the; upper plenum,: they pass into the piping.
L 1 3 landfthe-pressurizer, We have_again for'these:particular
~
41 ;structuresLlumped those as two structures, one representing
,-: 5 S the'pihings one'the pressurizer,:and from.there it goes'to- 1 6 the containment.
~ '
' In some earlier [ analyses we had broken down the
~
' ~
~7 1
8 upper 'pl$num in a' series of series-connected structures , 8 Teach one containing_a-s' ingle structure based on the limitations '
~
-"- 2 ~
2 of the earlier code.- IEam_n'ot-. exactly:sure.which is the 11- .more correct way, but particular: treatment we have here
- n : now implies that the gases in the upper ' plenum are relatively -'
13 _ well mixed, as opposed to' being a straight plug flow-through. 14 ' I-think perhaps I-should put.thisLup. These are ;
~
3' LMERGE-calculations of primary system temperatures and [ structure; temperatures after they come out of-the MARCH
~
~
16-b' 17 code, have( gone _through the MERGE Ecode , and Nave -undergone
~
!181 some processing.and-averaging to' enable.the TRAP code to 18 - -handle.
- The significant~ points -- and this is the. question
~
L2 lE
. fI.think that-Bill!Kastenb' erg raised a minute.ago about E- differences between MARCH 1 and MARCH 2, and~they are not I E entirelyLMARCH 1 and 2hdifferences but there is also the .
m
~
24 choice of options in.the models.- 25 This -is the gas temperature at the exit of the L t . - - 4 , 4
+ ~c y- 4 y - c,*- + w'f m -
em- + , asw' -+ , c'--"c.,-- - - - -
l 41 e 1 core.- You notice that.it does not really get very high, 2z -something like 2200 degrees Fahrenheit,- before it turns
.around.
~
4 In the calculations that we had performed the 5 firstLtime through with MARCH 1.1, using a different code 6 with a different set of options, with a dif ferent set of 7~ Linput parameters -- and this is where I have a great deal 8 of problems:because it' is ' difficult to sort out what is 8 most-important.- Certainly, the differences in the code are 10 important but the' differences in the structures are also 11 4 important. 12 But anyway, in the MARCH 1.1 analysis, I think 13 ans are predicting gas temperatures in the 4000 range and 14 we are predicting structure temperatures above melting. 15 -As the gas temperatures have stayed lower, obviously, so 16 have the structure temperatures.. 17 Now, again' going back to the MARCH 1, MARCH 2 . :18 c'omparison, in the MARCH 1.1 calculations we used what I 18 frequently call the coherent slumping model where we told
#- the code that there is no' fuel movement-until three-21- quarters' of the core is ; molten, then the core. drops'inno
'E .the bottom head.-
8 That is one of;the principal reascns for the 24 -high temperatures. If you holdfthe core in place that long,
# the1 temperatures just keep going.up.and up.
I-
- ,- _e -, -
w, -
. 1 42 l
2 ' 1: The .other. reason for. the high gas temperature is 7 2 -in MARCH.l.1 as1 opposed to MARCH 2 was that the heat 3 transfer coefficients between the gases and the core in
-4 ' MARCH:1.1 were based'on a relatively simple correlation which
;gave_relatively low numbers, so that when the gases heated
~
5 6 up they stayed hot.
-7 The heat transfer _ correlations that we used in s ..
8S MARCH 2 are considerably. more elaborate, other things have
- 9. 'been'added. The gases follow the' core temperatures very 10 closely, both on'thefup side as:well as the down side. We 11 still get very high' gas temperatures inside the core with 12 ' MARCH 2, but as the hot-gases flow through the upper portions 13 - of the core which are not as hot as the center of the core, 14 -they cool back down.
15 ' This is the reason why your ': temperatures at 16 the exit of the core never. reach the temperatures that we 17 ' have previously predicted. 18 - The abrupt drop ~in the exit' gas temperature, which 19 ' is' kind of. obvious .there, occurs when the core starts to
# slump.. Now,-let-me again back up for a moment.
- - 21 In the current calculations we are not waiting 1[ for three-quarters : of 'the core; to : be molten before we start L21 -: slumping. We-are starting to slump the core when the lowest 24 -node ~'in'any region issfully molten. ~When that happens, the 25 molten' fuel:in that region is, allowed-to fall to the o
g g 'W -
'y .x -
3= 43-
"1 structure below'it.
p: '
~
?.' . r .. (2J J As it falls'to the'. structure below it, it sees
! L3 'waterfycreates?a:large'. puff of. steam which does a number of things,LoneLof which'is brin'ging the gas temperature down,
~ -
8? ith'e_-oth'er'is.providing steam to.the cooler portions of the 6- ' core-toEreact with the-hot zirconium,'and there is~a rapid
- 7I: success' ion'of events;-that us'uallynleads to a fairlys rapid s .collagise.of:the entire head and the bottom: head-once slumping 8f iisEstarted,-but~a' Iso l leads-to?a large-puff -- puff is_-not
~
._10 a .the: word - _but a large! rate'of' steam.' generation associated ill 'with:the. core. slumping..andLtends'to drag-the temperatures g.
12 down'.
- 13 --
MR. WALKER: On.thatr. curve, are.those both gas 14 ~ ;and temperaturcs? :I. don't% understand --- 15 'MR. CYBULSKIS:- That-is the gas-inlet,_that is
'I am, sorry, 'it is~ just a ' mislabel. on the
~ ~
_16 the; gas' outlet. - 17 ' figure..:Then, the;four' structures. This~is~in the upper
- 18 :' , p lenum. -
? [ 18 iMR. HILLIARD:(;Pete,'can I-ask a.. question?
'^
'20 ,
=MR.'CYBULSKIS: S ure . --
+ 21 [MR.:HILLIARD: JIf-those temperatures:are lower than
' 22 :
- the} MARCH 1.'1, what else happensLto.the heat?: 1Do you have
- 23 theTsame' heat' source,Jdon't you?
' ~
' ~24 MR.-CYBULSKIS:~ Yes and no. 1 We have the same heat
~
L28 source as 7 far.Tas the: power generation rate. .0bvicusly, the 1- ,1 J
.m. _ . _ _ _ _ < .
P 1 l _ -44 , t 1 amount of: metal-water reaction which is equal or larger than i i
- 2' the decay heat is somewhat different in-these calculations
. ~.
3 and-because of the differences in the heat transfer ; i 4-coefficients, numberione, the differences in the surrounding 3
~5
_ 1 5 structures -- I mentioned that the input to the surrounding 6:
-structures was somewhat, or the description of the surrounding 4
7' structures was:somewhat-different than in MARCH 1.1.
.s- E We also'have radiation to the core barrel in i 8
MARCH -2 which .we -did . not have in MARCH ' 1.'1, and a number of 10 '
- other things. The heat--source is approximately the same,
' ll
.though it may be somewhat-different -- I guess in essence it 12 is distributed differently. It is spread out more than it
-i 13 was in-the MARCH l.'l calculations in terms of core 14 structures, what have you.
15 So that, one, you don't hold-the hot core in
- 18 :
place as- long as - you. did because the changes- in assumptions ,
- 171 1
so the gas temperatures never get as high as they did, at 18 ileast the gas temperatures at the exit of the c ore. In core e 18 : they are as high.. 20 -.
.Perhaps the simplest way-to look at it.is, you '
. 21 3, are' distributing lthe heat:somesthat diffarently. or the eneray.
4 3 . tgl. RSILBERBERG: State your name, please. Do you ,
- 23 1
'want'to use the mike??
24 -
'MR.fKELLY: ' Kelly from the University of Virginia.
M fI:am still confused;about your nomenclature. The 4 s e e-- r - , ,-o- -- - ,-a wnv- ,-~
r .n - 45. 1 op! curve ? represents ;the temperature of the gas inlet to _2' ithe upper 1 structure, aor :the ~ gaslinlet to the - core?
> 13 < - MR.:CYBULSKIS: No,.it is the gas inlet to the l 4
.4 upper structure, it Lis the exit of the core.
lMR..' KELLY:? And.the next curve.then,.the 5 8 6 i. triangles,~is' inlets.to-the~ core? 7- 91R. ' CYBULSKIS : No, this is all out of the' core.
'8 We pihklup the gases.:at=the'~ exit, . the top of.. the core ' and lookiwhat 1happens-to them as.they pass through the
~
19
- 10 l structures above the' core.
'11 MR.: KELLY:~ So, theLtriangles represent the gas U: ' outlets 'from thE upper structures?
13 - :. MR. CYBULSKIS : That isicorrect. 14 MR.; KELLY: ,The gas outlets. 2l MR.;CYBULSKIS: Yes. -That is mislabeled there, I 18 ' am'sorry.
. 17 ~ NR.fWALKER: hete,-could:I'ask you one,.what 18 you' assumed? You talked-some~about the' core slump model
- 18 .being different, the material dribbling down into the pool 7
20 J -ben'eath the- core and _ generating : steam. 4
' ~ '
/
What- I' did' not understand was', whenithat first
~
- 21 ~ -
- s ' 22 Laaterialicomes[down,Jdo:;you assume a. reactor vessel failure,
.- --23~ 'Ldolyou: assume you.have steam generation that sweeps the 24 vessel and carries the material? .
1 N:
- MR.[CYBULSKIS: Oh , no.- When the' first : material . falls P
l a
46
'l. downtit'can:only fall as far as the first structure, which 2 :is not very far because the support plate is right below the 3 . core..
16 ' But typically, if-the: support plate has water
.5- around it,;it'will evaporate to water. If it does not have
- s water around it,'it will attack the support plate, the 7 first support plate;and fall down to the 2next support plate.
4 8 MR. WALKER: So,'you-are basically still assuming s 'there is Lnot -- sweep of. the vessel before vessel failure 10 as a result.of the material falling in the-pool.
- 11 MR. CYBULSKIS: I am not quite sure I understand L2 the question. LThe slumping fuel generates steam,' typically.
13 -Then the steam!is allowed-to do whatever sweeping of the 14- vessel;it would do, based on how much steamLyou generate, volume of the system pressure, et' cetera. 16 . That is all calculated consistencly. MR. WALKER: The question is, how much time do you
.18 Lallow the material to interact with the water and generate 19 steam before the vessel fails and you lose the material out 20 of'the bottom'of the' vessel.
21- - MR..CYBULSKIS: The time is not constrained in the 22 : sense that'we do not say that it has to happen in this time 23 or that time. When'the fuel falls onto a structure, the 24 ' > code checks to see'if the water: level:is above or below the 25 . structure.
n- ., -
.f q.1. , -<.:
47
~
fil , 'If the' water 11evelLis -abo'v e: the structure , you will
~
-2 evaporat'e ithe ' water. lNow,Efor the evaporation of the water F ~
!s .- 'I think'in'all these_ calculations we used something like a 4' debris bed model to control the ' rate of evaporation of the
, [5 water,;;as opposed 'to ' MARCH .1.1 where it was evaporated in e- one. time step. 7;
~
So, it1may-take a.while. Again, that is based on 8 the3 heat transfer' energy,,etl cetera. (si . When the; support's'tructure becomes uncovered, then to thel debris is allowed;to' interact with the support structure,
, j 11' Now,1that is. basi ~cally' an1energyf b alance calculation. But Et again,.the water. level has;to have receded before the debris zg 13 can attack the structure.
F 14 kThen, if the debris fails the structure, ' the debris
.m' is allowe'd ta) fall to the next structure, which-would be the is second grid, as we call it.' 'Again,fthere is water involved.
1t7 Once the second grid is failed, then you fall _down to the
- 1s bottom head.- So,.there is a step-wise progression in these n 13' calculations.
! 20 'MR.-WALTER: -Well,'then.-let me just ask the question 21 infa different way. .In'the initial Surry calculations on
- nl TMLB prime.what was happening.was~ --ithe question is, has
's:
that situation changed'in these calculations?
. 24 .- .MR. CYBULSKIS: I'am not sure that I know how to NJ 25 L . answer'.the question directly. lAgain, the flow out of the
7 _ _ b ' 48 F .1
; relief valve is a function of how much steam you. generate 2- Land the-pressure setting..of;the relief valve. When you get 3
core. slumping into.the bott'om head, you generate a burp of 4 steam which will. lift the~ relief valve and releese some 5 fraction of whatever is there. 6 I can't answer.the-question in general as to how
-7 much does.it sweep, it obviously will sweep some fission 8
products and some will' stay _beh'ind. Perhaps when Mike gets 9-to his presentation, I think he can.shed some light on.that 10 . question. 11 - There is no a priori determination that the fission 12 - products ~ are either swept out or retained. That is why we 13 are _ going through the exercise. We are trying to calculate 14 that consistent with everything else that we calculate. 15 I-think Mike will get'to that when it is his turn 16 since I can't answer the~ question. 17 MR. LEVY: I-havoca-question, Sol Levy. 18 You know, I think this shows some real major
. 19 issues.= _You know, the first one, it ,is apparent that the 2 one where you assume the; core _ melts makes a lot of difference 21 ~
in terms of the entire behavior. It actually seems to be 22 a major impact in terms of temperatures.- 23 ~ I.think it raises.some key questions about_which N- one of these you should use. 2 The second one, you know, that one has to raise a
y
~
;q ,
- <.1 R a; , i
~
_ y ;, - j 49
)
11 so'me-questions ~about is,-as~you melt this material, where
.o,;*
81 idoes it stop?- I think'you probably would. catch it the next 3 setifof:spaciers rather thAn;letlit go pretty far down.
~
s u:- ' 4-so, I-think'there could be some. arguments-about 57 ;where . the ; stuf f istops'. . .
- 6 TheJthird point, I just can't believe the gas-gets 7
cooled. .You'~still2have production'above it. 'Unless we 8, even svented -a very high,~ gas' lieat transfer ccefficient there 1 8 is no.way;to'..get that'.gai_coole'd, you know. 10 So, it keepsJcoming back to the key issue, how
' ' 11-
-well'is this' MARCH 2 checked _out;and verified. Then we 12 -
compound it ~by.. changing the wh51e melt process, so we can 't 18 4 -even-compare _one to the other. 14 ' MR. CYBULSKIS :' I:think some.of the points that is ~
~
you have raisediareLobviously perfectly valid. There are 16 -
-. major differences between what we were doing some time ago
- 17 anc what we-are doingLnow.
. 18 1 From 'our' viewpoint, we; fee 1'a' lot-more comfotable-18 with the current slumping models. ..There.were many_diffi-
~
20 1 culties' with the1. slumping models that were used;in the~past.
-.21Y ;
m Remember.that this gas cooling-down:here, whatswe.are showing
' 22 -
l:11s an integral effect, it is not only the gas through the 88
? hot regions'and the co1d regions.s There.'is:rather dramatic
~
1 24 decrease-'in the temperature'.
. 25 X I believe, without getting into great detail, the i7, .
j.-
Y, E ,; '
, Kyygr y ,- p .
ro'
- a. >
^
EN p p . 50
- simpA #e ans,wer for the dramatic decrease in temperature is
~
.1 y y 4<
-that you.are generating a' lot-of' steam that is much cooler /.
~
2 4
, 7;
~3 f,- f
( / .' .. y than the steam that)you havelbeen putting through the core.'/ y, ' .~l\ 4 . 4 MR. LEVY: You know, I'could raise 'a-lot of heat 5 : transfer issues on that. If you are going to get to that 4
., s
~
6 sophisticatiari you'should treat' all the gases as a simple . homegeneous\ $ l~ . I think you are going to get some l ,.;; i 7- mixture.'
'e. . i,
,r
. channeling l in /that core and you (are going'to 'ha've, 'to put that 8
v 9 in.
) .. .. .
A, d'
~ 10 7 think you have to decide.to what; degree of ,
Y },i; .,.
~11 sophistication yo1 go andsa hen-do you cut it off. I am y
~( p-
-t, ..
3 ~ ffrank'.y.very!greatlyuconcerned about some of the trends we 12 a-are not shir$/ing.
~
13 I think particularly when you start to . ,7
~. ]
14 h show thath. cme of the gas - I think[Ifsee a reversal of ,
.,: >N
. temperature ubetween inlet. and outlet, which I din't- figure
~
15[ y 16 out,on your graph.
~
- 17 ' d e al4 of a sudden have -the LNeat/ ;oing the . ;
i Lopposite way. 18
-3 19 ,
MR. CYBULSKIS: I suspect that:obviously the
' ;7 heat doesEnot go'the other way, but something-gets lost-in 20
( 21- the.trinsiation t of the detail'ed calculations into an w
;overall picture.
;-l 22k
.As far as we know,-we don.'t violate any funda-
.l23 M( .24 - mental". lays .
! 25 y .MR.. LEVY: I hope you don't violate the first; law.
t t
~ - 1
% w ,
p: ,, . y j- .,f x
~
,h, , ,,
N N-a 51 u. 1 MR. ROWE: '.I have a question, Donald Rowe. , 2 The question is ~ related to that and also just to 3: the presentation and the report. I notice that all reports 4 include.the very simple nodal or control volume diagram
- 5. '
.that shows the core, the primary system, and about four 6
boxes.
'I I continue to be frustrated when I read the reports 8
.to' figure out what inIfact is a control volume. When I come 8
to.the meetings I. find that the structures are identified or 10 mentioned, but'I don't really know where they are. I don't
= 11 know where a control volume'for a thermal calculation is 12 located.. I don't know where a control volume for a hydro-i
~ I3 dynamic calculation is' located. I don't know what flows 14 from one volume.to the next. I don't know how I would ever 15 close a heat balance.
+ 16 Somehow, infthe: presentation 1I think that a lot 17 ' of the. questions like Sol raises, the answer in the minds 48 ~ of the reviewers and theLreaders.of the report is a clearer l' . and 'more concise description of the control volumes could be i 8
-included in the report.
21 I am not asking.for-all ths. nitty-gritty details 22-of the calculations to be shown, but I think there needs to be;something on a conceptual level'that is readily identifi-iMale with, here is the primary system or here'is the
" reactor system; here is how it is subdivided and here is what l l
l l
SIN
'1 goes-in here, comes out there; heat flows from here to there, 2i et cetera.
3 MR.'CYBULSKIS: Well, again some of us that perhaps 4 are.too close to.some of these things take too much for 5- grantedLin our presentations. 6 . , In the report that we ~ circulated prior to the first, (7 -
' previous Peer Review meeting, there was a series of diagrams 8
for'the Sequoyah plant and in particular where we tried to 8 indicate the flow paths. I. don 't know whether that' would 10 T
'try to answer the questions: that -you raise or not.
.11 They were primarily addressed, though, I believe, 4
M how does the' flow.go from the primary system and where does it 13 go in the containment. I am not sure whether that would 14 answer your question. , 15 MR. ROWE:-'This is a comment, I am not sure that 16 - itican be resolved - 'Ilthink in terms,of presentations and 17 ' reports as this process evolves, I.think that is something 18 that'is worthy of some:a'tention. t s 18 MR. - SILBERBERG: Don, you have' expressed'this
-8
-before, this_is'not the first tin}e. ,W e will have to'give
' 21 ithat.-some dttention as to-how we can do a better; job of 22'
+d'isplaying:--;unless.we are not' communicating.
23 :: MRi-CYBULSKIS: Well,;ILdon't know whether this M' is ' the appropriate time to ' back up or not, but.this is a , 8' ~
- picture' of . 'that, that we had-shown in,the previous presen-5?
e + nw - , ,
52 1" tation. I' chose not to go back t'o it for purposes of brevity. aa - 2 But'perhaps.it might be appropriate. This is our schematic. t: 3 MR.-KASTENBERG: . Pete, what Don is saying is that 4 for example you show four structure points, and on a diagram 5 liketthat.he would like to know whether.those four points (6 ?are that you calculated. 7 'MR. CYBULSKIS: Okay, 8 MR.~ KASTENBERG: I_think that is what he is saying.
'9- MR. CYBULSKIS: Would'you like me to try to answer 10 that right now?.
11 EMR. ROWE: .When you said it verbally, I know what 12' you are talking about. But'in the report it says,.four 13 structures ~and I don't know what the four structures are.
' 14 MR. CYBULSKIS :. I'm sorry.-
.15 MR. ROWE:- That. raises the question --
16 (Simultaneous . conversation)
'17' MR. LEVY: . I'would like to see a picture of how 18 you are melting the ' core because to me it is fundamental.
19 -MR. CYBULSKIS: Now, need I go-back to the diagram 2D '.or has?the question'been1 answered?
'21 MR.;ROWE: You have answ4 red -the question verbally 12 alreadylfor.me. 'But Isthink -- well --
121 MR CYBULSKIS:- I get your point.
~
, 24 'MR. ROWE: .Okay.
25 MR. . LEVY: Can you' describe your-core slump model F._..
53 1 in a' lot of' detail about where the multi material goes
.2- and stops, and'why it stops where you say it stops?
3 MR. CYBULSKIS:. Okay,.let me try again. In the *
'4 . current analysis we nodalize the core and I think it is 5' . described in'the report in ten radial regions, 24 axial 6 . regions. There is a power ~ associated, power level associated ,
.7 1 ~ with each node'.
LTypically,.you start out with the core covered
~
8 9' and' balance and make up, and heat generation, the'1evel
-10 drops. .Atlsome. point you'get to the point where the core
~
11 uncovers, the level-decreases further, you start to heat up-H -the core. 13 We calculate *the temperature of the core and the
~
a 14 ' temperature of the gas next to it-and each one of these . 15 nodes,.and we calculate the melting at each one of these 16 nodes,.using an input' melting temperature. 17 Slumping. When the 24-axial nodes in the region, 18 when th'e lowest node in'that region'-- which means the
~
19 bottom node in the region --'is fully molten according to 20 our' calculations, we say that that, the molten nodes in J
- 21 that' region are allowed toLfall.~out of the core.
H Now, again just to backtrack for a second, 2 - typically the melting will' start near-the center of the 24 core'and proceed downward. So that when the bottom node is molten.it implies that there is some number of nodes
~~
25 t
54 11 above it.- maybe all of them, maybe not, that are molten. 2' Now, when the-nodes. fall out of the core region 3- they' will fall to the lowest cnr to the nearest structure
-4 below.it. '
5L In the-typical' calculation we represent the structures below the core by two support structures and the 7 lower vessel head. - So, there is a. total-of three structures 8 belowJthe core as we have described them. 9 When the molten fuel falls out of the core it will 10 stopLat.the.first~ support' structure it encounters. That is 11
-basically an assumption-in the_model. ,.
12 In the earlier versions of the core, or actually ' 13 I;think there are_ options in the~ code that would let you 14
' drop the - fuel all the way to the bottom head. In the is . calculations that I presented, we have said that the fuel
-16 would stop at the first structure that it encounters.
17 When it stops ~at the structure,.it tries to 18 '- evaporate to water at that structure. If the structure is 19-already uncovered, it will attack that structure and fall to 20
-theJnext-structure. This may-take a number of-time steps, 21 :obviously.
H~ In the meantime -- or it may even take a long tsme, 23
- depending on the-accident sequence. In the meantime, the
'N fuel 1that is still in its more or less: normal configuration, 3
the-history of that fuel is being-followed, the heat-up and k .,e
;1
'55 i
< 1 :- / melting:of_that is being-.followed.- When the next node or 2 whenithe next region-reaches'the~ point where it is allowed
'to slump _it will- go through a .similar progression.
~
3-4 1When we get to the point where an input fraction,
- 5- 'and typically (it.is'three-quarters of the core, is molten,
~6L
- we sayuthat1the whole core is effectively molten and we letl I 57 ~ it drop ^into'the bottom head-and transfer, in effect, the 8' < control:ofothe calculation ~ to-a:new portion of the core.
9- LMR. LEVY: You know,-that model bothers me quite 10 a bit' physically, and its answers. Let me tell you why.
- 11. .First, you put so few structures that when a Et material gets molten-and goes-all-the way practically out 13 the core, I think in the normal process it will be caught on 14 lthe way at spaciers.
Lis ~ At those. spaciers , you will: form new obstructions . 16 - The gas will- redistribute itself,you would get less gas
.17 - flowing'there where you.have those obstructions. You will 18 : get_much higher temperatures in real life than what you 19 -- calculated.
'# There is noneiof that in there and I think the model
' 211 you have, I have some' serious. doubts whether it is giving the
. 22 - -right high temperatures.
23 MR. CYBULSKIS:. I-am sorry, I think you misunder-
-M stood what:I said. We do not drop the nodes out of the core
- 25 when'they-first' reach melting, we drop them out of the core
= ~ , - I 56
~
1 when the bottom node in the region is molten. E 2 So there is, if you will, significant holdup of
, '3 fuel'. You have to wait until the lowest node in the region 4 is molten.
5 . MR. LEVY: Do you redistribute the gas to 6 account for the fact. that now there are a lot of these
.7' obstructions?-
8- MR. CYBULSKIS: -In these calculations there are 9 no redistributions of the gases. That is a valid point that 10 : .you raise. 11 MR; SILBERBERG: Excuse me, Sol, are you saying 12 that the' gas temperatures would therefore be hotter rather 3 - than colder? 14 MR. LEVY: The gas temperature-has got to go much 15 hotter-than what we calculated. I have breached this issue 16 before,-we have to check this MARCH 2 carefully. 17 - I have called for some kind of a qualification, 18 validation of MARCH 2 before.
'19 MR. SILBERBERG: Well, the subject of the status 20 of MARCH 2 qualification and validation is a subject of
- 21. our Element l report which the Oak Ridge people have been
- 22. pulling together. That subject will come up tomorrow, I 23 : : believe, on-thefagenda, Tom Kress.
' :M - But I think your point is well taken. The dilemma Mi - and,--if you will, the concerns about' MARCH -- even up to
=
+ ,
57 1- . MARCH ~2'-- that methodology is with us. Certainly there are, 2 ~1f.you will,~ improved more mechanistic treatments that are
.-3. Lunderwayfas part of our-on-going.research program, particularl y 4
alprogram'at Sandia,-working on a more mechanistic approach 5' tofthis. 6 But they all obviously' await some form of validation , 7- if you'will, physical' evidence. That is another subject.- I
- 8 -think we could trylto put that a little bit more on perspective 9
on status during. Tom Kress ' presentation. .But it certainly 10 ' 'does not help you at this point. 11 -MR.; LEVY: I think the point I am making is, we n have to be .very careful with these models that we do not 13. put'some extra sophistication some places wthout bringing the 14 : sophistication to the same level in others. 15 I1think once you start to' move the molten material 16 I.think you have to also recognize it is blocking and what 17 t it does to the gases. If you do 'not do that, . I have some 18 - concern. 19 - So, I think we 'seem .to be doing this quite a bit
. 20 in this-' modeling, that we move in one area without moving t
21 .the comparable effect in this particular case. For all I
'E- know, you know, maybe ' these answers are - right, but I am far
-2 from convinced. -My first intuition tells me those gas
'24 . temperatures should be. higher ~if you have some molten material 25 L that Lis - running very hot;- Where it is hot, the gas won't be ,
u
g. I
~
58 l
;11 -
;able"to.get-there.and so'therefore you are going to get some
- 2 chigher temperatures. Eventually,.if you try to calculate 3
'homogenous . points you are. going to get higher.
4' But I am bothered with some; elements. We seem E 5
-to take a lot of details without. moving the others to at s'
.least some kind of'a step' forward.
MR. CYBULSKIS: Without belaboring the point, I
--8 think-your, concerns are:well taken',IDr. Levy. The MARCH A _so-called slumping models which obviously are not the final
. 10
-word'- ILam not sure if there ever will be a final word.
11 .g, are obviously concerned about them, but in 12 order to do-the calculations we have.to use what is 13 available. We are"very much concerned about things like 14 the slumping models and-our inability, if you will, to
- 15. -
Ecalculate, things in a more mechanistic fashion. 16-
;I.think perhaps one other thing to be kept in mind 17~-
is, let's wait and see what difference :it makes, if any, 18 to rewrit'e the. program.
~
-before.we decide In some cases it 19 may make a big difference, in other cases it may not.
8- MRZUMWALT:
. Zumwalt,' NCSU.
21 I would like to go backEto the four control 22 : volumes here on the upper plenum. I realize you may have r 23
- some problems.here of. proprietary nature on the plenum description.
26" But it seemed to me these do need to be described
y q PO , .i 59 s
', . 1 -
;1; -inltermssof the thermal,and hydraulic,.and also material
- 2 jdepositione or- absorption, or chemical interaction point of is: ;yiew,1
. 4 'I was wondering'how you feel about that. Do you
' ~
5J , feel that'is adequate 1or is.~this going to be covered later, N- too?:
' ~
v7 MR; CYBULSKIS: Well, let me)just-comment that the s? : structures.that=we use in'the z upper plenum,.the mass-18- u is'rfaceiarea,1 hydraulic characteristics, have been provided
~
10 - ' torus by,.in-the; case of the'Sequoyah reactor, by the
- 11 reactorjmanufacturer, Westinghouse.; We have'been. told that
[ U we are not at liberty [to; discuss theIdetails as being
~
13L ' proprietary. .As;far.as we'know, we have.not misrepresented
.14 'anything;.that they have provided.tonus1on the one hand; 1
;M
- on _ the other hand, they have not provided to aus ' detailed 16 drawingsLof th'e structure so that.we.could. independently '
- 17. . -review-or 'independantly.model'.
'M~ They ave provided.us with, shall1I_say, an 8: - interpreted picture-of.what is up there.' We have used that
'#' ' input in these. calculations. -
21' MR..ZUMWALT: ,Thank you.. (22 - . MR.~ _l WALKER: .'So, with respect'to the question you lE- asked,-_let,me'just'trylone. :The two curves he has there,
-r 24 there were'quite high. temperatures at'_the core exit, very 3 '
. E. )lowftemperatures.at theffinal outlet. The: structure s
60 temperature -did not ' rise' very much. 1
-2 The initial Surry calculations they did had very 3 lowisteam flow,.-just a'little bit from radiation boil-over, 4 this. boil-off.from the lower plenum.- So, the temperatures
~
5 'did not raise very much. 6 This seems to indicate that MARCH 2 calculations 7, ;are doingLthe same thing. There-is basically no steam flow. 8 MR. CYBULSKI: There is relatively low' steam flow 9~ ~up to the point of' core slumping.and then, when the core 10 . slumps.there is a lot of steam flow. 11 : MR. RITZMAN: I think we should point out that
- those-temperatures there are probably core exit average, 13 average core exit temperatures. .They are not temperatures
~ ~
14 at the hot. regions. 15 MR. CYBULSKIS: I _ thought:I made that point.
' 16 l MR. RITZMAN: It is an average.
17 MR. CYBULSKIS:- It.is an' average, you are correct, 18 Bob. It.is an average across the core. 19 MR. RITZMAN: Towards the end of this -- my feeling 1 3' is theyJare on the low side, too. But I wonder what the fuel 21 - . temperatures'are at the top of the core at that time, at 22 those times.t .I don't recall us getting fuel temperatures N' that'. low.
- 24 MR. CYBULSKIS:- Unfortunately, I don't believe 3- I have -- wait'a minute, I do have some plots of fuel
l e
'61
~
,1- ; temperatures.
- ^ A. ..
2: . MR. RITZMAN: I might.say,.our calculations with 3' .MARCHJ2, we don't.use the model'in MARCH that calculates'the
,l' 4z sgas 5 temperature at -the top .of- the core because the version is. -that wethave, when-the bottom node-in'a region becomes 6: ' molten,-then the.; code sto'ps: flow through that region and it
~
.7' uses-a saturation gas temperature aus the temperature.
s; So, we _use the 'calcul'ated.' rod temperatures. at'
~
.9- the-top of_the pore 1an'd'assune the. gas is within-50 degrees 1 10 -or something.like.that.. We-do get higher temperatures towards 11 :the.end.
u? .MR. CYBULSKIS: JThisLis a' plot of the core
, . , 'u Etemperatures, selected l nodes. .I believe there is some. mis-
.a
' 14 : labeling on the last. node.
- 15 : These are all nodes in the central region of the
.16 . core,fdesignaged by'one. 'I believe this, the last item, is
- 17 supposed to be the-top nodefin the core. But most of the-18 . _ nodes--in this central region are molten, and when'the
.19 _ . temperature drops is the-core slumping in:the bottom-head.
^
an So, what it is saying is that the very- topmost 21 nodelis still relatively cold,;but the rest of the region is n- ' molten'at.the time of.coreLslumping. !- MR. SILBERBERG:. Sol,.I.had a question. If the L u - gasEtemperatures were indeed higher,'= based.on the description
- m you
- provided-in~the core, how might.that manifest itself in -
t l
L i , t 62 15 l1; theiso-called melt process ~and,Lif'you will, subsequent 2 falling:of molten material to;the-bottom of_'the vessel?
'3: In other words, if thatris the' case, then what 4- -happens.in the other case?
~ 5: MR.-LEVY: . lI'think its impact will'be on fuel 6
temperaturesJwhich is'what,. Bob? And then finally in the
'7 releases from-there.
8 MR..SILBERBERG: You mean you expect the tempera-9- .tures to be low,.the fuel ~ temperatures would be lower? 10F -MR. CYBULSKIS:'JIn this particular case the temperat'ureJof the' fuel will-be lower. 11
-u MR. SILBERBERG: Lower.,
13 : MR. CYBULSKIS: 'I-am about two hours behind 14 . schedule, should.I --- 15 MR. SILBERBERG:_ No, we are not'so far behind
-16 -schedule.that we can't~ enjoy the coffee that has been 17 provided for us.
18 Are you.done on your part?. 19 -MR. CYBULSKIS: No, I am not done. That is why 20 I was wondering as to whether I should try to wrap it up, 21 or should we.take a break, or what. 22 MR. SILBERMAN: Let's take about a ten-minute 23- -' coffee break. 24 f(Whereupon, at 10:30 a.m. a ten-minute recess 25 was taken.)
}
e + 63 p ./ 1- MR 'SILBERBERG: .go ahead.
%<Q',.'
~
25 -MR. CYBULSKIS: -Jim Gieseke has a suggestion here 3- : that1we, want to throw out to. the : forum. 4' ' At the January meeting, which I do not believe I 5 attended,;there'were1a' lot of comments, I understand, that
- 6
- the L temperatures ~ in the upper. plenum were too high.
We heard
--7: on a number of' temperatures today that they were too low.
~ 8
- What Dr.'Gieseke would. propose, that we split the 8 difference.
_ '10 1 (L'aughter) . ;
- 11
'MR.'CYBULSKIS: -Now that we got that problem out
- 12
. of the way, .there were a couple'of' questions asked of me
[- 18
- during the' break.that perhaps I should address before I go on.
14 Ed Fuller, IDCOR, asked about the-hydrogen 15
. generation in-vessel / -In,particular,-he was_ concerned about 16 :
. how much hydrogen we' generate-during the core slumping 17 -process.
18 ' I.do not have the' numbers at my fingertip as to 19
'how much hydrogen.or how much of the cladding is reacted
#~ . while the core is in place.and'what the actual numbers are 21'
~ while the core-is slumping. Bu* in a-typical set of 22 :
calculations we do .get significant cladding reaction as Lie 23 core slups. 24 The reaction takes place primarily in the nodes 25 In other words, you will have that.have not' slumped yet.
==- - --
- y. -
~
' , _ ;" ";=_ ,. ,
y .
, ,nf '
r
- x:
[g ' J ,' D' 64
- -; " ~
< regions; of fthe : core which ;are.. relatively high temperature
?M '
,m * .
IN -
^ $ 2I liillt:the reaction hasinotLprogressed very far because the J'4
~
- 36 ; nodes:Lare1 steam starved.
, , p :4! 4ASTthe coreistarts.to slump into the bottom head
~
15: you generate thisjflow of steam which:tends-to fuel the c :e [ Ideta11ottery action..
~
-58 So,:there isi-appreciable amount of
~
- $7 icladdingfreactionipredicted'toward'the core' slumping process.
L8' -
.Weldo attempt'to~ calculate the extent of fraction l9^ that? takes place in,the lower head, but' that is typically
- i lo small unless youichoose to go' to very, .. very small particle
* '11? ' sizes for the debris in thenbottom head. -Most of that
; 12 = . increment of' cladding reaction during the. core slumping
'la i ; processes Lis with: the nodes that?have not slumped yet.
14 Ifdon't know whether that sheds some light on the 151 l question.;that you. raised,'Ed. l 18 - 'MR.' FULLER: Yes,-it does.- But I w'ould like to
'17 ;followLit upione.smalll step.further.
~
18 L iAre you implying that a lot of material-slumps
' 19 ' priorito significant: involvement 1of a large fraction _of the 20 ' core?1 - Are we! talking about something .in the order,of 50 21 ! ' percent. slumping.when the core. collapses, or-is'it much less
.- 22 - ithan:that?
~
- 23 : MRO CYBULSKIS:'"The exact amount that' slumps
- 24 'is Lobviously a functionof the core nodalization that you 26 : ; choose;to use., wechave: typically; ten regions in the core.
, s N r j.
65 1: .There can be equal regions or unequal regions at the choice 2 ~ of the: user. 3' .Butfl et's take the: case, if you chose to make 4: - the - core ' aiten -equal region representation , then the central
~5. portion;which-is typically the one that goes first would
.6 represent ten percent'of the core. Sometimes it may be
, 17- less than that,' depending on-how you choose to model it.
~
8 So that when'the.first region slumps in the-example 9 (that I cited ten percent of the. core falls down at that time
'10 step,190 percent of thefcore is still-standing. Now,-when
' 11 -
the first-ten percent of the core falls down, creates steam
^ 12 - which-feeds the metal lottery action, which may lead to
'13 1 fairly rapid heat-up melting-of-the. successive regions.
14 - So, frequently once melting starts it tends to 15 - progress, or once slumping starts,'I am sorry, it can is progress very' rapidly to the entire core falling into the 17 bottom head. Now, again to shed a little further insight into 19 .the question you raised, at the' time of' initial core 20 - . slumping-you-will have in the neighborhood of 20 to 30 21 percent core melting, and it will vary sequence to 3- -sequence, but'that type of neighborhood if I remember
# looking over the numbers, over some cases.
M MR.lRITZMAN: I am sorry, Pete, I want to inter-2 rupt with one question again.
67
'1. ~and I th' ink we kind of'--
2 MR.-CYBULSKIS: . We slipped here and, as I say, I 3- must' apologize., that ~is something that adds some needless
- 4. confusion, obviously.
5 Let me go.on to the containment response for the
.6- ^TMLB' prime sequence -- and~if I don't hurry up we will be 7 talking about-TMLB still this afternoon.
'8 ' Basically, this'is a TMLB gamma sequence where
~
9 we said that the hydrogen is allowed to burn if flammable 10 conditions exist. We did~notispecify-what-the ignition
; 11' source'is, the just said, if'there is an ignition source and 12 ~ 'the conditions are favorable to burning, let it burn. This 13 is really a very simple - example .of that. It.just so happens 14 that'in the calculation flammable conditions are not 15 ' -reached until immediately after head failure.
16 There is a lot of hydrogen, the burning takes 17 place in the upper compartment and you get a big spike which 18 fails the containment.. In the particular calculation illu-4 19 ' 'strated the debris-bed was not coolable, so you continued
. 20 some hydrog*en generation, and there are some successive 21 spikes but they are.obviously not as significant as that.one.
22 .Now, if there are no ignition sources -- I was
, 23 going to say the other extreme, it is not necessarily an 24 - extreme. For the particular sequence as it is defined it 25 may be a perfectly' plausible scenario. You get a totally
n . .-. - - -
~
I 68 1: :different picture in~the' containment pressure response. 2 fAgain, it is the same sequence but low ignition. 3 Remember,?the ignition took. place at this point.in the other
- 4. .' sequence.. What happens is, again the debris is uncoolable
~
5 'or if it.is.coolable in this sequence it will eventually run 6- outrof water and you get a slow progression in the containment
, 7 press ure . -
8 Inithis particular case I had used 60 psi for a
- 9 ' failure pressure and you get ,there in 500-some . minutes .
1(L obviously, the. difference in the consequences -- I guess we will here the results in a little bit -- could be quite 12 significant between the two. 13 ' MR. HILLIARD: Is that dependent on the type of 14 concrete for. some -other reactor with other concrete? 15 MR. CYBULSKIS: That would be--dependent on the
~
16- . type'of concrete. Obviously, if.you have a significant
' 17 fraction of limestone you would tend to pressurize a little 18 ' faster. .If.you had basalted concrete you attack the concrete.
19 I think that the rate of concrete attack is faster than L# basalt is but you generate less . gases. lli So, it is-conceivable.that you could get melt-JN through before over-pressure failure. That is one area that 2 is,;again,' highly uncertain as to exactly how the concrete 24 attack'would progress. 25 Let me go on to the next sequence which was the
+;
~ - - - ~
* , - en _ ,
l [ ~ m _; ' y, f. *~ r s a-.
~
69'
~
I' J 1' - ~ TML sequence. Remember, thelprimary system here, the primary jsystem behavior is' identical to.what we had in the TMLB
'- LIS ' csequence but the containment safeguards are available.
.4- lIf I'may, let: me go back quickly to my very
' ~
1 LS 1simplelevent tree _ notation and indicate-some of the possi-
]
t6 k ibilities that oneimight.want:to consider if you talk about
# 17' ithesTML sequence, j y -
i8 ~ Now,rI'said the1 containment safeguards are'available .
-9: =That means the sprays are on, the. air circulation fans are E
10 So on ' the :the hydrogen igniters t are on, by definition. t il it hatLone,ino hydrogen burning' cases' disappear. Presumably ut ._the hydrogen would always' burn-in these cases.
~
D Again, the sameLargument'that I used previously,
-t
- 14 ' the; core melting: takes-place where there is a lot of ice in l 2- the. ice condenser so:thatL the steam production from vessel
[ k . is - :intrerac'; ions:probablyiwill*not. fail the containment. i F 17 The cases thatiyou are left with,..or perhaps the p '18 ; morellikely ones I have tried to indicate and the point I ; i
'W' 'tried to make here,-there are'several what I have called
- 18: ' "no l fail" cases.-_ 'If the igniters work as they are designed !
1 11 - and the burning' takes. place in'11ttle chunks or continuously, I s
'E- and.~in?this case'where the. spray is operating,_the ice - 1
~
se i . melting, Tyou have _a' lot _.of water. in the cavity, it is , 1 88 conceivable ifLnot highly likely th'at you.may form a debris j bed and inifact you may.come out with one of these scenarios
~
~8 j
~ u + A
+ <-w-- * * '
-F* - * '-frei n -WW = -
's-*T--s -
***---'e+-"-*--'*
A i' $s[ M [ ' '. %%W .,
, w ~ ~ '
'~ gg
,e. a: -
-g Th.e temperature curve you were showing before the 7' .
; break on upper. plenum temperaturesJ--
_3'.
- 3' a '.MR.-CYBULSKIS:. Yes, sir.
-s
~
- y. *Q -
' MR.- IRITZMAN: .- Dit starts at 45 minutes. That is
'31 l45' minutes what?;
e "
*" ~
a MR'ECYBULSKIS:
. -I'L have to refresh my: memory here.
-.7i iThose' comments I:made before about timing on these graphs
^ -
;s- h
.have not been-clear. ~I mustHapologize if.there is confusion
,of ;on-that; to MR. RITZHAN: Whe' nj we lookx at.'the events table 'it.
7
. says zthe . melting" starts at '121'.5' minutes. ' Is that time after-
~
~ 11 -
13 start of'. l melting?._ c .
. 13 ,
-MR.cCYBULSKIS: .In'the-event table the times are r-1 14 measured--from time:zero,"from;the start of;the accident.
13 : ThisI-is something(that unfortunately fell through:the crack, 13 -noti. intentionally.
'17 : In the primary system temperatures, I believe those
- 1s 'are measured from the time of core uncovery. I .must '
13 . apologize'for that, that is'something that should have been
.m' made clear or corrected. .
' 21 ' ,The-people that do. fission products'do not recognize
-ss .that there.is'a world before core melting starts. So, the m ' clock is-Erecalibrated.
se MR.'SILBERBERG: As I reca'll, Pete, we had corrected
.- ss : some of those.- There were difficulties in the BWR reports y ,
.y
p _ r 70 1 owhere the' hydrogen-burns but the debris is cooled. There 2 is no concrete attack and you' terminate t$.e accident. 3 LIf.the debris does not cool in the. cavity for 4 . whatever reason,_you-may have-an over-pressure failure s5 ultimately if it keeps attacking the concrete, similar to the
- 16. other one. You keep' generating non-condensibles-and if you 7l keep doing that. long enough, the containment may fail.
8- -Or the other casefthat I have talked about, if the
'9 igniters do not work' exactly like you have them because of 10 whatever' differences in hydrogen production rates from what
- 11. . you anticipated or if a burning happens to take place in the 12 ~
wrong place at the wrong time, it is possible that you would 13 still fail containment,'so there is the-gamma or the hydrogen 14 burning case. 05 There'areLseveral variations of these cases 18 possible, obviously. We have chosen to look at one particular 17 over-pressure case and one particular hydrogen burning failure is case. There are, of course, a number of possibilities for 19 - containment isolation failure cases that would exist for this 20 sequence.
- 21.
This particular containment pressure temperature
' 22 response is identical to what I presented at the last meeting.
23- Again, all I am saying, the particular calculation was done 24 for a two-volume modeling of the ice condenser containment, Mi assuming ignition at a. volume percent hydrogen. In this
- = -,
72 41 3 ^ 7
-: started out modeling the containment as a two-volume system.
2 We ' looked at some variations of -it. We modeled it as a four-volume system. We looked ~at the igniter threshold, six and 4' eight-volume percent ignition and'in all the cases considered 5-weigot what I would consider appreciable pressurizers.-- 6-
~
Jappreciable being several to many times the design pressure 7- of . the containment.
-8
'Since the.last meeting, and since some-of the 9
comments that were raised by the reviewers.as'well.as the 10 - observers 1at the last meeting, I have done some additional Il calculations where I tried'to explicitly model the upper , U plenum of the ice condenser and put the igniters up there. 13 Unfortunately I' lost the slides for-it. Let me 14 see if I can find them. 15 In that particular_ case-I' subdivided the containment 16
~into vour volumes with the upper or lower containment being 17 divided into the main lower volume and the dead-end volume, is and the upper compartment being divided into the upper plenum 19 of the ice condenser and the dome and the rest of the upper ;
20 plenum. 21 . What I am going to show you are two particular 22 sets of calculations of pressure response. This is four volumes El TML. sequence with the upper plenum of the ice condenser 24 modeled explicitly, assuming ignition at eight volume percent Ei hydrogen. If'I remember correctly, this comes out to a peak i Y- _..L-
- -CC A- M, , 73 1 .' pres'sure'of'"about:46 pounds per-square inch.
/ , it-If'I; repeat _thismsame calculation and same modeling 3- ' assumptions,1or'same. compartment modeling assumptions but
}
n4 'I use a"six-volume' percent ignition threshold'for the f t
'5. igniters, we get this. kind of' response, pressure at-just-6 '. 1about-40.? psi or something like'that.
_ :7- The1 reason why the pressures are as high as they
!8 - :are is'~that you'do'--:I. don't have the detail that shows it --
9 - you-do t'e nd to get some propagation of the burning ~in these
- los Lealculations.from the upper: plenum of.the ice condenser into
-11'. .the_ main volume. 'It makes the hydrogen burn pressure larger,
' ~
12 ; yo.u burn more hydrogen.
.13 If you can assure yourself-that allithe burning
= 14 takes place in the ' upper. plenum of the ice condenser, you is ^ would-have.relatively small pressure rises because this is a-large volume and it'has a' lot of expansion into the
' Is 17 other volume. But if you look at.the possibility of HB -
propagation'into the other volumes, you can still get N' _ appreciable pressure rises. 3D - -MR.. CASTLEMAN: I Pete, just a quick question.
< -211 Cast!1eman here at the table. Just in order to'know, are 23 .these all adiabatic burn calculations, or do you take --
- El HKR'. CYBULSKIS: No,'they are'not adiabatic burn 36 calculations,Jthey take into account heat losses to structures. ;
C lui .they take into account the~ effects of the sprays in those
;- p gn+ , - >
N "T, ~
, , _ p ~,O' ,
}Q 'r , .?. , _ ~ 5, ~
- I Q%,
~
y y , 7 , , 74
, s, o i . , . . . ; . ,
7 , 117 c;compartmentsfwhere they?are operating.. The one thing that
'.,.]' .
,x ~
, s . .
, + .
- 2. ,i
- theyEdo'.:notitake'into accountithat is-included in some other Li f ' L3 ;
i: codes is' radiation from theiflame~to the structures. b_ - 4; s ;I do;not'believe that.would'have'.a significant Q. ,' ' '
, , , (5 ^ 'offectp on' .the } peak pressures . It would have an effect on I8' 'howfast(itcooled:down..
? ; .c - O<'
- L7 Questions alliover the. place.
o !1 _ 18- }MR.-7HILLIARD ' :Dottheyfassume a hundred-percent [ [8- [ burning fficiency, f say, Tfor the six -percent as well?: 10 . - _MR.?CYBULSKIS:. No. . ^In the. code as we use it,
- 11) ~we'have. adopted or)Sandia National Laboratories has provided M 112 lusLwith their burning models to:use in the MARCH. code which 13 '
~
i I1 understand7 to:be=.very similarLto what is in the. HECTOR 4 1 g 14 -? Code . . 7: s [ b 18 There are~some built-in controls that are based on experimental observations, analyses,-what have you.
- 18. - .
If
- 17 :: ignition takes place 1at eightLpercent or.above, you get
(, . complete burning'of1the hydrogen.
~
- 18 If -ignition takes place 18 below eight percent,1there is:a correlation which I believe so - _'Lis ' based on experi:nents that only ' leads ; to partial burning.
21 i In;the case'of the:six-percent ignition there is l 2B e not complate' burning.- If-you go.down to the limit of, I 88 ithink Litiis 4.01^ percent _as it is ' programmed in the code, -
- i. . w--
84 Jfor:anfignition threshold; I think you would only burn the L>> '88' .01? percent,and,get many,=many little-burns. S 4 5 ' # q , &
3 71 1 particular case we predict that the containment would fail or 2 might fail and proceed from there. 3 I will come back to the question of igniter 4 performance in a minute. 5-If the igniters do their thing and burn off the 6 hydrogen before it accumulates to an appreciabeloctent and 7 you will not fail the containment due to the burns, you may 8 get a succession of burns and.the magnitude and the timing 8 of those burns will obviously be a function of your hydrogen 10 production rate and the assumptions you make about the 11 igniters. 12 In this particular case I assumed, again for the 13 purposes of the calculation, that the debris bed is not 14 coolable. You keep producting hydrogen, noncondensibles, and 15 eventually you will fail the containment. I did not carry 16 out the calculation long enough but you can see that it is 17 going to take quite a while. 18 The alternate scenario is that you survive those. 19 burns. You obviously probably survive these without any 20 question. If you form the coolable bed you will cut this 21 off and the pressure will probably level out somewhere close 22 to atmospheric with the safeguards function. That would 23 represent the no-fail case. 24 On the question of hydrogen igniter performance, as 25 I discussed the last time we met, we had looked at -- we
'l- 'R.' CASTLEMAN:
M For the-case where the sprays fare on,ldo'you takeLyour product as~ water vapor or do you J2-
.3- have the water condensing in theLpeak pressure calculation?
4 MR. CYBULSKIS:' The products of the hydrogen 5 : combustion as well as the energy generated, a mass in 6 . energy balance and you calculate the resulting effects. 7 -The results of the combustion would be water vapor, s- .But'in the. overall energy balance it would be distributed to 18- . what'ever -liquid water is 'in'the atmosphere in' that particular 10 . compartment at.the time, as well-as all the other gases, 11: whatever is there. U- MR. CASTLEMAN: I amojust thinking of the peak 13 pressure. It is going to make a big difference how long 14 it takes to. condense whether you assume always instantaneous 2 vapor-liquid equilibrium, or whether --
> 16 - MR. CYBULSKIS: lit is not instantaneous equilibrium, 17 The assumptions that you make about heat losses generally UL will not make an awful lot of. difference on the peak 118 temperatures because1the burning rates are- generally fast
'EL -enough that-the heat transfer processes don't make'much 21 difference.
23 .The gases thattare airborne, if I can use that 8' eypression, will mix instantaneously, if you will. But se anything that is not airborne will'not have much effect on 25 the peak,fas'I' understand it. g , b . _ _ . _
, . ~ . - - - - . N ge ,
~ '
' =
l
% g
- 1~ 4 77 .l q - i
- ,g - ,
'515 -withhlimestone concrete, you probably. would have taken a l 'k..; -
t .
;2^
" lot? longer toiachieve the pressure levels that you indicate !
, - .3 thNre..
, y -
'4 (MR.nCYBULSKIS:
I guess -I don't know whether I 8 ishoul'd; comment"one way or another. . 48' ijp - iMR.L.GREENE: ;You think~it'should be done, though? 7-
-It-could-increase theLtime to: reach peak # pressures that you 6
. ~s ; :
show - ati-;thef en'd - of your' slides by 2 hours .
. 8/ IMR'. CYBULSKIS':1.Well, I am not sure how significant - - >
the point.may be-totally correct. I am'not sure how i : ll : i significant it.'is.1 Rememberj att least in'the. context of
' 18 '
lthis sequence'.you'have the sprays operating.-:We are talking-18
'about;.800 minutes or'something'like'that out-here.
14'
'Whether the' containment fails'ata1,000 minutes or 2',000 minutes may not really'-- atileast from the context of 14 .'
overall results. -- may not really be JaL point of concern. - 17 MR.'GREENE:: . One-other-little thing, since you put
~
E' that:up there.. Can you explain what that :non-linear hump is I# out past all your hydrogen burns?-
'8
'MR. CYBULSKIS: Here?
-L 21 J
.MR. GREENE:- On the.way.up there. . e t
IE - MR.ECYBULSKIS: This one? E MR. GREENE: Yes.
~ 8' '
MR.-CYBULSKIS:' I am not sure .I can -answer the
'8 o
question', I!would .have!t o look at .the calcuations is more [
- m .'
4 i^ y
" .- ' ;f '- i 4 -
9 i
. )_ l _ _..-L--------'
- n. _
-~ -
, -m. .r"
* ~TJx'
=
~
F {
- T D.R ' -%g? 37
?
m ' r,
,e
.Ds
- 4 t,v .p_
. ;4 y =:
1 + ',i 1
~ T !"i
- [# h~
a: 4 .76 '[ 7
-i
;4 gi e -
i? y,
'
- o
_;p_, . _ g_
+ .
.g- -q. 7 s&
M'.u.g. 'MR.:. GREENE: ~ : Geor},ge - Greene from Brookhaven.
,,' j1j s -
$g J 4
- i-un. s
, 1% .-
. m . X
- g fy 2.. -A. ' Pete , J on:' your :le ft. islide - .. withathe periodic burn
'l; v7, z
,A g W _ g; g]q!
i
~
, .; 3 o y,
, w si;7a-continuous base-line;iterease in-the containment pressurel ,
t :' '
,,. z.
o : c - g, - --4 u Now);for clakification,gJise this due to a debris: interaction of .N: s
&:m
;s;
$this concrete .or molten core concrete ' interaction?
s La _ _( % .'
- 6-
.y, -.MRL CYBULSKIS: 1 In,that particular calculation that '!
27-
,s ,
hshEbredit}ismoltencoreEconcreteinteractionthatsis .} s "r ,
- - ; 3 ,
1 generating.. hydrogen.as welixas other non-condensibles. .
& 1,8 l Thatf. ;
k - lis whyfthere is'a continuousSincrease.- N 8e t d [ to
. . fid\
If .you 'did not have concrete tattack, you would not
. a .
' l11) 1 fsee that continuing 1 increase.
y
%, i > W
- 12 . ' JMR. SILBERBERG: } George, use-the mike,!please.
\
' I
- 13 . MR.'GREENE ? Is that'. calculation performed with 114 >
.the enter y code 74 [, " i s - -
'q *
- t
, [ .,
- '18 "
- MR.ECYBULSKIS: s.That is correct. '
}-(
.16 t fjhR.GREENE: ?And have yea'done maybe one check to 4
.c 'n '
U 17I ? find out.if"the base-line pressure-slope would beLattenuated: 18 'if'you used;.the CORCON~ Code?- .,o
. , , y
~
18 MR.~C ULSKIS: Ihave'notpersbNa1 experience - 4 ( 5 \!
,M-b, 30 i with the;CORCON Code. I thinkLtihis question has ~ been raised s ',
I wonderiif anybody from Sbndia would
~
> 21 ) frequently enough.
as? . care to make a comment'about the relative gas generatio N
~
t* a y.
- 38 ' irate.1 T
'f
., , e -
y j' .. y i g e J S4 ' MR. GREENE . . I tihi.tk I lcaMtart it and somebody. (re- a'
~c .
. cs.
2B /canit' ell me I am wrong. If you had' used the CORCON Code i j, m _. ,
- 4 .\ .
7 , g Y,', I f
't ,.
, b. > *
, y i__
s , . . , .. 'I' ( ,, ,
, 1 .
t , e
78
'1 detail thAn I have.- I suspect it has something to do with t
L :2 the ~ debris inversion in the concrete interaction, or some
'3. such thin' g -like that,' the debris layers flipping.
~
But I do not know for a fact.
~
.- 4 I would have to
- 5. -
~1ook at-it'in more detail.-
16 ' Bill Kastenberg-has a question? 7- , MR. KASTENBERG: Two of them, actually. 8-
'First, it-is not clear to me what is more importantI 9
=the change in the percentifor-ignition ~from six to eight ICL percent or the way.that you compartmentalized for the 111
-different calculations,.and'how~you follow that through 12 Ito your end point.
13
.The second-is, if the compartmentalization is 14
'important in how you'do the'modeling, should you not do the 15 -
same for' the' TMLB prime sequence or other sequences for the
- 16 ice condenser?
17,.
.MR. CYBULSKIS: Again,'I-don't went to answer the
- 18 '
question totally by saying that we can look at only so many 19 things in'the context of this-thing.
~
20
~ C1early,- the compartmentalization is important
- 21 because ;the way we model and .the way most- people model 22 Lhydrogen; burning in the degraded core business is, you look 15 ~ at the entire' compartment.
_M ~Now, if you carry the argument.to some extreme, 26 you'can say, "Well', if I make my compartment small enough and
n-- ,, - 79 1 . only burn in a limited number of them, then I will get no 2 pressure increase." 3- One thing that I personally try to look at when I 4 talkabout,compahtmentalizationis,whatarethephysical f
~8 restraints.that kill create compartmentaliza' tion, and at a some point you come to the conclusio '. hat there is no 7 point in ' dividing the volume into more nodes , it is just a mathematics and may and may not have any representation.
8 But_it is an important aspect. In the calculations to thatLI have shown,.what I tried to convey is, I tried 11 several different nodalizations and in each case I got what 12 ;I considered a significant pressure increase. Obviously, I 13 did-not exhaust them. 14 I tried perhaps a nominal ignition threshold is and.what might be perhaps a more realistic threshold and 16 ' still, the pressures did not go away. .But they are both
. ' 17 ~ important.
18 However, in terms of how far do you go in modeling, 18 - I think it brings up the point that Dr. Levy made earlier, 20 ' how far do you go in the finer. detail before it quits making 21 sense within the approximations in here in.the analyses. 22 MR. CASTLEMAN: 'I guess my concern is that you 23 lhave chosen this failure of 60 psi for these examples. By se - doing this.nodalization or changing the threshold you 25 . straddle on your. cases for'some reason that threshold. So, b
80 t it is important. 2 MR. CYBULSKIS: Well, let me interject. I have 3' chosen to.use.60 psi as'a failure pressure for my calculations . 4 That is not necessarily a threshold function. Again, there 5 is a working group that is trying to address that question a as to how much a containment can take. i 7 But let me just point out that if you look at the 8 Zion containment design, it has a nominal design, I think,
-s of 11.8 pounds gauge.
i. 10 MR, GIESEKE: I think you mean Sequoyah. ( 11 MR. CYBULSKIS: I am sorry, I mis-said it, for
- 12 Sequoyah. The only thing I am saying is that in all these 13 . burning cases I am getting several times the design presssure, i 14 therefore to me for my perspective it seems prudent to l 15 consider-failure.
,1s I am not.saying that Sequoyah will fail at 60 and 17 .not fail at 50 at this point. That question remains.open for is the time being. But to do the calculations,-I have to have 19 - an input number.
20 I do not.know whether I answered the question or
'21 evaded it,. Bill.
22 MR.-KASTENBERG: I am not sure either. 23 I' guess the answer will come when you show us 24 the bottom line,when you show us your bottom-line calculations, mf MR. CYLBULSKIS: I think there is a necessary
'82 1 ' higher, a few psi. Again, it depends on the range you are
~ ' '
~2 in. . They are not dramatically different from the actual 3 peaks.
A 4 MR. RAHN: So, you are essentially saying that 5 the adiabatic assumption does not make much difference. 6 MR. CYBULSKIS: It does not make a tremendous 7 amount of differencein most of these cases as far as the peak
.a- pressures are concerned, in my experience.
9 MR. REYNOLD3 : Reynolds, Virginia. to I had a question about the peaks in the two cases, 11 TML gamma and TML delta. Why are they different? The peak 12 in the.TML gamma is about 66 or 67, and in TML delta it is about 52 or 53. 13 -I would have thought the'TML delta would have 14 the same initial peak and just differences after that. 15 - MR. CYBULSKIS : I am sorry, I neglected to perhaps
.make the point t1at in the PLM gamma that I presented and
~
16 17 'I: assumed containment failure, I used the eight percent 18 ignition threshold. If I had used the same threshold in 19 ~t he_other-one, I would get the same results. 2 Be'cause I was'looking at a case where the containmen ; 21 'did not' fail, I chose to-lower the threshold to six, which 22 .gave me a somewhat lower pressure. I should.have made that 23 . point- clear.' 24' MR. WALKER: With respect to the Kastenberg question , s just let me _ put a couple of things -- so you are aware of them --
p=~ -
~.
e t a P '- L:' . 97 I' perspective"that is not available today as to what is a
=2" meaningful failure pressure for this reactor or any other L :3- reactor. I think that 1:s a very key question. An awful lot 4 of the Tthings that' we .say will hinge on that. That perspective
.5) we do-not have'available to-us today.
6 .MR. RAHN:- Frank'Rahn,'EPRI. c
=7 'The pressures in the hydrogen peak seem to be L -
8 ' fairl'y h'igh. -You mentioned f that you did not use the l
~
l 9 adiabatic-assumption doing the calculations. r
'10 -Did'you do a calculation that did have the
- 11 J adiabatic. assumption, and if you did, what was the difference?
- c 12' MR.1CYBULSKIS
- . Typically, the adiabatic -- let l
[ 13 me' rephrase that. h 14 In the MARCH. Code we have the option of not burning
^ 15 ~ .the' hydrogen but~asking the-question, what would the. pressure 16 .beDif.the-. hydrogen burns.
= 17 So, then the code'-goes on and calculates the l-l 18 ._ adiabatic pressure (at any point in-time.- But since you said p - 19 . not to; burn _it,'it does not' throw itLin the energy balance sol :and you-can print it.'out'and look at it as a point of interest,
- . 21 - and .then Leome back on th'e next run'-and say, "Well' based.on ,
I. 22 - - the previous run fit looks':like ' burn may do ~ something, so .I i I .' . . . . , - .
, 23 willsigniteTittthe next time."
24 Typically, the adiabatic b' urn pressures are somewhat L , L El - Lhigher.than:what youfactually calculate,_but not dramatically
~
i
+
'e
} 41
-.-A - .a - g.-
83 1 on the record. 2 "The ice' condenser. containments, there were
- 3. calculations ~done and put on the'ocket d for both Sequoyah 4- and for :the FMP on ultimate capacity of those containments 5' around.60 psi.
6 Also, Bill, there were a number of compartmentalized 7 burn calculations done on the Sequoyah docket with the 8 Classics Code, and they indicated the same effect that Pete
.'9 ~ has noted here.
10 MR. KASTENBERG: .Just to clarify a point and not 11 ;to belabor the-issue. On your table which shoes the two
.12 ~ cases you did consider, the TML gamma and'the TML delta 13 ~ epsilon, on the TML~ gamma you,show a containment failure 14 at 157 minutes, and on the TML' delta epsilon you show 15 no failure.
16 Is that the way you proceeded with~the calculation, 17 then? lll EMR.-- CYBULSKIS : In the TML delta. epsilon I did i III -not carry out'the' calculation far enough to get to-either 20 a natural melt-through or'an over-pressure failure. I did
- 21 .not think it' worthwhile to worry about it under the
.2 . circumstances.- >
-2 What I'am implying there, I guess, is that if you
- 24 ' 'let this thing proceed you will get either over-pressure .
- 25 Dor melt-through eventually. I was not overly concerned 4
- p T w P
84 Again, I am perhaps
~
1 as to which would happen first.
- 2. showing some prejudice but in this particular case you have 3' a containment with'the safeguards functioning for that 4 length of time, if there is anything left airborne other 5 than-the noble gases, I would be very much surprised.
6 MR. GIESEKE: Mel, I have a question. This is 7 dragging on pretty long. How do you want to handle it, there 8 is a lot of discussion on the thermal hydraulics. Do you 9 want to keep going on and extend this to go on Friday, or 10 do you.want to terminate some of the discussion and go on 11 -to other items?' 12 (Laughter) l 13 - MR. GIESEKE: I guess the main point is, if people 14 have gotten their-major points across on the generalities
~
15 - rather than details sequence by sequence, do we'want to 16 continue going-through? I:think you~should decide because 17 . we.are already what, an hour behind?
'18 MR. SILBERBERG: How much more material do you
'19 have, Pete?.
20 - EMR.'CYBULSKIS: As far as what I had to present 21 on Sequoyah, I have.one, the S2HF sequence, and then i 22 depending on the amount of detail that people want to hear.- E MR. SILBERBERG: It.seems to me that based on 24 the' time we have' allowed-for discussion and comment, we 25 would have allowed.it this morning, as well as some time in
A
' ~ ' ^' -
- 85
+ -
~
'l the afternoon. I.am not all too nervous'about the time.
- 2
~ '
But -why don't' you finish the next sequence. and hopefully many of the questions will already have been 3 '- 14 answered, asked previously. We can try to pick some of that is- up again. 6 .-I'should say,,it was mentioned before, with all 3: d'ue respect'to the. question of containment failure analyses, si that the:Containnent Behavior Group 1will treat in detail as 9 a standard problem the Sequoyah situation, the whole 10 questionc of hydrogen L from. beginning to end, : getting a panel 11
- of experts to put their analyses and opinions on the line.
12 So, that will serve to put, if you will, the
., 13 'Sequoyahisituation in perspective.
1 14
.We alsd will go beyond the_ simple assumption that r
M- says,; "When I hit Pressure 'X L everything comes apart." The 16 - Containment Capability Group .is' looking specifically at the 17 possibility' of local:-failures,fmore likely_ failures of seals,
* ' 18 be it -hatches or a purgelvalve, . vent valves, cnr whatever, in 19 .
terms;of which of them are likely . events, I mean leakage 20' failures from-these kinds of_ loads.
~
21< They will'actually tak'e'into account the fact
~
'M
~
- that the; leakage mayHoccur well:before a containment 23-failure'which:in'effectimay tend.to. turn around the pressure, 24 .too.-
; 25 - __
Allsthose' things will be-put in' perspective in due
86 4
- 1. course that would ultimately provide a basis for refined 2 ' analyses (ni these questions sometime after December.
,I would -hope that- these issues will be addressed
'4 by the. group of people studying it.
5- -MR. BARI: Bob: Bari, Brookhaven. 6 In termsfof putting things into a. broader
~
7 perspective, I go back to the TMLB prime gamma'that Pete
-8 -showed1where he~got early containment failure, I think, at 9 -about 160 minutes or so due to a large hydrogen burn in'the 10 ' upper-compartment.
11: Here we have a compartment without any electric 02 - ypower available to it. The question is, what:is the1 ignition 13 - sour'ce for that~1arge burn at that early time without power 14 available? 15 In some of the early-other TMLB prime sequences 16 - for large drives I-think we have talked about the core debris 17 serving at least as an ignition source, perhaps not burnable 18 ' at that time.- 19 But here I do not see the ignition source. ' My 20- question is, is somebody going to be putting that part of 21 - the problem into perspective ~, where are these ignition sources M and so forth?
-n 'Other questions.thatJseem to be falling into the 1 14 - cracks -are , 'a' gain for the' TMLB prine , say .the pump seal 251 failure question, no power, :no _ cooling..to the pump seals
n, . . . .. r - , - t - t W s , 87 r s L1 1 due,to'no; power' .
,r; 2-The; scenarios change,Tna do not.get the accumulator 3~
, , -ldumpfaLheadifailure,ctne accumulators come.in earlier. ;
4- iThere~is'an_ interplay between-accumulator water and ice N^ 5' .~ melting, and so forth. _ 6. . Somewhere, . hopefully, :that is going to come into , e-7 ' perspective.
~
8- MR. SILBERBERG:L Thank you, Bob.- I.believe it 9 will, in terms of.tryingfto lookLat' things realistically. 10
.IDsuspect some of'your people will:be helping us to do it.
11
^
- (Laughter) - ,
H-101 1CYBULSKIS: 'Indont;know whether that was a
- 13 commercial-or not.
14 l(Laughter) M MR. CYBULSKIS:- Back to'my-event. tree S2HF
~
16 ~ sequence again. I' tried to illustrate, :as -I tried for the~ 17
' previous sequences, on.my more or:less general containment 18 -
Jevent tree for:this sequence what is possible, perhaps, and.
~19 ' E what is not'possible. Remember, S2'HF, the' core is initially 2
cooled and.then the recirculation systems', both"ECC.and
+
.21' While:thisLis'.not a sacred-conclusion,..the
; spray fail.
c ' Bi
.way we 1 ave analyzed it, the ice is.all.gone by.the time M'
'the. core starts to melt.
24 ;
. Given that type:of background, and we do have
' 8 -- ligniters.:in the air return fans'in the.. sequence. .So,.the:
r
,n'.'-
_ 4
7 88 I cases with no burning fall out by the wayside.
,. The cases 1 21
. with;over-pressure,. avoiding over-pressure fall by the 3
-wayside.- The reactor cavities cry, so we have calculated 4 it,' so you don't have lany steam production when the debris 5- gets - there, so those falf by the wayside.
6 Some of the principal sequences that you have 7: left~-- and there may be others that I have not really
-8 explicitly indicated but'these are the principal ones --
8
.given the sequence as it is defined is, you either fail by to
_long-term over-pressure, you fail-due to hydrogen burning, 11 or you have an isolation ' failure.
-12 we-have chosen to analyze that particular one there.
13 Again, just: to define the sequence, it is a emall-14 break LOCA with failure of.the emergency core cooling 15 recirculation and spray systems. 16-The flow paths in the primary : system, the fission 17 products and the gases, come out.the core through the hot let,
' 18 through-the: steam generator, and out. I:believe in all the 19 analyses that we did 'we Lassumed effectively a break .that would
# .;1ead'to fission products through the steam generators. You
. . 21
- could also have breaks, obviously, - that wculd not have let
- # '- everything flow through'the~ steam generators.
.2f In.the primary system treatment, again,-this is 24 the. upper plenum =-- I will give you a_ sketch of what we did-8 with.that; piping steam generator.
J L g --+-- g -s v.,
~. - .
89 1 MR. . FULLER: This is a quick question, Ed Fuller 1 2 .from the IDCOR Program. 3 Did you assume that the-ice was all melted by 4~ the time'you got melting, or.did you calculate that with
'5 -MARCH;2?~
-6 MR. CYBULSKIS: The melting of the ice was 7 : calculated as a function of input into the containment. In 8 the p' articular set of. assumptions that-we used regarding the 9 ' containment spray flow -rates and ECC flow rates which are 10 - really key to the timing of'this thing, it turned out that 11
-the last bit of ice was melted at about the time of start of 12 Lcore melting. Alternate scenarios are possible.
- 13 It was calculated, but obviously under a set of 14 assumptions.
15 MR. FULLER: Where did you obtain the mass 16 inventory of ice? 17 MR.'CYBULSKIS: From the safety analysis report 18 - for.the Zion plant -- I am sorry, Sequoyah. 19 MR. RAHN: Frank Rahn, EPRI.. 20 ~ How manyLgallons of water would that be, Peter, 21 if all that ice melted and-where would it be? 22 MR. CYBULSKIS: The number of ice that we used 15 is2.45'. times 10=to the._6th pounds of ice. I' forget what 24 the. conversion, the gallons is. I think'it.is eight one
- 26. way or.another.
J
90 1 -- -Inhthe sequence as it-is defined -- and we used
-2. Lthe -definition 1of the' sequence that was developed in RSSMAP --
3 ? the ECC works initially, the containment sprays work 4 initially ~. After.the refueling storage tank is exhausted, 15- the suction switches <to the' sump because the return line is;
.from the upper' compartment to the-sump has been left closed, 17 presumably, after a refueling incident, the containment 8 ' sprays pump the waterffrom th'e cump into the upper compartment .
'8 The water cannot return to the lower compartment.
l 10 The: sump runs dry,'the pumps fail, and then the sequence l' L 11 progresses from there. 12 ,
- MR. SILBERBERG: Three-hundred thousand gallons 13 of water.
14 MR. CYBULSKIS: ~I seem to have lost my slide, my 15 cryptic slide. 16 .MR. WALKER: Pete, can I ask you a question? It 17 :just occurred to me, I had not thought of it before. 18 l In the S2HF, what you basically do is,.you melt 18 the ice out ;and . dump the RWEST into the lower compartment. m You.do not try to go on recirc~until you have all the RWEST 21 down there .-
- 22 MR. CYBULSKIS :: That's right.
-# - 'MR.' WALKER:' Is the~Sequoyah. design such that at L iM - thatitime -you float the' cavity? ' ; Because if you- have , it will
- 25 ' ~
probably stay flooded. 4 e W I $ f +- -
i . b 91 1~ MR. CYBULSKIS: 'In our calculations we attempted 2- to' represent the relationship between the cavity and the 3 containment sump, and in a particular calculation we did not
'4 reflood the cavity.
B
'5- MR. WALKER: So, the pump cavities up on the 6- ' containment floor, if you ever flood the cavity you are l-7 never going to pump that back out again.
[ .. 8 MR. CYBULSKIS:. I am not sure I quite follow the
, '9 point that you are making.
10 - MR. WALKER: Well, when you dump the refueling ,
- 11 water storage tank in the ice, it melts up on the containment 12 floor. . There is usually a-divider that you have to over-L 13 flow to put the water down-the cavity, which is a lower 14 level.
15 Okay, the recirc pump takes suction from a sump 16 'which is up on the containment-floor. 17- MR. CYBULSKIS: That's right. c18 EMR. WALKER: So, if'you_ever overflow into the 19 cavity you will never pump;that water back out again. That f 20 -is why I asked the question. j' 21 MR.-CYBULSKIS: Yes. In these;particular t- ! 22 calculations the cavity was calculated to be dry. f
- M L I seem to have misplaced my slide. But again, b
l= 24 in the primary' system treatment I indicated the flow paths
'l'
- before. .We treated the upper plenum'as a single' volume
~
15.
.~
?
92 1- withifour: structures in'it. We had a separate volume for 2-the pipingLfrom the. reactor vessel to the steam generator, 3 andithen we had aL separate representation of the steam
' generator before We went on into the containment.
5 LGoing on to the containment response -- and I 6 believe this isfidentical to what we presented the last 7 time -- this is aLeontainment pressure response. 8; ~ At the time of core melting, again there is no 8' ice, the sprays have; stopped.- The' fans are' recirculating, 18 (the hydrogen' igniters are on.
- 11. -
What you see here is a: series of very small 12 burns , all in :the lower plenum of ' the containment. At 13 -
- this point in time, they are.relatively small burns. The '
14 ' pressure expands into.the upper compartment and there is 15 : _ no problem. 16 As you get into the post-vessel failure and II the core' concrete attack,4you are generating hydrogen and
; 18
. CO . At this peint-in time, you' start.a. burn'in the lower 18 - ' compartment which propaga'tes to,the upper compartment, and
~
- you get a very large pressurexpeak which inithis particular
- 21.' calculatio'nfwasl assumed to' fall. containment.
. 22 - The alternate scenario that I alluded'to earlier
~
' 8 would be that either you do;not'getLthis -- if you don't 24
- geti.this kind 2 of. burn ~or if it-is a2 smaller burn, the 25
. pressure would go up andlupidue to the generation of non-J k '
r V
E , i I 93
- 1 - condensibles and it would probably fail at some later point 2 in time.
3 MR. FULLER: -Ed Fuller, IDCOR Program. 4 -That seems to be'an awfully short amount of time s' from the time of vessel failure until the point where you 6 have your global burn to generate that much carbon monoxide 7 in'whatever' hydrogen'you generate, even if you are treating 8'
~
tne' cavity as not having any water in it. 9 Do you have any handle for the relative magnitudes 10 of J the -hydrogen generated and carbon monoxide generated from 11 the core debris concrete attack relative to the original 12 hydrogen produced in the vessel? 13 MR. CYBULSKIS: Those numbers were calculated. I 14 do not have those numbers at my fingertips. This burn here 15 1 dia involve burning of both'the hydrogen as well as-the carbon lis - monoxide. 17 - -
- I can double check and find .out -how much of each there 18 was.- I don't have those numbers. I think it would not be 19 appropriate to speculate but I can find that out.
20 'I think that does it for the presentation of the
' 21- thermal hydraulics from my viewpoint. Are.there any other 22 questions?- Everybody is tired..
23 (Laughter) - 24 . MR. CYBULSKIS: Mike, are you ready?
- Mike'Kuhlman -2 will-now talk about-fission product release from the fuel and
7 94
'l-
'from the' primary system.
2 MR. KUHLMAN: I guess before I really begin talking 3 about the specifics of the sequences for Sequoyah that we 4 considered -- and there really is not'a lot of detail to be 5: presented regarding Lthe primary system . transport and 6 deposition --- there are a couple- of ~ issues I wanted to address 7. generally and quickly, which sort of-vary in connection 8: with Pete's presentation. These regard the model of.the 8 core slumping used in the' MARCH analyses. 10 - A number of the-implications of or the results
~11 that this selection of a model has on the MARCH predictions 12 -
of core temperatures we talked about'already. But that is not 13 where it stops because this continues on through the entire 14 primary system behavior of the accident and the primary system 15 response to the accident loading. 16 The question was raised in earlier Peer Review 17 meetings about the CORSOR rates being.potentially too high, 18 too low, or-whatever. Jim mentioned this earlier. I want 18 to elaborate ~on-it briefly before I get started. 20
.I think the questions related to CORSOR fall into 21' ' three categories.
2t' OneLis, how good are the MARCH core nodal temperatures + 23 which are used to predict the release rate; that is the first 24 item that needs to'be addressed. This, of course, is what 25 also is-related to the core slumping model. F )
. m
95 0 1-TwoLis what effects of' geometry are taken into
~
-2 . account,:and the answer is,fnone are taken into account in
.CORSOR.
13 The core begins'to slump and lose geometry such that
.4 - you.have small pellets and' globs of molten fuel scattered 5' - about in'the-system.- Is this going to give rise to a
~
'different : release rate -from that node than if you had y so'mething more ' representative of what was actually taking
~
g_ placeinLthe experiments upon which' the CORSOR' release rates
, are based.
10 The : third question related to the CORSOR output,
- 11 just how good are the-release rates themselves at approximating
-12 the experimental data which are available.
13_ Yesterday I received a draft document from Oak g4 ; Ridge.. I am not sure whether Tom will be talking about this, 15 about CORSOR in his presentation or not, naybe a litt'le. 16 -But in any event, that is being looked at, and what 17 was done was an attempt not really to model anybody in 18 particular's ' data set but using Oak Ridge 's expert judgment to come out with what is the best or what would be the
~
ig_ 20 consensus if'you get such from the experimental results that 21 are available. Zt S much for the CORSOR question. g The other effect that this core slumping model has y on the primary system responses to the accident is in the g amount of steam generation that takes place and the concurrent 1
- g. _ _ . . ..
c 96 2 1' mass flow rates ~thr'ough the system. 2 I think we tend to have a picture of the primary ; 1
- 3. system as being a nice quiescent, steady-state sort of a
-4 pro' cess'where you have constant. rate of heat input to the
~
5' water, a nice steam. flow rate through the system which is 6 easily.characterizable and gives you' simple to interpret 7' results. s -It turns out that that is not a bad picture if 9- one uses the 75 percent core slumping model, if you want to
~ 10 . call it that, such that the core h'angs out there until it is 11 75 percent molten and suddenly falls into the lower head of 12 tne vessel, which does not-strike me as'particularly 13 realistic. ,
14 The current model, which has the intermittent 15 ' dribbling of molten fuel' pieces down into the water, which 16 ' remains in the' lower head, gives rise to bursts of steam
' 17 and a-very erratic mass flow rate of gases-through the 18 primary system.. It also gives rise to fairly erratic 19 . generation rates of aerosol.
20 I do not think there is any question that by the 21 time' core slumping begins most of your semi-volatiles -- 22 being cesium and iodine -- have left the region of the core. n But as far'as the aerosol generation rate is corcerned, this 24 is going to vary. enormously,-depending upon the core 25 . slumping'and the partial quenching that occurs as the nodes
v. j 97 1- hit the' water. l l
- 2' The results.that will be presented to' day and 3 tomorrow for the primary system retention,.what you are' going "4 to;see againLand again is the effects'of varying flow rates
~
5' during the sequences , that we looked at. There are periods a sof stagantion, periods of high flow'through the primary 7 system, and obviously.these have' great implications for the 8 aerosol ~ retention.
.g ESo, without further ado, the sequences that need to 10 : be talked about for Seguoyah are the TML and TMLB, which 11 are for'the. primary system indistinguishable, and the S2HF tt ~ sequence which is considerablyfdifferent.-
13 The rate of emission of the sequences modeled 14 in TRAP MELT is presented here for the S2HF sequence, where
~
- us the scales have been adjusted as indicated on the plots is themselves. -All the times that I will be presenting, as 17 .Pete alluded to, start.with the beginning.of core melting 18' as predicted by MARCH.
le one can see here a fairly steady rise in the m' . amounts. These are cumulative plots, so'the total mass
; 21- released increased up until this point where there is very
- 22 ~1ittle release remaining to take place from that point on.
.g; So, most of the primary system response will take 24 P lace 1through here. ,
25- The.end of the data presented in the plots is J
, - e- m +- r ,,,e-
98 1-concurrent with the-bottom head failure, at which time any 2 suspended materialiis assumed to leave the vessel and enter 3- the containment. So there is a slight difference in 4 appearance to:the rate of generation of these materials for the- TMLB sequence and for the TML. 6
.In both cases the' amount of aerosol generated over 7 the course of'a core melting during the in-vessel failures 8 ,of_th'e melt is between 1-and 1.5 tons.
9 I put this up!for one reason. This is the sort 10 : of input which:we give to Sandia to use in their VANESA 11 calculations. As you can see, the cesium and iodine are
~12- exhausted, the inventory is. exhausted.
This is what remains 13 in the melt as it leaves the vessel. 14 - The tellurium content for these two melts, however, 15 - is considerably different_due to the-extent of oxidation of 16 the zirc in.the two~ sequences. You will probably see in 17- -Dr. Lee's presentation some results for tellurium release 18 from the primary system which will at first seem pretty 19 - surprising until you recall that most. of the tellurium was 20 still in the melt for the TMLB sequence. That was considered 21- in'the figures that_he will present as being released from 22 the primary _ system. It is at least'available for release 23 during'the VANESA code predictions. 24 For the S2HF' sequence'we have -- what this table 15 is , as 'you. have seen this type of . table before, for each of
p -- t
+.w 99
;1'
~
m the - materials considered"in. TRAP MELT we have the total which
,. <2- 'hasTheen emitted by.the core. This is not: released from
- 3' the prima'ry: system,-just emitted into the primary system.
4 Here we-have'the total mass retained as a function 5-
.of-time:since the' start of core melting.
6 ~ What takes ~ place in this particular' sequence is, 7: you have ' fairly low flows up until.approximately 650 to 700 8 seconds after~the start of core melting.
~8
'We really should be presenting more mass-flow
.10
- rate residence' time: type figures:than we have available right 11 now. .But.from this periodLto about 1200 seconds and a little
~12 bit beyond you have very high flow. rates, about three orders 13 of magnitude higher :than previously,1 which results in 14 st'opping the amount of aerosol being. retain'ed, really, due is -
to the maintenance of a' low concentration of aerosol by the le ~
;high gas flow rate.,
17 Secondly, .this aerosol,. what is emitted during 18 this time, sees a fairly short residence time in the primary 18 - system. . So,.there is not as much time avaIlable'for it to-
~
. deposit.
21 - One'can seec again towards'the end of the' accident, , ,' 8 .the end of the'in-vessel phase-at~1 east, as the-flow rate 8-
-drops offLagain,;the aerosol retention begins to increase. ;
24
.These are' presented on-a; fractional 1 basis in the 25 ~
-next? table. :It'is probably~alittle hard to read.. But;the i
+
w c- c y -,,w,, , e- --nw e- , s ,e,, ,w- r-g a
100 1 totalsprimary system; retention factor for each of the 2 species, as' presented in'this first column for each of them, 3; 'you can see there are 75 percent retention of the cesium 4' - -iodine;182 of the hydroxide; tellurium I will talk about in J5 a second, and . 84 percent lof the aerosol for the S2HF 16 sequence is predicted by TRAP to be retained. When we go into the tellurium, we are going to see
~
- 7, 8 :in all cases today and tomorrow.almost complete retention in 9 'the ' primary system of whatiis released from the melting core.
-10 ;This'is before the core leaves the pressure vessel.
11 MR. CASTLEliAN: .These are cumulative percents up 12 to-that time, are they? 13 MR. KUHLMAN: Right. 14 . MR.. CASTLEMAN: So,.what this means is that some m' has been lost again when you come down -- 116 MR. KUHLMAN: Yes,~_I mean some re-evaporation of 17 cesium iodine has occurred at this point.
-18 So, to1 characterize the accident or get the bottom-19 'line numbers people love,,we used this as the retention 20 factor _which one would apply to the-sequence.
- 21 Okay, for this one you-recall that there was about zn. 20 percent of the tellurium inventory still remaining with' 23 -.the core, so the other'80 "ercent is retained with .98 24 ; ' efficiency. If you are worried abou't that, it can be added.
5 'MR.' CASTLEMAN: Somewhere it is probably worthwhile
s 101 1 (putting'downfthe:words " cumulative percent." j
~'
2~ MR. KUHLMAN: Okay. 13'
~
, MR.' CASTLEMAN: When you usually.look at these
-{ -things;;you.must. assume th'at.
It is not always clear in
~5 _the pre'sentations..
Si ~ MR.jKUHLMAN: . The.only. point to be made here
'7 is, a little bit of.the aerosol' behavior:is made, I think, 8 -a little' bit' clearer.
9L What is going on, we-have the high retention-here
- 10 : and it stops off. We can;see that? the coreLis no longer
.11' . accounting for quite as.much of the retention during this
'U- ~perico of~high flow-rates that I talked about.
13 : What is happening isedhat the material is no longer 14 - residing;in the core, itiisanot available for deposition 15 2 there any longer, it :is 'being swept out into.the further 16 reaches of the primary system.- During this high flow rate 17 period'we see-it: transported to the~ steam generator. 16 Later on in the sequenc'e', as we continue'to have 19 s aerosolEgeneration and the flow has-stopped, this material is 1B , stagnating, if you-want'to-call it that, in the core 321 Jregioni and the' core again begins to -become a. more significant 22 contributor'to1 retention..
- 25- If one were able.to keep;-from failing the vessel, L
24J 'of course, this 22;.and 25!would continue in that-trend and
'~
15; . give7y=ou ' higher an'd higher -retention i fractions in .the' core. u
~
;x ". _ _
,m-102
,a.
' Il
?MR.'BARI:' Bob'Bari',EBrookhaven.
+,
2?
, : Myjquestion is k Erhere is the silver from. the 3'
; control $ rod.in.your CORSOR: analysis?-
41 - MR. KUHLMAN: from the'CORSOR analysis all the
~8~
msterials"other th'an noble gases, cesium, iodine and
; tellurium are treated asibeing lumped together in a generic 17:
- aerosol',Cif you ;-will'.' .
8' We arelnot tracking'.the~ silver.in' dependently,.nor the. cadmium...:Biat I do"know that it is.all'added up into I
- the/aerosoll and for. at least one -of the - sequences that I
;11;
=have examined in a little morefdetail thanithe-rest, the 12 Laerosol..is' composed of about.70 percent by mass, silver and
[. 'II - Nadmi % from the control: rods. I' This is using.a, I:would say,(fairly' primitive ,
.15 model of ' release of.. the : control rod materials; thatf-is all 18 '
that' is:available.. But what'was released from the control I ~
' rods. was approximately one -half .of the silver and cadmium
~
'nventory i as aerosol.
19 . ~
.MR. SILBERBERG: .Does the silver release model 8
- follow'a'. temperature 3 time distribution,.to some extent?' '
21 MR..KUHLMAN: It 'is: ~a real first cut: at it. It-
;22
.1,cb'ased.on the recommendations, again,-from Oak Ridge.
'8
- which in an attempt to approximat'e. experimental results 24
;which I~think they have; attained.- I believe there is some
~ 8 e e. -
"SASHALresult{factoredintoths
'5-
-- ( #
.. . - . . . . . = . . _ - . .. . . -
- 103s i
- 1H
- Without belaboring _ the~ point, what takes place 2- .i~s'.thatLduring the bursting of the control rods which is 3,
_followed- you : release: five percent of. the inventory. From
}
4-
'there until the time'of' control rod melting, you release
.5' anotheri.50 percent of the inventory of silver, cadmium
;6 Lind'i um,: and ' from that temperature up' to some maximum 7: l temperature -- which escapes me now -- you release the I 18' : remainder of.the inventory. '
9 There-is.no " time-at" temperature based dependence 10 in this release model, which -is _a definite shortcoming. 11 - -MR. RAHN: Is;there any particular assumption s 12 =
~a bout silver iodine formation in all of1that? i 13
-.MR. KUHLMAN:: The assumption is that there is 14 ,
4 s 3 none. 15 - MR. CASTLEMAN:' These-numbers under aerosol, they a ;' > ' ' 16 are : aerosol to the deposited out. on the surface?
; 17 '
MR.- KUHLMAN: Correct. 4 18 I MR.' CASTLEMAN: Would it not be better to put that
't
;19 - down because everybody-usually uses the' word " aerosol" to
'20 ' '
~
. mean -- the word " aerosol" means' particles in a gas.
MR. KUHLMAN: . Righ t .- 122 ; MR.' CASTLEMAN:
~
-It seemsLto me'it'might be better T 23 i just Rfor_ clarification forithe future, people. reading this 24 thing get;very confused between what is really airborne 125 aerosol and what:are aerosol particles'. You could just
- g w Y
.a-.
~,, 6- g- , y -
.q.,-- - -,e ,- s- e,.- oees-,o-. g -- g ,q ~ - sp -
' ' " ' ' " ' ' ~ '
+ ,
^: < 2:
~
1' ' ^ "~~ ~
~
y qf . . cv
..104 .
_. _ q , :1 Eclarify'. tElat. .
- r u, .
_ . g, W } [2 . MR.iKUHLMAN: .'Okay,;. call it "non-aerosol." ik - 3! IMR. SILBERBE'RG: Settled aerosols, u '.;; 4 That would entail. going.back through
^
MR.:KUHLMAN:
)5 severa1[ earlier volumes as'well.
It.-is.' confusing when_you first
. 61 5"MR.l CASTLEMAN:
7.~ re dd ..' it . . w. ' " ' t-n _3 .MR.LKUHLMAN:--That is-something that~will have to
" ~
Ibe made clear'in the' text,'I-think; Ewe have too many of:
~
i 9.! I 10 "these t$bles'to~: correct already. ^ j ~
.MR. CASTLEMAN: Just a footnot with a star. t 12
-MR.-SILBERBERG: LI don't think that:is a big j problem ~ to : correct 1those tiables .'.
i.
' 14 , . MR.' ~ KUHLMAN : Do you.'do typing?
15 -(Laughter) i-
. 16 - MR.-ROWE:.< Don-Rowe. -
Just a small point'of- '
- - 17 clarification. I tried'to add.these numbers up.to one,
- j. t18 'maybelI-.am missing afpoint1of;* definition. .What'am I missing?
4 s b 4
.18 MRS KUHLMAN: That is something-I~also should have 20 s
pointed lout. These numbers'should sum approximately to
- 21-
.this --.they may'not'.. The reasonLwould be.l deposition'in
. 22 Lother parts.of the primary system.
23 ! The RF is the total' fraction'of the material 24' .rele'ased-from:the core up to the-stateditime which has been 1 -
/
- 255 ' retained somewherein the primary.l system.
f *
+ ; ,
.e g .
m 5 .g T -' I **g 1 'j- e *" I * '- a-
- r. , ; ;, , , y= ~* '
}h .W ,
O', g; '
,- ;;i; j(
\
~
1 -
+ .
s
, . -4 105 i <v 7' -
v- . 17 .MR.iROWE , 'And those four ' items, then, to the right? u
~
L2_ ;MR.!KUHLMAN:. TI;would'say25" percent of it has been L't a 3 ~- ret'ined:in.the core, an-d,so;forth. On manus this number 4' : tells,you'.how , y muchxhas~been m emitted. ae and is either airborne, steamborne - -however you want to call'it -- in the primarys
- r /
- 16) system:or-has already been injected-into the containment. ',
.7' '
'MR. ROWE:q thinkafooknote'orsomethinglike s *%
.i a that1would" help. -
*2x ? ,.
s.
- s , LMR. KUHLMAN: To reiterate, theseLare much higher
> ~
10: rates'of~ retention than we have seen pr biously, partly. A 11- due'to the gas flow rates-predicted by the~ MARCH Code. tu N s. .
- 12 Let me also' mention'a change.in:the modeling from
.13 th~e first analyses' that we performed in which we did not 4 14 . permit the gas reactions at surfaces ~to occur in the core 3
15 region for the' cesium hydroxide anE tellurium. That was 16 pointed.out as being an. error, which it was, in the earlier
- 17. analyses. -
~
, r, is ; :Once one includes'the possibility of these chemical ~
is , reactions taking place in the core with'these veryllong 20 residence times that youlhave-for theEgas in the core region,
. 21 you wind up1with a great deal of'this reaction ~ retention of 23 the-cesium hydroxide and the tellurium.- Tellurium proceeds
-,. J ss atia._much higher rate'of' reaction than does.the cesium 26 z hydroxid'e .: w 25 The net effect of'this=is, I'believe.in the Volume I S
'l 4
s 106 t: , 1 .Surry; analysis,:for example, we predicted most of the
- 2 : tellurium ~ to be :available ' for emission into the containment 3 at-the time of pressure vessel failure. This would no 4- ' longer be the case since this material would have reacted 5 .with'the~ surfaces there'and be retained.
% 6 MR.'SILBERBERG: Mike, isn't the steam generator, s7 atlleast the cesium iodine and aerosol, is certainly playing 8' ' a f largeirole here, . about - half. Maybe we will see it when
~
.9 we.come to the Surry recalculated.
10 But how did.that. compare with where you were 11 originally'in'Surry on the S2D,.wh'ich is also a steam 12 'generatar control; is there a . difference?
- 13 MR. KUHLMAN: There is some.. Again, it is more 14 1 dominated by the flow rates and residence times you have
. 15 -here.
16 Without looking at the flows from the earlier 17 ~ sequences--- and:I have.only done that for one in particular, 18 looked at'the_ actual previous flow rates' compared with 18 these -- you really_can't make meaningful comparisons. 20 What would be interesting to do in a sensitivity.
~ 21 L _ sort of a study, which we are not supposed to be doing, 22 ,ould be to use some of the old flow rates with the new
. 23- I ; temperatures. While it would not be very mechanistic, it 24
- would(show you what'the' effects of the,various changes have 25 :been.
,. e , ., m .w, -
., w.. ~
n - -
- =,: -
( , 107-
'1 'MR.' FULLER: .'Ed. Fuller, IDCOR Program.
3-. M:: T
'21 .po'you, Mike,Emodel possible revaporization of
-3? ideposited fission products?
+
14 .MR. KUHLMAN: For' cesium iodine, yes. Cesium
~
, -= :s iiodine._isj predicted Lto'be retained either by condensing on
'8 : surfaces-(ofs the system;or on particle surfaces, i?[ <
- In'either.' case,.if:the surface on'which this
[ material is depositedaheats up enough, the cesium iodine 8 5 i8' 'willibeLre-evaporatedLinto the. system..
- ~
L 10 ' LCesium-hydroxide proceeds via.two mechanisms, s ,
, ill: - a.: chemical: reaction which;is. considered? irreversible, and 12 2
;via' condensation as well.. So, 'you can .see cesiuin hy'd r6xide -
la '- :which condensesfcan be re'-emitted ^in thesejanalyses if-
- 14 1 Jthe temperatures,get high enough.
15 : The-tellurium, however,Lin all the sequences we . 18 ( have/ examined,'the resctibn has^dominatedLthe' retention of
- 17 x tellurium--so: completely that it-is~not' worth talking about- .
la -s Econdensation. 18 i < iMR. FULLER:' What do"you cons'ider to be the heat-up l mechanisms, do yourallowjthe. fission products themselves 21 - leofheati.up th'eEsurface,s? .
.y -
j. 122" N' MR.' . KUHLMAN': That'is-one of.the. things that Jim
- 23 - presented' as beingTan: item .yet to; be done for th'e sequence
~ ~
i # because weMid;.begiri?toelook,'at': that in the first attempt atn J20 : Siarry . . ~ g. 7
'~
c , r ts + ,. s .v.t
. mh '
.tr, _ .' ,
v - 1 0-1 108 1, From what-we could tell then, it looked like there
- 2 twere certain' conditions under which it could be important .
3 For an S2HF I would not. expect'it to be im portant, you have 4
-the st'eam generator available to pick up what is going to 5' ;get shot down the system.
6 For something such as an AB, I think it would be 7-
-more important if,you~could have the fission product heat-up 8
~ of the'~ surfaces which we do not have in this analysis-and
~9 we are looking at for selected cases.
u)
. I 'do not know'whether we have the thermal hydraulics 11 :
yet which factor in the decay,-the fission product decay heat n~ back into1the system. But right now it is just the gases 13 which will'be heating up the surfaces.
- 14 MR. CYBULSKIS: A point of clarification, Mike,
- 15 not-all S2HF: sequences involve the steam generator. In a
' 16 particular calculation we assumed the location of the break
- 17 such that the steam generator. is there. Obviously, if it 18 is there, it is going to dominate.
19 ' .MR. KUHIJ'.AN : Thank you. - 20 MR. ZUMWALT: This is not considered in the present 21 program, but it seems to me the chemical interaction of 22 tellerium with . aerosol, especially- with silver and things 23 'like that in.it, could.actually put a lot more tellerium.in 24 theJaerosol~and. carry it with it. '
- 25
_I: understand why you don't have that, but that is-a 4 i
)
- 109 4
- '1 possibility; isn't it?
2 MR.-KUHLMAN: That is a good point. What is making 3 its wayfunder the aerosol in certain sequences could be
'4 actually released into the containment more efficiently than 5 H we are seeing here.
6
. We really don't have any information regarding i7' 'the deposition of tellurium on-materials other than what we L 8 consider the interior of the ICS to be made of at this time.
m
'9 These were based on experiments with the stainless steel and
~10 Inconel shims which were done at Sandia.
.11- You are_quite right, it could result in emissions 12 of some of1the tellurium as an aerosol . into the containment 13 ' atmosphere .'
14 . MR. THOMAS: . Thomas from Yankee. 15 on the previous. slide you showed.the amount -
. 16 available in a melt. -I am curious as-to why both on a 17 mass - center. and fraction inventory there is more noble
' 18 ' gases.than there is iodine.
19 'MR. KUHLMAN: That is a good-point. This has just
~
~ 20 ' got to be a 'round-off problem because the release rates of
.these' critters-is not. distinguishable, indium and krypton 135 .are completely gone from the S2HFLto any degree you want
- 2 to a talkl about. ]
24 'The TMLB,.I don't know where that -- that "three" j l
- 'cannot be correct. .That is a. good point, it is obviously an
4 110 il error.< w '
.2~ :MR. ROWE: ' Don Rowe. .I have a question on 23 representation of the : system again for control volume.
~
R '4-
'I notice steam-generator,fthe hot leg, and various components
} 5. ' mentioned.
. 6 =How>do you' represent this task through the primary j7 ' system,-how~do you distinguish these various' regions? s- MR. . KUHLMAN:- Th'e order in which they proceed' , 9 '- the flow path is s as Pe'te presented for the sequence, you
~
~
-10 'know,-the core plate, the' guide. tubes,--top hat and core 11- J barrel, hot Eleg, ~ stieam -generator. -
-:12 .. ~ Now, these are characterized in-. the TRAP MELT
- 13 : . code only by a few: vary simple geometric parametsrs, these 14 .
.being a--length, a' cross-sectional flow area and. hydraulic 15 . ' diameters are the-key ones~. There are'others that are involved, '
;16 theLgravitational settling.
; 17
~ But that -is alli the TRAP . MELT code can do.
There-Lare about four volumes, then? 18 I MR. RITZMAN: 19 1 -MR.'KUHLMAN: There are seven in this set.
' 20 : .MR..RITZMAN: -Seven volumes along the path?
~
21 MR..KUHLMAN:;.Right. . 22
-MR.'RITZMAN:: Just like ,in' the steam generator you
'23" 'haveMa1different flowjarea than you would have in the-
~ 24
- primary piping, 26 ~ MR..;KUHLMAN:
- g. Right. This-gives rise to,.of course, 1
s -
/ l .
# w , . 4 .-r +' , 4 ---# me.6 - - , . - . %,. -+,-
111
'l-different flow velocities and different degrees ~ of turbulence 2' ~
and so:on in each of the-control-volumes. But.throughout 3-any given control volume you are stuck with that one value 4: at any given-parameter.
.5 MR. RITZMAN: Well, what was the reason that
-6~
tellurium was held in the core? 7-MR '. KUHLMAN: Wh'at we-are doing in CORSOR is 8 on a node-by-node basis MARCH.is predicting the extent'of
'- zirconium oxidation.
10 Based upon the extent of oxidation, the tellurium
~
11 - is re1 eased according to one. set of release rates or another 12 set'of release rates, again based upon recommendations from
+ 18 Oak Ridge.
14 There is about a factor of 40 difference. If you 15 recall, the 0772 release rate coefficients, we have there A and 16 B sets for three different temperature regimes. We have II' such a set for-tellurium, assuming high oxidation of the 18 zirconium and low oxidation of the zirconium. High and low 8 being, 70 percent being the cut-off point here for the two-
. side. oxidation that we expect in the melting core.
21 If you notice a split here between the- zirc and 22 zirc oxide, you can: see the difference in the extent of 23 oxida' tion, so that this does explain why most of it is held 24 ~ up for the TMLB and not in the S2HF. 3 MR. RITZMAN: -I was not referring to this table, I
~
nu '
~ ~
q,W ' 112
~
'e 1
- was f re ferring - to ;the "next'- one , . thi~s Table-:74.
There
. 2 .'
. tellurium --- the retention:' factor .Iis high and most of :it '
~
3
~
~-fis? occurring in;theTcore.
4 .: . .e MRJ.'KUHLMAN: jThat,is-correct. .'By " core" I mean
- "5' lthe'controi. volume'which contained the actual' physical' core.
/
-6'
-MR. RITZMAN: .Does that mean it is because of 14 7 iaerosol= settling?
?:_? 8: lMR. KUHLMAN: .-This is.due'to' chemical interaction 8> !with surfaces.in this volume,-.:the core-barrel-.
- 10$ - MR. SILBERBERG: Well,.-there iscalso.the question 11- -of zirconium, right,-being. held up in zirconium clad?
112 5 lMR. KUHLMAN: This isfa. separate' issue from that,. 13 that.would : prevent 'its ^ release ' altogether. 1 14 MR.'RITZMAN: Okay, so.it is reaction with'other.
- 15
- st ructures .-
,16 :
MR. KUHLMAN: Right. .'This is' mat'erial which 17 -' -1 escaped 1 the : fuel. -
~ 18 '
MR. RITZMAN: .Yes. 18 MR. KUHLMAN:-~-Then.you would. find-it'.in-the primary 20
- system after-the-core has-melted:.through'the pressure vessel.
. 21i MR.,' CASTLEMAN: )You jus't used a term which is 22 also~used in:the. text, and I did not quite understand. - What. ;
i ( 23 Lis'one-sider anditwo-side oxidation? 'l 24- -
' (Laughte r) f
~ '
'28 lMR.:KUHLMAN: Would.-our friends.at Oak Ridge. care 1 -
1 6 d Y 1' c
- c
n
~
113 l-
.to elaborate'on.that?.
2
-Dick Lorenz is unfortunately not here.
3 _
-: MR . CASTLEMAN: I just don't understand what you
~4 meant by'one-sided and two-sided.- Inside, outside plant,
'5 is'thatLwhat you mean?
6 MR..SILBERBERG: Yes. 7 MR..KRESS: Tom Kress, Oak Ridge. 8
- Yes,-that is the inside or the outside clad. It 8
depends'on whether the steam can get to the inside or not. 10' MR. CASTLEMAN:.-That is worth clarifying in tue text. 11-It1is not totally clear what one side, two' side means. U2
-MR. KUHLMAN: For-the TMLB sequence, which is E3 characterizable as high pressure, most of the . time it is a 14 very :lowiflow rate sequence- but you will see in. these results 15 that ' there is evidence of a surge' in the flow as slumping 116 begins.
~ 17 Again, the total masses emitted for each' species 18 -
and'the total masses retained somewhere in the primary 19 system is a function of time.
' 8 The flow rate is quite low up until in the vicinity 21 of 1,100 to 1,200 seconds. This results in a fairly low E'
amount of cesium iodine retention since most of this material 23 has not iescaped the core _ region as of yet and because it is 24 far-too hot in the core region for condensation-to occur to 8 any great extent,'either on the particles or on the surface
y_
, s -
.j.
^
. s .,
s 114 1 c.within ' that ' control ~ -. volume . . 2 , CesiumIhydroxide, on"the other hand,-is being retained;even-though:it.'is-experiencing the same temperatures
~
3
~2
- . 4 ' . idue to.this chemical _in'teraction with surface within the
's-coreJregi'on.-
Tellurium, again, is retained almost with-a
~
-6:
. .e f:- y s hundred-percent efficiency, .and one can see the aerosol
. Esi being retained quite effectively _through-this' period'of low s ': flow.
10 < 'Later.'on; fat 1,250las.-the flow. increases fairly 111- . dramatically, cesium. iodine-retention actually. increases. L 12 ; I Tne.-cause.of this, when you look-through the output in-13 . detai1~,iis that.the cesium iodide-is' reaching regions of the 14 = upper plenum which are. cool'enough to permit condensation 15 on particles which are subsequently deposited.on surfaces is . in. the.- upper. plenum. Likewise,-we-see-relatively.a' decrease in the
~
17 18 : tamount of. particle retention because the material is being
- 1s swept through more rapidly, through.-the primary system, and
~
, .l 20 - .resultst.in aflower concentration:of; aerosol at.any given n
.' 21; Ltime; Tas Lwelllas 'a- lower instantaneous residenceitime ,- if
- 22 tyou'want'to call-it.that. 23 . So,- this' material has less o,f a chance to deposit' LanaT.is=not" driven as strongly by ccdgulation due to the lower
~
24
. 25 < concentrations.
W x t' y ., d -- 6
-115
't. The same information presented in the same manner s
2- 'as before. We have three- cparters of the cesium iodide 3 being retained, approximately'all of the cesium hydroxide
~4 and: tellurium,.and.90-percent retation of the aerosol. Most 5 Lof this being in the core due to gravitational settling in
- 6. that-volume.
- 7. Perhaps these aret a little - bit easier to pick up 8 the trends.as to what is going on. What we have here is 9 -the cumulative mass. retained in the primary system as a 10 function of' time since the start of core melting. This 11 . broken curve here represents the. total amount of cesium 12 hydroxide which has left the core region.
13 The solid line is.the amount which has been retained 14 in the core. Theulong dashed line is.what is retained in-15 the core and in the upper . core , plate. The small difference
- 16. here is accounted for by the remainder of the primary system 17 -and,. finally,.this amount would represent-what is suspended 18 at this time.
19 The ' difference here at the end of the sequence 20 indicates what has been actually injected into the containment 21 atmosphere.
~
.22 . Recall that cesium hydroxide is permitted to 23 : interact with the surfaces in! het core region, whereas
- 24 cesium iodide is'not, which results in a peculiar profile
' 25 :for the cesium iodine.
r y := m 116 il
^
One can-see'some initial condensation on particles
- 2 in - the: core region, veryllittle, and there the re-evolution of 3 .this!materialiat_this! point. A drop'in these-lines would be 4' re-evolution _of the cesium iodine. You:can see the. transport
-5 into the' control'; volumes-downstream'of the core.
6- ; Finally, the entire primary system response or 1 7 retention is_ characterized'by that upper-most dash line. s' - Again, this_is the' total which has been: released from the
'8 ~ core'.
-10 'The aerosol behavior is as one would expect
.. '11 through here, the nearly complete retention that we showed 12 on'the tables before. . Flow rate increases, and we wind up 13 .getting a bit lower' retention for a while which causes the 14 - divergence of these two lines.. So,Lyou actually-do have M some suspended aerosol at the time of'RPV failure which 16 would be expected-to be injected into the containment at 17 that time-. It'is a.veryfshort duration source term. 18 Nevertheless,-theltotal retention, as you recall, 18 was .about 90 percent of the total aerosol emitted. El MR. SILBERBERG: 1As I recall,~ Mike, again back 21 tofthe;old'Surry analyses, you were at low flow. You had M lowfflows'fo'r MARCH, ILthink.. The temperatures.were higher, 23 Inknow. . j But were :you retaining 'as much . aerosol back in the
~
. 24 8- old Surry?:
tn5-
a :- 117 i
.1 .MR. KUHLMAN: We were retaining just about the
'2 Esame amount.
3 .but. SILBERBERG: Just about the same. 4 MR. KUHLMAN:- But what is_ interesting is that
'5 -in the' previous Surry analyses,.at the time of the pressure 6 vessel: melt-through which I think was a little bit later 7 going to start a core melt than in here, we really had not
- 8. . yet . injected anything _ into the containment, whereas L here we
.9 have included in1the analysis an inert species just to' trace 10 what is happening.
11 . We have" injected about a quarter of the material
~u which is injected, which-is'realeased from the core early, 13 - ihas made it into.the containment before RPV; failure for 14 'these cases.
15 ThisLis due to the fact we have a period of several 16 ' hundred seconds where: we do have a ~ pretty - significant flow 17 rate through the primary system. 18 So, whereas before the whole TMLB source to the 19 containment would have been a puff, if you will, at the end
- 2 of the sequence, of the end.of the RCS part of the sequence 21 anyway, here we'have a low-level source up until the time 22 of RPVjfailure. 'Then we have this additional puff.
L2 MR. WALKER: Mike, do you include in your modeling 24 . .the vent line'between the relief valve and the quench tank,
~
' 25 : deposition therein?
~.
x -- -- ~ .,.. , % 33 7,
~ ^
7 __ ' [:^:88. , ' s-
~ . :
?. :,
7 - n .,
- 118..
x . , c. o z~t # , .
-. - l'g , -
.- 1( - ' :MR.IKUHLMAN:. We have'Lgone only up- -'to the -
_l2 ipressurizer;?andiassume:that'from~the. pressurizer out. ~So, the y [ ;' .k i3' 7 actual! piping from the. pressurizer - to the~ ' quench . tank ; has 241 1 noth bleen[ included. H e, . . _:
^
25i ' h !MR. , BUTLAND: . ' Butland , - UKAEA. - m,?
~
;t:- : 7g; DILknow:the:. TRAP. MELT code:does-not include any y
,,- ~ 17/ /models of resuspension,5 particle resuspens' ion. from tlieI walls. -
j M .8
^
,You:haveLspokenfabout these steam spikes,shall 18 we call them,~these sudden;> changes in steam' flow and 110 pdrticularly~s'o!when molten; core drops-into the base of the
^ ~
!!!L ives sel~. .
- 12 ; .How much of an approximation-dolyou think'this lack i
, , 13l :of resuspension might-be-i~n view-'of these slightly changed
~ 14 : . circumstances and in view'of-this drop:into the-bottom. head?
15 MR. KUHLMAW:-LWeactually have.not analyzed this is - 'at'all, the possibility of resuspension; For-one, the
! !'? ' situation 4is'that"the:particlesJwhich-have been desposited
- 18:. rangefinLterms ofJa; mass median diameter..from three upwards- l 118 s Tto1about as large'asysix to eight microns'iit some cases.
20 i - This should be'afdry: system which you would expect 21 - to' have. a good possibility-: fora resuspension, especially of
' = -
sthe?largeriparticles. .ButLwe have not analyzed this at all. e 23 , 28 .I-see it as a potential? shortcoming in.the'modeling
+
26 o'f-;the primary system..:Onet thing-to keep in mind, however, 25 - '.'isithat barring a pre-existing. cont'ainment failure, if you :
'm ,u I
4
' = .. , ,
'_{',
s A
x ~ . .
, , . y_ . .- - . -. . _.., . . . ~ .
~ ,
.;l
- s
'~' '
~
-k -
m w ._ y, 119 y 1. - 14 ~ ~
- will,:'the: particles _that are likely to.be resuspended during
=7 - 2:
- ~
- these j pulses; of highiflow- rate through the primary system .'
3: ~ are very likely - .wel1, they:have.been deposited in the y lo: firstOpla'ce.because.they have-an enormous. settling velocity.. h *i 'L8i i.
~
Becausecof that, their: impact on-the airborne mass:insthe. 4
;6 -
4 containment, ILthinkt--would be; appreciable only very ~early '
-7'
'n i the history;of'their' inject' ion into the containment ~
-f 8-h.
1
^
.- :They .would:be eypected to be settlingjout:there
- e
'. 9 more.quickly than anything else, but.they could represent a '
- 10 significant part of-the mass available for a,very, very early- '
' O
.11 containment failure, f 12~
MR. GIESEKE:. Jim Gieseke, Battelle Coluinbus.
. la
+ ' ,' I would like to make~ a' comment' that -I made at one 14 . of the other meetings. regarding resuspension of : particles - _ .~ - 15 ' and material from the su'rfaces.. ' h . A 16-Although it is dry.in.a lotiof respects, especially i 17 e in terms of water, we are condensing the other materials onto 3 the surface-7 18 t L. I' think ?we -are: forming sort 'of a liquid cement 18 f.. on the surface rather than a deposition of dry particles.. [ 20 - i: It is-interestingjeven in-the.MARVEKEN - experiments ' [ : 21' where.they have steam and someLof.these materials, the cesium 23 ' } hydroxide complexes with;the water. - Itimay have-been as a j- 23 - particlerat~one time,-but it is as a liquid running on their [ ' M. surfacessin;the pipes. '
- _ .s ,
SS { - I think what:we.probably have on the' surfaces _is not= x 1-x'Ar g d I _
, t wy y , , i1---.y- 3 , , , . ev < w re 1-
-w..w-.+ -,e- ,, ,,..,-,,-9ee,3
- F , , . - , , , w,,,s , y w . - + y , e.g m.e_,.++,p-r
r .- v a-120 I' 1- a collection of dry : powders but is at best a paste or a 2 real. thick liquid..
-3 MR. FULLER: Ed Fuller, IDCOR.
4 I would like to-continue on the resuspension 5 discussion for a second, and that is in the context of when 6- the. vessel itself fails for this particular sequence you are
,7 . blowing down from a couple of thousand psi.
s Are.you still. confident that material won't"come 9 off? 10 MR. .KUHLMAN: Not at.all, I am not. We really 11 - have not analyzed for this, we do not have a basis in TRAP 12 MELT code for calculating any resuspension. We can get 13 at1 the gas flow properties which I . think you would need in 14 any attempt to model the' process. But it has not been 15 included in TRAP, not for want of desire on our part to 16 incluce the process but there have been other factors we 17 have been dealing.with. 18 MR. SILBERBERG:' We have an experimental program 19 that has gotten under way at Oak Ridge. This is another example, another area where,:if'you will,.the speculation is 20 21' much farther ahead of the experimental evidence. 22 Bob?- 3 MR. HILLIARD:. I was just going'to say that.from 24 some experiments that I have been involved with, it is. sort 3 of;like. Jim Gieseke has said that'any time aerosol,that has.
~
a
- m. .
y m g-q eg .g * ' ' [ - y + ' ' m t ?le . .!been:' deposited.'in pip'esfisires'uspended by high velocity M
~
L-2 _ gas, . tihey 's'eem to be ' dispersed 'in very, .very 'large: particles J 3' iof . several~ hundred microns land they settle- in just a few 41 ; minutes in'a largeicontainment-volume; 3y :So; my observations.. substantiate that... MR.LKUHLMAN: 'That is! sort of what one would expect
~
, , , 4 79 from,looking<at the'old data that people had,-in trying'to --
- 8 when:you'are_looking at the actual. hopping.along..of dust
- o[ : particles 1. down . tubes , and'~ so . on ,-: that 'it = is relatively' easy to . to get;very large ' globs off, ;even if it hit the surface as:
it _ - these ten: micron and smaller-size particles, they do associate is with one_.'another there and form a1 bond:there that can't be 13 . broken.up=by a-simple air velocity. 14 : It:is_. going to_affact=the ultimateLres'idence of this material whether it is in the containment-or'in
~
13 - le_ the primary system. We:have notlgotten into that sophistication-
-17 _yet, I don't think, to thefextent itiwouldibe nice to.
fig - B'ut really, when you are blowing down;in the TMLB 7-sequences you would certainly expe,ct to liberate some 'of
~
- 13 3 3 this! mess tha't has.-accumulated in'-the' primary system. But
- 21 whether11t hastanLimpact..on..witat.would:make it in the 3; -environment,LI..have serious doubts..
t ' 3: MR. REYNOLDSI LReynolds, Virginia.
.sg ' These . retention b:-factors, seem ._to 'be . pretty : high ,
m- which;is; good. .Are1you very sensitive to theigas temperature? L-
.r 4' j t # y
r a 122 ! You are assuming that the _ outlet gas temperature is what,
~
IL 2 the temperature of the gas going into the primary system. 3 There wasia lot of debate on whether that was too low or too 4 high earlier. 5 Are you very sensitive to that, do you thing? 6' MR. KUHLMAN: Personally.? 7 MR. REYNOLDS:- Well, from your experience, too. 8 It seems like earlier you were assuming higher _ temperatures. 9 MR. KUHLMAN: Also in the earlier analyses -- and 10 this is a key difference as well ---we had less aerosol' 11 available.. 12 We also had different flow regimes. But'what I 13 was driving at in the introduction of this talk was that 14 the biggest difference that I am seeing from these new MARCH 15 - analyses is due not to the surface temperatures or gas le temperatures that are being predicted, but due'to different 17 flow rates being predicted. 18 The gas temperatures would have to be quite a bit 19 different before we really would see the effect on retention 20 of cesium iodine. That is where you would see it'first. 21 The numbers here, while they are greater, show Et greater retention of cesium' iodine, it is in large partLdue n- to-the condensation on the increased loading of aerosol. It
.M is also' partly due - to the, increased amount of surface to
.25 volume ratio -- if you want to express it that way -- in
123 a. 1- Ethe upper plenum. 2 While there is some sensitivity, there really have 3 been just too many changes made in the input to the code to 4- back out how sensitive are we to temperature. 5 MR. SILBERBERG: Excuse'me, I missed something.
,6 Why do you have hi'gher loadings'of aerosols.now than in the 7 past?
8 -MR.-KUHLMAN: We now have the control rod materials 9 being released-as well.. 10 - MR. SILBERBERG: Okay. 11 MR..KUHLMAN: In the early Surry analyses we did ut not include them. They constituted, in at least one of the 13 sequences, 70 percent of the aerosol mass loading. 14 MR. SILBERBERG: Now, okay. Bob? 15 'MR. RITZMAN: I'want to.try to understand why 16 cedium hydroxide and tellurium are retained in the core. It-17 is reaction with solid surfaces, right? 18 What is used for temperatures of .those solid-19 surfaces in the core and what is their geometric distribution, 20 and what is the mechanism by,which stuff gets from the gas 21: phase to those surfaces to react, that is not being swept out? El' .MR. KUHLMAN: Okay, what is going on"in the core
? is, it is characterized as having a hydraulic diameter which 24 is appropriate only for the core region at the start of.the 25' accident sequence. As ' the core ' loses geometry , we no ~ longer l
r -- ' 1 l 124 1 1- have an appropriate hydraulic diameter, which of course 2' ' af fects you surface-to-volume ratio that you are usihg. 3 The surface temperatures that we use, we are using 4~ one surface temperature which is predicted by the MARCH 5- code. But that really does not enter into this process. 6 This is considered to be a temperature-independent process , 7 this vapor reaction with the surfaces. Dana at Sandia 8 has provided these deposition velocities which we used in 8 the code. l 10 MR. HILLIARD: Temperature independent? l j' 11 MR. KUHLMAN: That's right, i 3 Over the range of the conditions that were 13 . examined in his data, these are not precise data by any L 14 means. But over the range of temperatures he looked at and 15 . .the states'of oxidation of the various materials, it does not 16 - seem to be dependent upon the temperature,- to the best that l - 9 17 we can approximate it. 7 18 I do not know if Dana wants to elaborate on that 18F or not. l 20 MR. POWERS : The temperature range of the data 21 was-700 degress to 1000 degrees. If you 'are in diat 22 temperature range, there is just not.much difference. 23 ;If you get much higher than that, then I would i 24 think-that the reverse reaction would become-quite important, 25 it quickly? dominates, especially for cesium. l h-
, 125~
-1 MR. RITZMAN:. So, if'you go above a thousand "C"
~
2 ~ you are extrapolating outside the range-of:this data; is that 3 true? 4- MR. POWERS: Yes.
~5 ' MR.-RITZMAN: What kinds of temperatures are used?
6- Is it getting temperatures above that? 7 MR. KUHLMAN: -Off-hand I would say, it would not 8 ' surprise me to see temperatures higher than that at some 9 points'in the sequence, especially late on. 10 MR. RITZMAN: It might depend on the structure 11' you are talking about. 12 MR. KUHLMAN: It'certainly would.. We only have
' 13 one surface temperature available to us in the modeling that -
14 we are doing. This could be improved upon but it would take is some pretty significant changes in the coding to do it. 9
- 16. " MR. LEVY:
Sol. Levy. 17 ' How do you change the characteristics of this 18 ' core structure? I understand how you started out. 19 MR. KUHLMAN: They are not changing.- 2 MR. LEVY: What do you. use for its hydraulic , 4 21 'diameterJand the surface to area-ratio?' 22 MR. KUHLMAN: We.use just the values which' 23 characterize itiat the beginning of the accident. These 24 are not: changed during the sequence at all. 25 .Let me get to the bottom-line numbers in terms of W f r m ~
h _ 126
'I the release from theEcontainment. If there are no further it ' questions, I guess Ken. Lee will take over at this point.
3- E MR. SILBERBERG: I think we are' approaching lunch, 4 so I think we will stop at this-point. 5 . Why' don't.we-try to return,by 1:45. I would 6- appreciate if we could promptly, so we can get started. 7 We-are somewhat-behind but I am not too concerned because
.a I think a -lot of the discussion- this morning was rather L
9 . crucial. So, I'am not worried about time at-this point. 10 We can always stay late. 11- (Laughter) 12 - MR. SILBERBERG: Thank you.
> 13 (Whereupon, at 12:30:p.m. a' luncheon recess-was 14 taken until 1:45 p.m.)
15 16 17
- 18 19 20
. 21 gg _.
. 23 24-26
127 1.
- AFTERNOON SESSIONL '
2 (1:45 p.m.) 3-MR. SILBERBERG:-: The meeting will please come to o 4' order. 8-
'Before we proceed,.I would like to remind everyone 6
of a rule of' order that we have established, I think, one
~
7-or f two meetings ago, having to do' with high-density aerosols
-8 from pipe smoking.
~(Laughter)-
8 MR. SILBERBERG: ' We .would very 'much appreciate it : 11~-
'if' people would-refrain from_ pipe smoking at the meeting.
'O Thank you very'much..
U Ken Lee? 14 MR. LEE:- My name'is Ken Lee,~Battelle Columbus. 8' ( I would like-to acknowledge that Dr. Chen who-is - 16 here contributed considerably on these containment calculations
~
' II -
and Mr. Reed -- who isinot here -- also on these . containment calculations. 8 To; refresh your memoryfquickly, the containment calculations , include the following aerosol behavior 21
- mechanisms. The major behavior mechanisms inciude --
22
. steam condensation onto particles,_ deposition to wall due to
" defusi6n,: and-of course sedimentation and accident conditions.
~
'Those are the mechanisms included'in the original s'
.' NAUA code and in' performing :this ' analysis z it was evident that t
' ~
y '
- 128
_ . ~
.1 ~ .w e.had to modify or. improve the; features,'especially for 2' the S2D . case :in : the Surry plant we included: this containment '
13 .- spray. model . Some of1the mechanisms'forithat collection of
- 4. particlesTto water ~ droplets include dif fus~ ion and impaction
+
- s which was covered in the last meeting.
e Then, after that, we' included.-- and for the i 7 isequoyah plant we:had this ice condenser model which was 8- 3 developed-for this program;by Battelle Northwest. "We took: [ 9 that model and combined 1that ice ~model with the NOW code HL :for analysis of the Sequoyah plant. , 11 Now, this fault ~ formation is something new~which l u- we have never t'alked about before. -Briefly, I will just l ( 13 touch on that and then talk'about the actual containment 14 analysis. L 2 - In the past, the mo' del we have is that' suppose 1s - you have dry particles and then suddenly steam comes'into 17 the containment. and ' the steam vapor condenses on solid' 18 particles. So, your particle will.be wet. l HL If you plot.this. mass or number-as a function of l
-m size it looks like this. Originally, your particle size j
,n defusion'looks like - ~and~then-because-of'the water
- m-condensation your particle sizeLgrows. Typically you'.
-2 out;with bi-modal'defusion and we have seen this many times
~
St . in1the-past analyses. as one of the. questions which came.up during the 4 2 1 -w I- = -g w re - n-y e, 3 . - +f v 4 + 1 t ww 7 - -
- r. --
)
i 129 1
.studytwas that some of the' experimental-work conducted ~at ;
~
i2 0ak Ridge, people felt that some of the- steam vapor could 3-nucleid without nuclides inside the water, especially if 4' you used this NAUA code. 5L We have seen a very high super-saturation ratio
..6 and we-felt that:it is not going to be too realistic, 7 especially'in some of the experiments we have seen water 8 droplets forming whether there are particles or not. For '
'9.
example,.in the NSPP experiment you injected steam when 10' the particulates are present in the containment,_ you see the 11 -
' fault forming or' droplets forming. Presumably, in the U 1 droplet you contain-this nuclei.
M Now, if you do not have any particular in the
-containment and inject.the steam, you will still see a lot 14 15 ~
of water-droplets in the containment. 16 So, one way to -- this phenomenon is to allow 17 - this water . vapor condensed on the existing. solid particles is as ' foam, - droplets , without nuclei inside. We call that 19 ~ homogenous nucleation mechanism or fog formation. 20
^ In the ca'se, smallest case experiment where you 21 saturate the environment with water vapor and expand zt adiabatically, then you have a nice condensation and we all 23
. agreed that the water vapor will condense on solid particles..
+ 24 But in the actual containment we have-seen many lE'
, cases where steam comes-into the containment and if you g w m
pn - 130
~
l'
'actually run the. containment code.we have seen many cases 2
where' supersaturation becomes higher -- I should .say a 3 lot ' higher ethan one. 4 ~ So, in this model, if the condition is met -- I 5 am going to talk about the condition very briefly later -- 6 / water.can form,- this water vapor can condense on solid , 7 particles. After that, this model can interact -- remember 8 ithat these are particles except that they are liquid 8 particles now without a nuclei in - there. Then they can stop 10 interaction and go.into this complicated aerosol -- dynamics. 11 This is'the model we have. adopted and the - 12 condition for the fog formation to take place is given by 13 ' this;upersaturation. .We define this, call it' critical 14 supersaturation ratio, which.is given with this simple 15 correlation. This is out of Greene.& Lane,'one of these 16
-classical aerosol mechanics books.
17 I think the main thing is, it is temperature 18 - dependent here. 18 I would like to mention that Ed Warman and his 8' ' team had a model which is slightly different than what we 21 have.- Now, I am not sure I-am correct, if I am wrong, I 22 think you can correct,me. 23 ~
- What they have done is that instead of a 24
. condensation mechanism they allowed this water vapor to 25 form droplets. But instead of uniform size droplets, they i .
131 1 define:certain size defusion and then let them interact-2- with the solid particles. 3 We found that out after we came out with this 4' homogenous nuclbation mechanism. 'Now, what is the fact of 5 this new mechanism? That is rather complicated. 6 Early on, I'think this mechanism will be activated 7 _and we' feel that it is a matter of how these solid particles 8
- will receive water. Is'this due to agglomeration.with the
'8 ' water. droplets or due to condensation?
10 The effect off this mechanism on the overall 11 ' release fraction,.I think, the release will decrease a little bit, depending on the condition, especially initial 13 condition. This mechanism has been 'implemeted in the NAUA 14 : code and the old. analyses we have done have'this mechanism
- 15 '
activated. 16 MR. REYNOLDS: Let me make a comment before you 17 go on. 18 MR. LEE: Yes. 19 MR. REYNOLDS: I think that the distribution in-3 size of these particles that are nucleating and then 21
- growing, I think that'will-produce a distribution in size lE like the second graph that you drew.
I was surprised when you-showed your first graph 24 of just 'one particle size, ' droplet size,' for your nucleated
^
# -particles. They will nucleate extremely small.
L
p -- ' F V - 132 L I; f'.. 1 'MR. LEE: Right. MR. REYNOLDS: 2 Very, very small and about the same 3 size. . But they will grow. I y 4 MR. LEE: Right. s
- 5 MR.'REYNOLDS'
- You will need to add --
L 6 MR. LEE: No,-this is a schematic diagram. t 7 1MR.'REYNOLDS: For just the nucleation and not the I 8 growth? L
-9 MR. LEE: This is a schematic diagram showing i
10 the nucleation, and right after this, this will interact 11 withLthis one~and interact by itself. M MR. REYNOLDS: That does not seem to me what happens , ! 13 it seems the other would happen. You would r.ucleate-and i ! 14 then.the particle would grow -- L M MR. LEE: Right. 16 MR. REYNOLDS: -- with further. condensation on the L 17 - droplet. You might simultaneously get condensation on the 18 solids, too. l - ! 2 MR. LEE: Right. We have a simultaneous condensatica e 20 - onto solid particles. 21 MR.REYNOLDS: I think that is probably rignt, but 22 'I.think the water droplet ~would also grow in size before it I !- 23 ever joins .with the solid. 34 MR. LEE: This water. droplet will grow, even as without this one it will grow. b
=
(3 133' t ', W 4 [. J p , p 1 -MR.'REYNOLDS: .Yes.. Do you show that? Do you t , . 2 have'a growth-model? i i 'MR.' LEE: 3- That mddN1 is . included in - the NAUA code . 4 MR.=REYNOLDS: Okay. ..
-5 MR., LEE:- Now', previously, what we,did was,'the r
6 only growth. mechanism we have was this. Now.we have this
~
.7 - and they are all tihere in the 'model now.
n .\ ! 8 MR.' . REYNOLDS : . 'Okay. But if you stopped ~at.that,
.9 that'does.not show me:what'you have because*that water a
10 should spread out~ like the water in the one.that you said 11 was Ed Warman's; model. 4 4 Itshouldlrpreadout, it should grow 12 in size- and so you should' have large droplets - of water in
- i. , ,
13 additionthdropletsof.wateronfthesolid. , h, 14 MR. LEE: I think Ed Warman's model, it is of a
,-w 15 certain geometric mean size -- '
3
^
16 MR. REYNOLDS: 'I think eitherlyou have to have a , t s
,s 17 growth model to get.to there - s N 18 MR.-LEE: Right.
- it j 19 MR. REYNOLDS
- --
or there may be enough data in i l 20 the-literature.that, given the initial conditions, you 21 - tould show 2- a s s . 22 t MR.3 LEE: I-agree with you. , Like I s' aid, it just 23 out of'one of.these clasically aerosol-. mechanics book. l_ 24 We . let them grow and the -particle size,' water droplet size 26 - will aventually'look like.this. . , -. w .s.
=
I (h k. , +- \ _ \Q L
n L 134
~
5 1 - JMR. CASTLEMAN: 'I am sure that you aware of this, but' classical homogenous; nucleation period is really a
~
.2. E 3 Every cru'de approximation ~ which of ten does . not work very well.
I4- . MR.' I LEE : Yes.: MR. CASTLEMAN:
- 5 But within the limits where it 6 does, as an approximation.. What sort of number of initial 7; primary water particles per cc would you get:at the onset 8 cof . condensation' compared to the number of solid particles 9 ~that might'be existing? ,
ICL MR. LEE: Right now, I'think we are using 25
~
11- -Angstrom in radius. Now, you'can. argue whether that size ft - Lis appropriate or not.
, 13 - -MR.-CASTLEMAN:- But if you use that, what number 14 , density would you give you, water droplets compared.to the U5 number of density --
16' MR. LEE: The number concentration is very high. 17 -- .Then, what you find is, they quickly grow. The number 18 . concentration will decay and the. size will increase. MF 4 MR. CASTLEMAN: : B ut forgetting about mass, just 20 ' number. compared to the' number of solid aerosol particles, is 21 it many orders of magnitude?. 22 MR. LEE: 'Right. Now, like I said --
' El MR._CHEN: Henzwe Chen from Battelle.:
24 'Maybe I can answer his question. I think our 25 model here, the saturation ratio now exceeds the critical
~
I 135 1 ratio, then the homogenous nucleation is not going to happen. 2 So, it does not matter really how much water droplets are 3 forming compared to a -- particles. It cannot be jumped 4- from 1.2 to two instantly. It could exceed a little bit 5 then starting homogenous nucleation, forming the 25 micron 8 size droplets. 9 7 So, I could not honestly tell you how many 8 particles, -I mean how many droplets are forming. That 9 depends on how fast the steam is coming to -- 10 MR. CASTLEMAN: The reason I asked the question is,
'11- I do not know of any nucleation experiment where anyone 12 has ever. observed homogenous nucleation at supersaturations 13 less than about three and-a-half and probably more like four 14 or.five, where on pre-existing surfaces it can often happen is at 1.1 or something like that.
18 So, I am just wondering in light of Al's question 17 over there, whether or not you 'are going to have a lot of 18 that down there, condensation. I am just trying to understand . 19 MR. CHEN: I think your question is, it looks 20 like, this distribution looks like a uniform size. I 21
,think according to the Greene e Lane book, they reported at in the homogenous nucleation, it is a kind of mulicular --
23 because the -censity of water vapor is so high. So, the
~
24 probability, if the molecular -- it is higher and they are 8 moving away.
136 1 So, the size depends on the critical ratio, what
- 2. your density is. I think if we pick up ten microns it is
- 3 a little bit bigger, it could be about ten micron or 4 five micron _ size. Clearly, particle size, according to 5 my calculation, isabout five or ten microns in that kind of 6 situation.
7 MR. CASTLEMAN: Let me just rephrase this once more 8 and then I will drop the point. 9 If you had defusion of the water vapor and to condensation at'a supersaturation -- and I am pulling these 11- numbers out of my. head -- something like 1.1 around particles, 12 so that locally you never got over three and-a-half, maybe 13 the homogenous would never happen, even though the classical 14 equation 3 # sort of looking at separated particles or 15 separated homogenous looks like_it would, but maybe the 16 other would remove water so rapidly. 17 MR. CHEN: Actually, the critical ratic we found 18 out, the functional temperature, at room temperature is 19 probably four or five. 30 MR. CASTLEMAN: Right.
, 21 MR. CHEN: Let me go out to 100 degres C, 1.8.
L22- You go up_to 20" degrees,'it is 1.2. El MR. CASTLEMAN: But how valid are those equations? M1 _MR;LCHEN: I.think those numbers, the 1.2-and 1.8 25 I.. picked , up from other experiments. t
.e _ 137 i 1 MR. REYNOLDS : What are those numbers, 1.2 and 1.8? 2 What do they mean?
'3 MR. CHEN: That is the supersaturation ratio, 120 4 percent saturation vapor, I call it 1.2.
5 MR. REYNOLDS: I thought your numbers would have 6- been a: lot higher for real high temperature steam coming 7 out into -- 8 MR. CHEN: According to the book by William Heinz 9 recently published, he reported something, 1.2, 1.8. I 10 . agree , the five, the Green & Lane book. These numbers are 11 s how they are reporting the experiment. 12 MR. KRESS: Tom Kress, Oak Ridge. 13 Before we leave the subject, I think I might shed 14 a little bit more light.on.it. 15 - One, the Warman model with the spread is due to 16 a completely different mechanism, that is water droplets 17-that get put.into the containment due to the initial blow-18 down due to flashing fragmentation breakup of the water 18 that is there. 20
.So, you would expect a bigger size, some sort of a 21 ? spread. Now, as far as the nucleation is concerned, I Et agree with Will that it takes pretty high supersaturation 25
-ratios to produce those. What they would'come up with is a-24
. size based on thetsupersaturation ratio.
E
' But tof answer Al's, question, the :NAUA code already
138 1 has condensation of the existing particles built into it. 2 -So, that part of the growth -- you put in an initial size 3 and that part of the growth would be taken care of and would 4 spread them out, along with the glomeration and other things. 5 The thing that bothers me is that NAUA does not 6 know anything about thermal hydraulics. So, when you 7 homogenously nucleate, you change the supersaturation ratio 8-ana you change the rate, and that had to be. input. 8 So, you' have a real problem. 10 MR. CASTLEMAN: Right, exactly.
- 11 MR..GIESEKE: I,think there is another point of 12 -
clarification that should be made. The do not do a nucleation 13 rate calculation per se. We do a calculation of super-14 saturation ratio after the NAUA code has put onto the existing 15 droplets or particles what it can. 16 So, it does as best as it can'to defuse or. move 17 the water to the existing' particles by defusion. Then it 18 says, " Gee, I could not.get it all there, it is supersaturated 18-by some value," let's say it is three. From the equation -- 80
- anacne can argue what is the proper ~ supersaturation ratio --
21 but if it is above what.you calculate.to be the critical 22 supersaturation' ratio, still, after you put as much as you* 23
- can or what is calculated -by;the gase phase .defusion onto 24 the dropletLit:says, "Okay, now'I can operate on these, on 25
~t his' water vapor thatiis left in'there. It-is still super-
139 1 saturated," and then it condenses it to m'ke new water 2 droplets-which of course then grow and interact by glomeration ,
-3 So, we do not do strictly a nucleation rate 4 calculation.
5 MR. DUNBAR: Ian Dunbar, UKAEA.
.6 Can you clear up for me where exactly your super-7 saturation input.comes'from because I have not been able to 8 find that out in the documentation.
9 I understand MARCH assumes saturation of each 10 time step. / 11 MR. LEE: It has been modified so there is not 12 any' input instruction for this. In fact, I.think you really ul L don't need any specific input rather than maybe, you know, 14 this droplet, initial droplet size. > 15 MR. DUNBAR: MARCH has been modified? 16 MR. LEE: No, the NAUA code . The' amount of steam 17 that comes into the containment is still from the MARCH 18 calculations. 19 LMR. HILLIARD: . I am puzzled here by the effect. If 20 - .the pure water droplet glomerates1with.your mixture over 21 on the right,-then-in my simple mind it seems that the water 3 would'quickly vaporate again.
- 23. MR. LEE: Even without this, it will grow and -1 24 condensation will'take place. Anoth'er_ thing is, they interact 25 ;So, now we . talk about very complicated aerosol dynamics here',
1 4 m v 4 w
140 1 a glomeration, condensation of these two-size -- 2 Now, what-is the overall effect of this mechanism? ~ 3 I'think that -- it is hard to say. I think the particle 4 size will be bigger if this water particularly goes on 5 solid particles. 6 MR HILLIARD: I guess I will just have to think 7 '
'more on that.
8 MR. LEE: Now it is getting complicated, like I 9 -said. Now, if the condition is met -- I mean, a computer 10 calculated all of these mechanisms, it is hard to envision. 11 - We looked at this effect, like I said, early on. Ut We are looking at this size'of the droplet containing 13 solid particles. They tend to grow faster, which means 14 they -- faster than the case where you do not have this 15 mechanism. 16 I am sure you are particularly interested in it 17 because you did some dry tests.and wet tests. 18 MR. RITZMAN: I still don't think'the gentleman-18 - from the UK got his answer. lE MR. LEE: -Okay, this is the modification to the NAUA 111 code. The steam input to.the containment'comes from the 22
. MARCH' calculations and we did not make any modifications to - 23 :-
the MARCH code. 24 ;MR. RITZMAN:
~~
It comes'directly out of MARCH. 25
- MR. LEE: Right.
4 9 9-
r_ _ 141 1 MR. DUNBAR: Ian Dunbar, UKAEA. 2 I understand that what is done -- correct me if 3 I am wrong here -- MARCH has an amount condensed on the 4 walls, and then an amount which it ad hoc says is now in 5 the atmosphere as droplets. 6 Do.you take that back, re-evaporate it and 7 convert that into a supersaturation? g MR. LEE: Let me just move on. g (Laughter) 10 MR. LEE: . The answer was , yes . 11 (Laughter) t2 MR. LEE: Now, the Sequoyah plant, we-lcoked at 13 it, there'are three containment -- two sequences and this 14 containment failure mode, the core melting time is shown 15 here in minutes.- TMLB time. gamma, vessel failures at 16 this time, and containment failure-at the sam'e time. 17 In the case of delta, the containment failure 18 takes place at a later time, and in the case of a TML gamma
~
19 all this event takes' place at the'same time as TMLB prime 20 gamma. 21 There are three separate, distinctive compartments 22- in the containment, the lower containment, ice condenser,
- 23. '.and upper containment. In the-case of a TML sequence'there
.u' is an air return fan available and spray.in the upper a compartment'.
. . - - . , . . . ^ ^ ' -
142 1 To get you an idea on the particle size in terms 2 of mass median, particle diameter, we are talking about 3 four. micron, and for the source from the core concrete 4 interaction we have a particle size smaller than one micron. l l
.5 MR. SILBERBERG: Where do you obtain the 4.37 l 6 MR. LEE:. That is out of calculation, and the 7 .other one is from VANESA.
8 Okay, this11s the key contamination factor of the 9- ice condenser based on total mass . for the TMLB prime. The 10 mechanisms we have in the ice condenser we already covered. 11 We are.looking at a number ranging from 1.6 to 12 5.3. I think those numbers will control the overall mass 13 release from the ice - in the case of.a TMLB prime delta 14 leg, momentarily we have a higher contamination factor.
~
15 MR. .RAHN: Could I ask a question? 16 MR. SILBERBERG:' Sure. 17 MR. RAHN:. Frank Rahn from EPRI. 18 Could you explain a little bit more what you mean 19 by thatidecontamination factor? Is it not true in Sequoyah 20 that -the . atmosphere 1 continues to recycle through the ice condensers?. 22 MR. LEE: Not in the case of a TMLB prime because El' we just assume that the return air fan does not operate. So, 24 the recirculation was not allowed. 25 I. guess you are.not asking this one. The
p _ - , 143 1-n dontamination L factor-is defined; the same way as we did 2 the mass entering'the ice compartment divided by the mass 3 leaving lthe. ice compartment. 4
.In the case of TML we looked at -- I guess I 5
_ought to. talk about that when I talk about TML. 6 This is the plot of the airborne mass as a 7 function of time. I guess you cannot see that, the time is 8
.at the bottom. This is for the upper volume and thelTMLB 8
-psime-gamma-containment fails at this point, about 150 to minutes.
Thisisthe-accumulatedleakl mass. 11 In terms of each species, this is one of the U' leaks out to the environment. This is the -- aerosol, M- cesium iodine,_ cesium' hydroxide and tellurium. 14 Again because of the different. source timing 15 the increase in tellurium is rather ~ substantial compared 16 to those for -cedium iodine, cesium hydroxide, although the 17 overall-magnitude is a lot lower. 18 I summarized all this with a table a little bit
- 18 later.
' This is for the TMLB delta, again,~ upper compartment ,
21; airborne mass as a function of time; containment fails at 22
,a later--time and we are looking at airborne concentration 23
;this way. By.the-. time containment. fails,.there is a sub-M' stantial decrease in the-airborne concentration, primarily 25
/due to sedimentation,. followed by ' particle size growth.
f
-144 1 MR. HILLIARD: Ken, before you take this off, I 2 noticed the. mass, suspended mass is about a hundred times 3- -higher than TMLB prime. Can you explain it?
4 MR. LEE: In'the peak'-- you are talking about this 5 one? 6 MR. HILLIARD: Yes.
'7 MR. LEE: Because in the first case containment 8 . fails, but now it.is intact.
9 MR. SILBERBERG: Excuse me, Ken. How many hours 10 - have'you come to? 11 MR. LEE: This is about 150, and this is about 2 -550 minutes -- yes, 150-and 550 minutes. 13 ' MR. SILBERBERG: Seven hours. 14 MR. LEE: That is from the~ start of the accident. 15 MR. SILBERBERG: You have, it looks'like almost 16 two orders of magnitude decrease; is'that right? 17 MR. LEE: .Right. I.will come to that point. You 18 - are talking about release -- < N -MR. SILBERBERG:.'No, it ceems.like a fairly large 20: . decrease and I' don't know, your concentration level, the
- 21 fact that you are getting less out of the primary system, it H' would seem to me'you have:a much diluter situation or 23
--dilute situation in the containment,;unless there 'is something
'M - ~
about the condensation _that Ilam missing, that'may be-25 - contributing to,that.
7. 145 1 MR. LEE: This is after.the ice condenser. i
.2- MR. . SILBERBERG: All'right, continue.
3 MR. LEE: .The mass going to the containment is about ,
-4 -I think-Mike told you, about 110 kg during the release 5 and then,.I think, about 270 kg. mass -- concrete interaction.
6 Particle size defusion is dependent on time. 7 Early.on,.we see this typical bi-modal disfusion. s 8- This figure shows.the location,of the disfusion 9-
. of each species af ter the 'ac'cident- is pretty much completed, 10 which means you'really don't have a whole lot of particles
. 11 suspended in containment.
12 This is_ gamma and this is delta, cesium iodine. 13 Af ter the accident: is completed, you go around to each
' 14 . compartment and you will find this fraction of core 15 inventory. This is the number Mike Kuhlman gave this 16 morning.
17 - This is in terms of fraction-of a core inventory, 18 so that can be slightly different than what.the number might
' 19 be because his factor was based on.what goesLinto the core, 20 ~
for example, we are talking about 26 percent.. 21- -You notice thst' about 14 percent _oficesium iodine 2t out'_of;25 percent'gets; collected in the ice bed and a 1 23-comparatively.small~ fraction windsEup in the upper containment . 24 TWe are lookingEat-a number:like..two percent release. 25 MR.-FULLER: -Ed Fuller. ~I am a little confused. a D
+ - m y
- N* v m
/
146 1 ' You-said these were fractions of core inventory, but yet, 2 you label them cesium iodine and cesium hydroxide. 3 .- How are we to decide what the fraction of the 4
, iodine original inventory is or the cesium original inventory
~3 is.when;you combine them like that?
6. MR. LEE: .We just simply took the inventory of 7 cesium iodine and iodide, that was about 13 kg, something 8
.like that, and get'the. cesium to come out with the. total 8
inventory ~ for cesium iodine. - The' rest of the cesium goes
' 10 into it. You - j ust add that'. We have the oxygen and II-lhy'drogen combined.
U So, w'e are talking about-120 kg. This is about 13 13 kg.and 26 kg, something like that.
, 14 MR. SILBERBERG: Well,.what he is-saying is --
15 MR. LEE:- Yes. 16 4 MR. SILBERBERG: What'he is saying is, forgetting 17 - acout the species, that'. the next"1eveli of reduction of the 18 - result would be, just.giving the elemental cesium, the 19 , iodine and so'forth -- MR. LEE: Correct. 21 MR.~ SILBERBERG: Subtract it.
' 22 :
MR. LEE: '.In the case - of iodine I can 'tell you
"# that'thisEfraction would change.
M -
~MR. BUTLAND: Butland, UKANA.
25
- I wanted to ask a sliOhtly more. general question.
2 w. t
,s ; .t .g
l 147 l 1 I am not clear how, in these calculations, you 2 take account of movement'of these radioactive species from 3 'the-containment into the environment in the event of a 4' failure because the containment does not just disappear, there 5- obviously is a. break or a hole, or something. 6: How is all that being done?
.7 MR. LEE: I am not sure that I understand your 8 question. Are you talking about this one? This is~after 9 -tne accident is completed, which means --
10 MR. BUTLAND: Well, you have a column headed 11 " environment," which I assume means -- does that mean outside 12 the' containment? 13 MR. LEE: Right. 14 MR'. BUTLAND: How has it gotten from the inside 15 of the containment to the outside? 16 MR. LEE: That is-the accumulated leak mass outside-17 .the containment. 18 MR. BUTLAND: So, how is-the leakage being modeled,
' 19 how are'you assuming -- what form does-the leak ~take? How 20 . sensitive -lis all of this to the form of the containment 21 Jbreak?
Ju- MR. LEE: It is directly influenced by the 23 - containment; failure time which you just see right here. - This 24 is~early: containment failure and thistis late failure.
-:B MR. SILBERBERG: ~ No, I'think --
- r. L
y
~.h t
148 1 MR. LEE: Now, on the dependence leak rate, again -- 2 MR. STLBERBERG: Excuse me, I think that Terry is 3- _asking a different question about, are you taking into 4
. account the specifics of the leakage and the leakage path 5- in the context of as you leak?
6L
-I ' think the. answer is , no, but maybe --
i 7 MR. LEE: You are talking about attenuation? l 8 MR.cSILBERBERG: 'Yes, that is what he means. 9 MR. UEE: No, we-don't consider any, we don't l' 10 give any credit on_ attenuation.
-.11 MR. SILBERBERG: Let me suggest that the assumption n' here at this stage is a failure'of some large size and t3 material going directly into the environment. .
14 - As our containment' failure. mode study squeezes 15 down.on the issue of what is failing -- and maybe it is 16 not catastrophic failure-but seal failure -- when one starts 17 to obtain a better definition of failure' modes, then I think 18 one will.want-to come back and take another look at how N important'those modes might be.to alter this. 20 But in itself, you know, it-is not a simple and
- 21 . straight-forward situation. It is again in terms of Et. experimental evidence and what'have you.
L 25 So, Jthe next level, I think once one can get a
~
.M. better -definition or characteriaation _of the leak paths, 25 che would try to look at that.
r: 149 'i 1 MR. KASTENBERG: Mel, since you won't be around to
~
2- explain to: each reader what that means, why don 't they just 3
-label that last column as the aount that is available to leak 4 -to the environment, and then people will understand what 5 it-means.-
6 MR. CASTLEMAN: The amount is leaked and there 7 is no attenuation. 8 MR. SILBERBERG: . Put an asterisk. 9 MR. GIESEKE: It is something different than what lo s is available to leak.: It is the integrated amount. You
~ 11 : take.-the leak rate and all that goes with time. Pretty soon,
- u. there.is a hole in the containment and-it may take some time 2 for'the containment to blow down and pressure comes down over 14 .some period of time.
M During that time,-there is-a flow and pretty soon 16 - it gets down' to one atmosphere and there is no flow. But 17 over all.theJdiffering flow r&tes.it.is integrated..with
?.8 time. But what is airborne in there.is taken out and is 19 integrated.over that whole time.
2- So, it is'really not what is available to. leak, 21 . which is at any point of. time.a different value because the n .. concentration -is dropping of f even while it is leaking out, even after the hole is formed there.
~
23 ---
.So, it is' constantly-24 - changing.
25
- But it is the integrated amount that.is actually
e -- - - - 150 1 leaked, . assuming no attenuation as it passes through the 2~ hole. s 3 MR. SILBERBERG: Fine. 'Are you saying.that needs 4 ito be-labelled or explained, or' flagged? 5 MR.,KASTZNBERG: ' Just a question also'while I am 6 up. .What is the effect o'f- the - fogging model that you E described before on this distribution?- 8 MR.-LEE: Was'it activated, do you remember? 8 Okay, that mechanism was activated in this 8 particular sequence at an early stage of this release. II
'MR. CHEN: Chen from Battelle.
12 - I think in the sequence, TMLB prime gamma we
.13 do-have'a homogenous' nucleation happen right after a --
I' failure because of steam coming. The steam . rate is so 15 high, it is exceeding.the saturation ratio, the_ critical 16 ' saturation ratio and that of droplets forming at a time, particular time. MR. CASTLEMAN: I had a related question. Does
# it. make. a dif ference- as - to what ends' up 'in tihe ice bed if
" thisistuff-has formed a fog and then can channel through.
. 21
- the. ice ' differently than if: the material' is as a vapor and 22-
-can be transported to'the ice and condense?-
MR.: LEE:- Right. MR. CHEN: I think we calculated how much particles _ 25 ~
.available go through the ice bed. In the ice bed we--have u
151 11-calculations from MARCH of how much. steam is coming in. 2 Basically, from the lower volumes through the ice bed we are 3 concerned how much the airborne particle -- so we can 4 tellLhow much.the particle goes through there. 5 But I think the particle size is on the wet 6 particle. 7 MR. KELLY: Kelly from the University of Virginia. 8-
- r think you addressed this question before but
~
8 I did'not understand your answer.. How do you calculate, or 10 how do.you determine from those numbers what fraction of
. 11 .
.the-cesium inventory escapes?-
U MR. LEE: You are. talking about this number? If U' . you add them up,.it is one. 14 MR. KELLY: What fraction of the cesium core 15 ' inventory escapes to the environment? 16 MR. LEE: This fraction right there.
. 17 MR. KELLY: Which?'
- 18 MR. LEE: The first. row here because this is
-U-cesium: iodine and we just assume that. cesium. iodine inventory -
20 -
.MR. . KELLY:- No, I am talking.about total cesium, Ken
) 21 How much'of all the cesium --
[1R. . LEE: . Total cesium, you.have to calculate that
.23
'from'the: cesium' iodine, cesium hydroxide.
24 MR.. KELLY: How do you.do that? 25 . 1MR.' LEE:- Just-look at'the-molecular weight. l ^ v v 9 $
152 1 MRE SILBERBERG: You have done this before.
- 2. MR. LEE: .Yes.
.3 MR. SILBERBERG: You have cesium and iodine previous ,
4 .MR. LEE: I think Jim Gieseke will summarize the 5 whole thing'in his bottom table. There we separate this 6 ' cesium from-iodine. But you just look at the molecular 7 weight ~of cesium. 8 MR. SILBERBERG: He knows how to do it. 9 MR. LEE: And do that same thing and get that 10 cesium. 11 - MR. SILBERBERG: Jim, do you go down, your summary U- will include for each of the elements? 13 MR. GIESEKE: Right. 14 MR. SILBERBERG: All right. 15 MR. HILLIARD: Since that last column is very 16 dependent on the leak rate,-would you tell us how you i 17 calculated.it? Is it a constant' fraction throughout the 18 ' - accident'or does it-vary with-pressure? 19 - MR. LEE: That is all calculated by MARCH code, 20 -dependent on pressure,.the containment. Thisbs not a ' 21 : constant leak rate in( any means. That is calculated by 22- .some hydraulic-code.
#' MR. HILLIARD:- Is it going to be described in '
24 .your report how you do'that? I mean the actual-values. 25 MR.. LEE: Actual-values, I think we listed.them.
153
^l n MR. CYBULSKIS: In the~ MARCH calculations, when 2
we reach containment-failure pressure, however that is
.3-defined, the MARCH code requires.as inputa hole size and 4-the coefficient. They'are essentially interchangeable.
15I In_the particular< calculations that are used 6.
.here -- 'and -I think that was described in t he report --
7 we used'a hole-size equivalent to seven square feet. 8-The attempt here.is to_say that you.have a hole 8 size which-leads to rel'atively rapid containment depressuri .
- 10 zation. In the~exercizes that were done over the years, til mak'ing the hole much bigger will not have significant 12 i effect on the release.
13 You could make the hole: smaller and it may have 14 some effect on the release. However, if you make the hole is size eco small, it will not arrest-the pressure rise and 16
-the: pressure will keep: going up.
17
.What we areEtrying to do is model-a representative 18 :
hole size. -I do not know that there is a. sacred definition , 18 fori.it, but we 'have'used_a--fairly significant nole. size and then we. proceed to calculate the time-dependent. leakage
- 21,
.through that-hole size, taking'into account what-is going inte
, ;the containment, what is: going out.
Y In the reports that you have.seen so far,-we didL 24
-not have._the opportunity;tol include whatlis, I ._be lieve ,
26~ [ supposed'to be numbered Tabl'e'6.5, which will present the i-
%-J. p- y y =p 9n 4 y *-'
r. 154 1-Ileak. rates as a function of' time in some time intervals 2 to-give you an idea of how much of the leakage takes place 3 .when. 'That will be: included in.the final report. 4 MR. HILLIARD: Thank you. 5 MR. ; SILBERBERG : Bob?
~
6- MR.-RITZMAN: A quick question. Ritzman. 7 -For.the gamma-up there, the tellurium, if you add 8 those numbers across there~they come up to 32.5 percent, I 9 would' assume'all the rest of the tellurium is in the debris, 10 core debris; isdLat right? 11 . MR. LEE:~ Right. it is not released. I see in 12 this particular one it would add up to only one; you are 13 right. 14 ' MR. LEVY: I think the question has'been answered, 15 what-happened if you' turned off this homogenous nucleation 16 and this fog, what would happen to the' environment? 17 MR. LEE: Well, that-is another sensitivity
- 18 analysis. I think:though what would happen is,-particle 2- growth'will be a little bit retarded. So, I would expect 20 .that;we-might have a slightly higher! release in the case'of 21 gamma, in the case of-acdelta there won't be any effect 22 1 because the particles will grow eventially-and -- out.
~ 28 . 1So,sI' don't see any effect-in'the case of a late 24: containment failure.
25 ' MR . LEVY: Do-you.know what the magnitude is for
r - . _ . 155 1 l gamma case, what'are weItalking about, a factor of two, or 2' 1l.2, or what? 3 MR. LEE: No, I don't see that. We are probably 4 talking about less than. ten percent. 5 MR.' LEVY: It makes you wonder why you go for 6 'all that complication. 7 MR.-SILBERBERG: _ Excuse me --- 8 MR -LEE: .I agree with your philosophy why we 9 snould go through this complicated -- the effort itself is 10 not complicated,_it is a simple thing *,d just a matter of 11 how you look at it.- 12 MR. SILBERBERG: Let me add something ta) that-13 sort of a vicious circle. 14 'You say that the fog formations were ten. percent;
-15 is that right? You.think-it would be worth about ten percent
- 16 in the release?
17 ' MR. LEE: Yes. 15: . MR . SILBERBERG:. By not including it. 19 MR. LEE: Let's go back and take a look at it. EL We are not' talking;about facts here. 21 MR.'SILBERBERG: The; point.is that this effect 22- was put forth in one of the earlier meetings as something L
-where you.would say, " Hey, you have tx) look at.the fog
- 25 24 . formation. Are you taking it into~ account? What is it El ; worth?"
I
156 1 So he has gone back and looked at it. Now he
. n.
2- ' is saying, "Having looked at it, it does not appear to be worth that much." [ 3 - 4 So, it is not the. cart' backwards. We went down
- 5. this' path because it was .put forth as something that they
[ Es believed that was requested to be looked at. 7 If I am'not stating it' accurately, somebody 8 correct me. ! 39 .MR.. LEVY: We.do not seem to know what impact 10 - it has. 11 MR. LEE: The condensation of a water vapor and 12 particle.is there. 13 MR. WARMAN.; : Mel, you~ asked'if you were not stati5g 14 : it? correctly, "please4 correct me." I would like to correct 15 you.
.16 Ed Warman from Stone & Webster.
17 Having been the one who proposed that this be is _ looked at'and prior to proposing it be looked-at having 19 done many-months analysis.on-this subject,.believe me, it-
- se . is wor'th at 'least a factor of two in reduction in leakage
~
. 21 to the environment.
- Et It is not a ten-percent. kind of effect, we would
~ '
23 have discarded-itflong ago if.that is-the kind of effect it 24 :was.. The analysis-just desribed here'is not the analysis 25 we proposed. .The information.forfthe analysis we performed
,Y
{. , . . . . . . . '~
157 1 -- has been made-available to the staff. We can discuss it
- 2 later, Lit will take a while to explain why it is a factor 3~ of a two-kind of effect. Factors of two should not be 4
ignored because if you get a few of them in series you have 15 quite a reduction on-your hands. 6 lMR. SILBERBERG: I do recall a factor of two in t
;7 .the' previous presentation that was made.
~
- 8- MR. LEE: We _ have to remember that the dominating 9 factor in this sequence is' ice condenser. You already 10 trapped 75 percent-in the case of iodine, cedium iodine in 11 the primary system and 25 goes then into the containment, 12 and 50 percent of the material gets trapped in the ice bed.
, 13-MR. ZUMWALT: I just have a simple _ question. It
- 14 -probably goes back to Kuhlman. But all this tellerium in is the core debris, from a simple-minded point of view, would 16 you say that is being held by the zirconium that is remaining 17
.in the debris? -It is not'all oxidized, I understand.
18 MR.-LEE: Dana.will answer it. 19 MR. ZUMWALT: Does anybody have it? EF MR. POWERS . ' It'is similar but it is not 21 zirconium, it is held by' stainless steel. E' What happens.in these calculations, especially in 23 _th'ese~1atter calculations-is, the temperature of the debris 24 - reallf. drops down very. rapidly. You have diluded the 2 metallic base of the1 melt so much, there is just not much
m:: - 158 1 -~- enemical. potential for the tellurium to vaporize; it is 2 ' trapped in the steel. ~ The circonium has all been burned out 3 ~ by that' time. 4 :MR. LEE: Moving on to TML' gamma -- 5 MR. REYNOLDS: Can I ask'a question before you 6- leave TMLB prime. Is'that what you are on? 7 MR. LEE: Yes.
'8 MR. REYNOLDS: (Let me.ask a rather general question.
8 You have 75 percent-of.your iodine, cesium iodine remains
' 10 in the reactor coolant system, and for the Surry report only 11 -
20 percent remained,180 percent got- out of the primary U system.and 20 precent. remained in the primary system. 13 ~ Why is there such a dramatic difference between 14 these two numbers? Can you' focus on.the central, main 15 ' issue that is causing such a-dramatic change between this
-16 and the primary system at Sequoyah and Surry?
17~ MR. SILBERBERG: Let me j ust . say, what they 'have 18 - to do is compare this calculation with the Surry recalculation , 19 I~think,-to an'swer the question properly.
'8 MR. REYNOLDS: Which will come up later in this 21 meeting?
22 ' MR. SILBERBERG: Which will.come up.
'E MR. REYNOLDS: So, we wait for that.
24 MR. SILBERBERG: Yes, that is my concern, too. 25 = We ought'to compare those and see how close they-are. If-6
159 1
.they are the same, fine. One ought to explain the difference i
2 back/from the first.Surry calculation. If they are 3 different', that needs to be explained. 4 MR.-LEE:- Okay, TML gamma. Again, the containment 5 failure is-at 150 minutes. The calculated decontamination 6 factor for the' ice condenser is ranging from 2 to about 15.
~
7- MR. DUNBAR: Dunbar, UKAEA. 8
.Those decontamination factors, are they time 8 dependent or: cumulative? I just saw one that went up and
' 10 '~
down. 11 MR. SILBERBERG: That is just time dependent. 12-MR. LEE: That is time dependent, that is not 18 cumulative at all. 'So, the overall decontamination factor 14
- for the material, you have to multiply the DF by the source
.15-rate.at:that time.
16 Tnis s the. airborne mass as a function'of time. 17
- The upper compartment -- remember that there are air return 18 fans operatiag until the containment'fatis. Remember that.
19 the NAUA code is a single volume model. 1 thinkIFrank mentioned that in.this case there 21-listrecirc.lation flow. So, what we have decided was to i 22 ~ use thisI NAUA code sequentially, covering the lower compartment , 23
; ice bed,.and upper compartment, an'd then incorporate that 24 recirculation flow; come back to the lower compartment, ice
_ bed,jupper-compartment. That is the way we took care of this-
160 It return air..
- 2 ; 'It turns out;that'the ai'rborne concentration in
, .'3' 'the.uppericompartmentLis.very_ low compared to.that in the 4 ' lower compartment, so that~the factor-of-recirculation 51 flow inythis particular 1
sequence was,not great. 6 LSo, we just. compared the first. calculation for 7 - the lower compartment with the second calculation for the 8; lower? compartment after taken care of this recirculation flow. 9- In fact,'it.was minimal. So,Jwe just stopped at that point. 10 - MR. DUNBAR:- Dunbar, UKAEA. 11 Which compartment leaks when your containment ut- fails? It is uppert I see.. Is that because you think;that . un is'the most likely one? 14 < MR. LEE: That is.what Pete has to answer. The , 15 leak takes place in the upper : compartment. 16 Particle size'dif fusion as a function of _ time, which 17 are' fairly dynamic. 'After' containment failed, we have this 18 kind of particle size' dif fdsion ~and can. change as time goes 19 o n .- Locational disfusion of each species-looks like
~
~ 2L
~
. 21 ' .this .-. It turns out that this ndmber'- . this: is .TML ~ gamma --
- 2. and,this number is HB-10.1 The numbers' for TMLB gamma: and
- 23 thoseTfor TMLBidelta late.
4
. ML The reason'is'obviously due'to containment sprays 25 in theEuppericompartment. .This number 1 remains the'same
'-- 4
, , -c. .,
, * . , , -_ -,e r , . - = , ,
u s 161 t
-l' because the primary system'does'not know the difference 2 between TMLB prime and,TML.;
3;
.I'think Jim-Gieseke will summarize all the results 4- in his' final table.. s N'
- 5. MR. GIESEKE: As I menti ^oned at the beginning, s 6 this is.-as far as :we have gone with the containment
, J
'7 calculations for Sequoyah.
8- Next week we-will meet here again, we will have-9 _the S2HF'for you.
- 10 (Laughter) s 11 31R. GIESEKE: The question came up about the L2 -iodine cesium and so on. This is'in( sequence what is released
~
from the primary system to the environment. I think you have 14 lalready seen, basically, these numbers. 15 There is no comparison hereiwith'the WASH 1400 18 kinds o'f: numbers that we showed'before because it is not 17 a situation that is really comparable. 18 As you see here, we saw-before the cesium is 19 - fairly effectively held;up in the primary system. Total 2 releases-are quite low. 21 For;. tellurium I will put this up and point out that Et the release here is very high for ' tellurium in contradiction
-# .to what you saw as far as:the percentage of what was
-M released,'from the' fuel th'at made its way'into the containment.
25 ' This includes a release that goes down with the melt. So,
, , _ . . e+
k
~
^
162 q
~1c ~ tnat-is the difference.
2 I'could'have; skipped all these slides but I p ll - wanted l to point this out because you are going to get this 4 handout and:will.ask why.is.that number like it is. 25 Again, the percent releases or fraction releases J 6 are fairly _ low, the highest 'one there being a total of 7- .one. percent. 8. MR. ROWE: When you say " fairly low," what do you 9- mean? 10 MR.'GIESEKE: Well,' compared with the kinds of 11 nhmbers:we saw for Surry and the usual conception of a PWR 12 release, these are way down. I mean, a couple of orders of
- 13 magnitude.
14 ml. ROWE:- A couple of orders of magnitude? Hi MR. GIESEKE: Of course, it depends.on the 16 sequences that_you look at, and there are.aL1 sorts of 17 qualifiers you want ' to' put in. 18 MR._ROWE: I do not have these ' numbers in my head, - 19 I am'trying to get a little perspective as to what these 2 mean.. 21 MR..GIESEKE: 'Okay. ~Well,~if you look at the 22 tables from previous meetings -- El MR. SILBERBERG: Fractions 1in' tenths as opposed
~
24 ~ to hundredths.
~
25 MR. GIESEKE: ^Instead of having 70 percent released t "
.b 163
-1 in:the environment, we have one percent. _That is a big 2
~
difference, I would'.say.
~
3 With that, I_wovid like to stop. talking about 4
?Sequoyah and we will'go on to some other, bigger and better 5- -things._ Back'to you, Mel.
6 MR. SILBERBERG: Let's take a ten-minute break, 7- and resume promptly.:
-81 (Whereupon,-at'3 p.m. a. ten-minute recess was 8- taken.)
10
.MR. SILBERBERG: Let's resume the meeting, please.
11 LPlease, be seated. 12-We are going to pass Dana Powers.to tomorrow 13 ' because one of the topics he is supposed to cover tomorrow 14 ~ Lhe'is n'ot going to cover at all anyway. So, this kind of 2 fits in-its place. 16 VOICE: Which one is that? 17 MR. SILBERBERG: -The one on steam explosion source 18
' terms. He will.give you a status report on that tomorrow, 19 but he wont' say very much.
E' ITom? 21 MR. BGESS: Mel: caught me by surprise a little bit.
~ ~
E Thislis just to let=you know who I am. l Ek 'I plan to,give~you a review of the status, a 80 s L status review fof - the / status- of validation- -- 25 _. :(Laughter) c: ' s i' . ,, m ..
164
'l
- MR. KRESS: You know what I am going to do -- of 2
the source term codes. Our review'of the. source term codes, 3 their status and validation. 4' I have some handouts here, Mel, if you want to
. 5 give.them outLlater. I am not going to dwell on this slide, 6
lthat.is just to.let you know there are a lot of people 7 helping'me and working with us on'this subject at different 8- institutions. 9 We are looking at the set of codes that we have
~
10 been talking about. I am not going;to say much about the 11 suppression. pools and the ice condensers because we started 12 late'on those and they will actually constitute a supplement 13 to the study and.will come in as a supplement to this report. 14 ' Just a quick statement of the status as of today.
. 15 We'.did produce. initial drafts that were-sent out to the 16 -
committee, they were' partially complete and had missing 17 parts in them, and hand-written parts of them. 18
. We reviewed those as a group together -- not you 19
-people but the' committee I have -- and'made a lot of comments.
20 Final: drafts were prepared'and are being prepared. I do 21 not have all those in quite yet. 22 ~' I have been wading through-these initial drafts 8~ Land.-final drafts and my-intention is to make a personal 24 , - opinion summary =. appraisal of my own that I gleaned out of 25 the various-reports and out of my own perspective. That is s _y --
165 1 in preparation, I have not completed it yet. The intention 2' is to have a first draft of the full report out by the end 3~
'of October that I can:present to you people, that will be 4 in. good enough' shape at least by then. I think it will be is essentially' complete. There may still be some hand-drawn
-6 figures;in it or:something.
- 7. '
I won't dwell on these either, but just to remind
.8 you that this is not a sensitivity uncertainty analysis, and
- 9. the: intention is actually 'to do this list of things which I 10 discussed before.
~ 11 ~ .
From my own viewpoint, after looking at the
. 12
. graphs,'they do a relatively good' job. I-think they were 13 a little weak in.what'may be the most important area, and 14 -
.that is really digging out the data bases behind all the is ~ inputs and'the models, and finding outJhow good that data 16 is. I think we ended'up a little weak there, and it was as 17L a result on the constraints on the time.
-18~
That really is a ' tough job, a lot.of digging 18 - tnrough literature. that has to be done and evaluating data. 20 So,:I think if it isEweak anywhere,-it is. weak there.
-21 Today I thought maybe I would give you a little bit 1 st aof a flavor of what~'is-in_these reports,-and mostly concen-Zl-
;trating in this area which I think is interesting, the 24
- subjective opinion of these people, and.perhaps in F.his area 25 ~
talking:about some.of.the~1 imitations that might be in these n
-g i
i 166 , 1: -codes.- 2- You;have'.to be~ ambidextrous to do this. 13; First a little provocation. This is my own opinion,
~4' and of course you can't really write these things to 5 contributors-to uncertainty without doing a sensitivity 6 . uncertainty analysis, which is the Sandia part of the job.
7 But just based on my perspective and intuition 8 'and reading these reports, I want to give.you a guess as to 3 how this will turn out,Jand you can grade me on this guess 110 ;when you get the. final answers in.
.11 - _Maybe:this needh a.little qualifying. I don't
-12 knowlif it_does or not, but the rationale behind my ranking 13 is,: MARCH, of course, is relied on so heavily.for so many 14: parts of this thing, it drives the thermal hydraulics.- It
- 15 drives .;the fuel temperatures . It drives the flows through 16 the primary system. It gives:.you the containment loadings.
17 As'farias..I am concerned, the single most important 18 element of source terms is containment failure timing which 19 ~ . we.areLreally not addressing in-this study. But if you 20 ;already use' MARCH for;that,7why, it would be put up here 21 -as first!rankingfall by-itself.
- m
.n 'So,-I think as far as level of validation and 23 ' contribution to uncertainties, you would find MARCH on the M top.
25i It is also; intuitive to me that the amount .that is
)
167
-1 . going to finally end up getting cut probably will have a 2- direct relationship to the, natural release from the fuel in 13 the first-place. That-is partly why I have these two up 4 as high as theyjare, and partly because level of validation 5 and lik' elihoodL of. these things having large uncertainties in
.6 them. So, that is-the. reason for their ranking.
7 You.will_probably come to a perspective as to how a these things behave in containment is very important, what 9 are these natural removal processes and how does the steam-10_ . affect it, and the fog, et~ cetera, so you might think NAUA 11: 1will'not be up'here in=that. region. But it-is my opinion
- u. ,thatithe level of1 validation of NAUA and the confidence one 13f can place _to;the aerosol models is high enough that actually 14 LI moved it down here_and placed the primary: system transport is 'before them.
16 You saw some:of the data today that shows the
. 17 - primary system'can be very important'in a lot of the J18 ' sequences. But that importance is not'due.particularly to 19 . TRAP. melt, it is due to what MARCH does to the thermal
'm ~ hydraulics and to.the-flow rates and the-terperatures.
21' -Finally, I do~not think I will get'any argument-Et with'that, I do:not believe origin is likely_to have much 23 _ 'of an uncertainty. j N So,'I am going to give you one slide-for code, keep l 25 . it short,'on my interpretation of'what we are: coming out with
' ~
l 1
168 1 in-this set of validation documents. 2 First off,- you are interested -in whether they are 3 validated. I think-MARCH 2 is essentially an unvalidated "4 coce for the most'part.
.- 5 A lot of the internal models use relatively sound 6 approaches, with some exceptions. There are data behind
'7 <a lot of-the separate effects models, such as the steam cire 8 -reaction; some of the heat transfer coefficients and
-9 ~ condensation coefficients,-and debris bed heat transfer
' 10 - modeling..
' 11 : But there are some areas where the modeling has ut. a big influence and I think it does not-have a high level 13 o f : validation . The first-Lone on my list, of course, is the 14 core melt' progression and slumping models. I think that has 15 ~ .A tremendous influence on'the source terms, and we have 16 seen it;today.
17 You saw how a relatively minor change-in the 18 slumping models -- well, it:was a later change -- but what 19 it did to the source terms.- Frankly,.'I.th' ink that is going
-20 .to be the area that really has the largest effect.on the it1 ' uncertainty. ranges in the whole_ study.
22 MARCH uses " enter" for. core-concrete interations.
- LIeput that one up -there because of my perspective that 1M basically containment failure has a really. overriding 25 ! influence on the: source terms.- If you were'to:use MARCH for 9
9
169
- 1. Econtainment failure, you would be using " enter" and-not
~
21 'CORCON. .I'th' ink it'is not as well validated or not as good i 3
'as:CORCON for determining. containment loadings.
3-4. By.the same token,-I think there are some questions
.5 about theEsteam spikes lthat one gets due to reactor cavity
.8 and.mostly.related to the size ~and distribution of the 7 .particulates.;that you put in there. 'I think that has an 8' :effectron containment1 1oadings, not particularly on' fission 9 = product-transport.
10 I think E I might add, I had it on'there and took it 11 ' off, and put it back'on,. hydrogen burning. I think I~will 12 : put ! it back on,-probably. 13 The MARCHThas some other problems relative to BWRs 14L and-I do not want to get-into the details on those. But it 15- all has to do with how you use MARCH'to determine the 16 temperatures-for the core,-the movements of the material, 17 and the fact that you do not have geometry changes, you 18 .do not have detailed' mechanistic models for the melt 19 ; progression; you do not have area' changes.
#- So, I will cg) on to CORCON. The thing about it is, 21 : ' it[isian evolving code, . it is not a thing that here and now
- H ~. is :still growing. I think it uses some-prettyfgood physics 2 internally. :I=do.not know ifithey are appropriate physics.
24 The' physics it-uses are' good, but I do not'know how appropriate
~ ~
s 25 ? they are. s 4
170
~
1 It has somellevel of validation, especially for 2~
'the thermal. hydraulic models internally in measuring the heat 3 --
transfer -coefficients due to bubbling up through the layers 4' and heatatransfer off of.the. top surfaces. So, it does have 5 some level of validation that is pretty good. It still needs 6 overall integral validation, and there are some areas we 1 7 worry about there. Some:of them are being fixed and being
.8: ;1ooked at, like it is not applicable when the debris
'9 : solidifies.
;10 E There are internal correlations that-affect the
. 11 heat transfer. - Really, what you worry about CORCON is how 12 muchL heat goes -into theconcrete compared to how much heat
- 13 ' 'goes out the top. It is those: internal heat. transfer 14 '
correlations that one worries about. These' things do affect U5
.them and.they are sources of uncertainty. It does not yet
- 16 model a water layer on the top of the debris.
That la s model that is being added and may even 18 be added'by now. I-do not know what effect it will have. 09 ' VANESA also sought to use. pretty good internal 20 .pnysics:andEchemistry.- The results have a basis in 21-experimental data, the large stainless steel drops cut -
.Et the -different concretes, out of which an empirical model had 25 been developed earlier.
M- .VANESA gives. pretty.goodLcomparisons with those 25 experimental"datarif you subtract . outathe. thermal hydraulics.- y 4 -
- r &
171 I'- If you :inpt't the thermal' hydraulics that you actually measured
.q 2'
Land lnot rely on CORCON ' to calculate it for you, then it 3 compares pretty well for the species that are given off and 4' the rates at which. aerosols and fission products should 5~ 1 give-off. So, it does have a basis, a validation basis. 6 -
- Some'of the needs for additional validation need !
l' to be compared withi the large-scale integral test using
-8 natural! melts rather than stainless steel, and they need to
~8 ne large scale.
10 '
-They have this~ bubble bursting release mechanism a
11- 'As,I' mentioned before-the water
. which needs validating.
12 pool. There is some question about, as you bubble up through
'13 VANESA you pick up fission products _.and other species and 14
-the' rate at which you pick:them'up partially depends on the 15 activity coefficients for these things in the melt itself, 16 in the different phases.
U I think at the moment they_.use-one for the activity coefficients,-and there is some1 uncertainty in those. 19 You heard mention also what effect would a crust 8F p on the top,:for example,-have on the release of fission 21 ' _ products. EIt may serve as filtration._ Maybe it. won't, but that is an area of uncertainty, also.
"' LCORSOR - the release- rate modeling for inside the 24
-core. -It is strictly an1 empirical model.- It takes'the 25 '
aimited: amount of data;we have,Jselected data, selected'on
~
n -- r , , ,,
^ < , ., ,
.; '+
, 3 ,
;g: -
+w : n :
s p 'g _, ^ l
; s ' -
~~
~ _ m ,
172
- ,%: f ,' e m ~ g e.c :-
_ m it vths bas'islofdhow applicableIit'.is under the conditions, and
. ,, s J ,..t 3 .. -5
- , ~ i ;2lf 1justifits?cufvesithrough.them.a You have-to admit that m- '3 C
, 1:tne.idata[sdat'tir;:- for Ethefrelease. coef ficients' themselves l
- E; " .
.n g, ,; ,
~
..0; is sprettyfgoodi It 'is:- better forf some .of_ the fission' products j '
[8$ :thanjforlothers, and it-is a-model where'the.only parameter a '[r ' a + - 8' linJit)isitemperature..
~~
- s 7- ,
~ sol intuitivelyfyou know the release rates.ought-A m8 -
i to.- betf a[ function? of l others things; like - the L geometry , surface-19 farea's',[ flow! conditions thatqgo_by; surface.to volume ratios-
. .~
10
; andieven 1 activity /coef ficients .: and ' things .. like that . -
"11" 3o, fit is a simple.model.
The question is, chow sis) much:fdifference;would Lit make ;if ryoulhad1alIot- better 13 1modeling.: LI thinkiit-is'mostly driven by: thel:-temperatures o
- 14?
th'at!MARCHstells you:that you-reach;. fit has a_. bigger
'15 influence > thanf thefmodeling ;itse'1f. -
;Itflooks pretty (good for' calculating the_more I
hvolatil~e fission products $because;they essentially get
' I8 '
releasedLorfalargel fraction'ofthem,.anyway.. 418 ;Where it.may. fall-.down.is_some.oflthe lower n L
~
..A fvolatilesLand:betteridata,1I?think, is'neededifor some of
]
21! ithes,e elements. I~myself-d'o:>notfthink.it is really_an w' # isti
~
lapproprfat'e.. type of mode 1Sfor: structural materials. 2 s
, .123X '
, s So ,[I i think : you ; need 'some thing 1 dif f erent for 28' Estructural mateiials-and[co'ntrolfrods. You need some sort
- SF;:
. _ 'of[af hdat ; tra,nsfer ivaporization l type - model . with areal changes -
, . 4 b 4 .(
y -
,. I' ~
. Er , fib - , ,. , ._ . ~
~
c.. I, h , ' A 'Y y , ,
,,?r- " -
- 4 s- -
s
,m 4 , ,. ,, s <
- as +
r ,=, , 173 l' andLso'forth.. So, I think additional modeling would be 2- ' nice for that bulk of the aerosol ~that comes out. And, of 3 course, i,t does not account for -- chances or even other 4 release mechanisms such as-when you-drop the stuff in the
-5' ' bottom of the vessel,can you get releases'then, and there s' may be~other.' types.of release mechanisms.
7 MERGE converts the MARCH pressures and flows 8 and' gas-temperatures'into thermol hydraulics, the primary 9 system.: It'is essentially an unvalidated code. It does 10 have internal heat transfer correlations that are just 11 like all~those codes. You go to - the literature and find 12 ouc,.here--is a heat-transfer coefficient for-turbine flow
/ 13 ~with= gases and so forth, and you put it in. The question is, 14 ' how good is it for the. conditions and' temperatures, and the 15 flows and the'gasesEwe have.
. 16 But it does have such correlations. It does not 17 solve the momentum equation.and it is essentially a one-18 - dimensional control volume system. Therefore, it really 19 .can't treat.3-D patterns, recirculation. It does not know
^
20 -how to do recirculation. 21 So, if:you?really have a-strongly driven natural n . circulation division' driven' flow-process in the upper plenum,
'n ~ for--example, it really,does not handle that very well --
24- even though it uses well-mixed assumption-that does.not'
. 26 take. care:of~it. .
4 f ,
. _ . . . . . - . . , .. . .. . . _ . = . - m.. ._
s^. , :. k. 174
;.c; ,
e
- m'
- h.a ag( . _ ,
l lit. gives ; you some id.ealof how good - the heat s
- , u
'2;
-tiransfer/ coefficients'are,' but it still- does not take care 131- ?of it;
[- L4 This;one"and this one are pretty important > t
' c. - ,.
M - A <5' ~ considerations-s
. We; don't.~really. couple.the heat transfer n , of Jin lthe l fluid :and. dynamics to the fission products as heat 1 s
i' 17: sources',:either!in thermal hydraulics-or.in some of the 1 :'
-ltransportlof1 fission products-themselves, the effects-of
~ 8' c
9[ :theftherma15 hydraulics.: l10 d
- I am getting1near the end. TRAP-MELT. Within
' 11 Ithe constraints'of!the. control volume formulation which it !
12 also'uses,'it uses pretty good l internal approaches also; . [y
? 13 - ' especi s1'ly 'for' the -~ physics .of the aerosols . . ;
14 -
- The greater part:of all the little separate effects
( L is .
'modelsLare;basedLon pretty good experiments with pretty 1-16 L good levels of-experimental. validation.
17-
.It has not had;a. good overall integral validation i;
' 18 set.:of experiments!yet. There are problems with it in my m
! Is' : t , :/ Emind.: The depcsition velocities for'some' of the things are ' f 20 ;
' based'on'a: sparse _ amount of data,fneed a stronger data base, t
- 21: ,more data at[ higher temperatures.-
, 32 ~ ,
once'again,yit does.not have.a good way'to treat g J . 28t natural ~ convection patterns in the upper plenum. region. Does 24 : not7 couple the~fiksion1 product transport to the thermal p
+
v . . . 235
' hydraulics. 'Those are basically the main' problems with it. .
I o,
. h e.:
- r 4
y p-4 n v 'c 1, c: a.. - , , , , _.:..=.. ....;.. ... ., _ ,,_ . .4,_ A.c , . . . , . - _.
175
- p 11 Those :are probably more minor problems.
- _ at j
' ~
. .-- 2 .NAUA, probably_the best-validated codes. There 3 has"been s a lot of experience with aerosols and with the This type of aerosol modeling has 4 -models-that go.into'it'.
5 pretty good acceptance by the community of aerosol experts. c6 'There are questions about things in it, how 7 'does diffusiophoresis affect-it, these fault formation s'- ' questions.: .I do not think it is that important. 9 -I think'if.you just use the glomoration models 10 ;and~the gravitational fallout models and forget everything v il' _else, you probably will get about the same answers. So, I
- n. 'think you can' rely on'those models pretty well with one is exception,=you have.to askuwhat effect the-well-mixed 14 assumption 11s on those things, and you haveL to ask how do you 15 treat multi-compartments.
10 MR. SILBERBERG: Tom, excuse me, a point of 17 - clarification. 18 - You are saying once one knows, gets through the i a thermal. hydraulics concerns of what material is available i 201 for condensation and then you have the condensation and
- 21 . growth,-thenLthe physics is okay.
22 Are you excluding from here.the thermal hydraulics 23 problem we were discussing before?
-M MR. KRESS: Yes. . ' I _ mn saying that now it dees
~ 15 . not calculate thermalthydraulics,!ao'it"is not part-of the.
I t *W u l _
. - 1 _ ..
176
't-
. validation question.- It'is,.how' good ~is the input coming
~ ~
2-
~
-from some other code.
;3 I think that has'an effect on-it, yes. I think
'4 if you had those~-things perfect, you would have a pretty is- good ccde calculation with NAUA.
6- I mentioned already - down to the last one, origen,
~
7 ---
.I don't'think the uncertainties there are likely to have a
.8 l big ~ influence on the source terms.
9 - Finally, to give you some more things to think' 10 about,-some summary comments on things. One comment,
- 11. personal comment. If we are really going to calculate 12 source terms, I don't think we have any choice but to 13 develop good mechanistic,-deterministic codes to do it. I 14 do not see any other way to do it.
15
' so,,if we are going to develop source terms we is are on the right track in doing it this way.
17 ~
. Basically the codes,'most'of them, have pretty is - good, sound physics in them,-good ~ approaches-to it. -It is a good.first start. But they are for the most part lacking
-in overall experimental validation and this is in several
~
21 important areas. I think the physics are good, with some 22 ' . exceptions. In a.;few areas,-I think, the modeling is crude 23 and1 simplistic, and non-mechanistic. I l M I make the comment that these phenomena where this ! , 25 ' :is the case are. complex and tough,-and diverse. But'I do u 9 r,. ,-- r
_ , 177 1 notsthinkLwe should let that slow us down. I do not think
~
2- :they are' overwhelming...I think we can do better. I have in
-3 mindshere particularly two major. areas -- that is the core 40 tmelting' slumping progression model. That drives a lot of
' '5 ~ this1 source term. We need to do better.there.
6 And,.I-havelin mind-the thermal hydraulics,
'7 particularly.in the upper plenum region where you have the recirculating' flows,; coupled _to the aerosol behavior 8-Tand1the fission: product: behavior.- I think those areas are 10 whers the modeling is_ crude, simplistic,-and needs some
~ 11 -
help.
~U-I_think we should do bettersin the future in 2 corre'eting [ these de ficiencies . There are a' lot of 14 L inconsistencies 11n the codes. .In one code we use one model U and in another. code we use another model to do the same 18 thing.
17'. You know, forf example , in MARCH to calculate
.M- ; loadings and'in CORCON to calculate the release. There are U ' lots of these thingsfthroughout the system and you really 8 ought t'oll remove a' lot of these inconsistencies just for 21~
study purposes,_if nothing else.
'8 - I think.weLdo'need to couple thermal hydraulics 28 -and fission 1 products a little bit better.
' se So,:that.-is a: flavor ofLwhat we are saying. While
^8J fthis,is personal 1 opinion, a. lot of.it I got. straight out of.
v
- 4
- ~ .
s 4 , m [, " [ ig ) } 178 y 11 ithb reports themselves. I am not even sure I know how to f+
,7' b e
-w '}
- 2:
--makeithe" division between.which was which yet.
3
- If-I'ican answe'r any questions?
f4- 'i-n<. LIPINSKI: Lipinski from-Sandia Labs . 8: fI:thinklone' area ofLuncertainty which may have w .
.6'
[been7 omitted in'this presentation is models for phenomena
'7- ~
iwh'ich.lareLomitted.from the codes which you are addressing, 8-fin particular;.for-example,'the melt that comes out of the l reactor.1 vessel,.how far does'it go? How far does it spread out ? iNo-code really has that in-its' jurisdiction and2yet, l11 ' fthat'is an":limportant parameter which is omitted. Therei
,ef .{ -
1are many.'others like that. So,:'the codes address what they 3
, are?l oo'k ng at^perhaps well,'but they are also missing.some 14 '
fthings which have>yet[to be put in, in the future. TMR. KRESS: I-agreeiwith Ron.. The: separate
' N-
; documents.which we are putting together do address' potential 17 b misshg phenomena.- .The problem.I have with:that right now 18 .
fir, I don'k know:how'toLquantifylthose. _, 28-lI do mention them'in'thefdocuments and even in
'my. summary appraisal.
~
You can name>several areas'.: Tor
. 215 example, this question ofs flashing.of!the water has not been 22 - ;in the models.
~
But'by the'same} token you could flash-fuel 8
.asfit' blows'down..:'Those havelbeen.mentionedLhere before.
'Ilam$not;sure.how to quantify those;and I-am not
~
E~ isure-how-Sandia will.in.their.U.ncertainty. Analysis. . But it g t.
- p. .
e u l .. . _ -- _
179 I x si- . would be well to quantify those missing phenomena. 2 :MR. ROWE: John'Rowe.
- 3J I have t'wo' comments or. questions. To what extent
~
' .4 are the codes.being checked'for consistency? Several people over the past several meetings have said things like, "We 6 .can't expose the h' eat balance to the mass balance," and things "7 of that' type.
8 The'other that-goes:along with it is that some 9 o'fLthese; codes could,:in a simplified form or strip-down 10 form,~be used to-calculate known analytical solutions.
.11 Some of.these might be viewed as trivial calculations 12 i but'what it does.do is, it does check to make sure th.at the
~
13 codes are internally self-consistent. You might have a 14 .modelLpulled right out of the' literature that might be
- 15
- accepted', but .'is :it really codes right; .is there a factor of 16
~
- two' missing? There are'various things that creep into 17 codes.
18 I am just wondering if these are being addressed 19 in'some way.- 8' MR.-KRESS:l The answer is, yes, in general you 21 1 would ask.the question of any code,-does it violate any of
. 22 -
- our physical-law,.does it conserve mass,-energy, et cetera.
23 i
- Most..of those constraints are built'into the. codes.
24 - (There are' comparisons with-analytical data. For . example , 28 MARCH during the early h' eat-up-phase is basically a heat
~ 180 1 . balance. It does'a pretty good job. MARCH has been 2 compared ~to 'the PBF experiments and it does relatively well 3 ~ during early phases-of,that.
- 4- Things like the aerosol codes, you can turn off 5 the fall-out models and see.if the glomeration conserves mass,
' 6' for example because the: mass'can't go anywhere. That has i
7 -been done. And the models have been compared with each
- 81 other and similar codes like CASS'and MARCH have been compared .
'9 NAUA has been compared'to QUICK and HARM and other codes.
10 - MR. ROWE: Are there catalogs for consumption in 11~ the report? 12 ~ MR. SILBERBERG: My understanding is that you do 13 have-a section in the report which tries to summarize the 14 status of analytical validation, if you will, or consistency checking. Right, there is such a.section? 16 MR. KRESS: Yes, verification and quality' assurance,
-17 .which is what I consider this to be.
18 When I said before it was weak in-the area of 19 digging out the data bases, the second weakest part is that 20 .part.there. That is just the way it. turned out. 21 -Kind of a broader comment on that-same point. We
' Et have a' Nobel laureate in this country by name of Hans Betam El who once made-the comment that no'one could use a computer 24 code until.they knew the answer to one significant figure.
25 I am'not sure that this is theLcase in the present 8 _ /1 .
o =
, 181 YJtj exercise. Il think- it would' make a lot of sense at. least 2 for a calculation to be-done on some of the more significant l3' Ph enomena, the back of'the envelope type thing, and one 4 significant figure before the computer codes were used. This 5 would serve two purposes. One, partial validation of what 6 is going on. At the-same time, it.would greatly aid the 7 reader of.the reports in-understanding what is going on.
^
8 That-is just a comment. g MR..Z'MWALT: U I was thinking that perhaps chemistry 10 - if in a little less good shape as compared to physics. One 13 : reason is, there is admitted. phenomenon, I_think, which was 12 - mentioned. 'An example would be;the inactionlof aerosols 13- .with the fission products. 14 In some'of this the~ problem.is.that.we do not have ni +'e data base to work on. I think that our program is 16 . dressing'this problem ~t hat is there, and two examples.
'17 One is the activity coefficients which Ifthink may also be 18 important to CORCON,.rather CORSOR because you are going from igF a place of a~well-defined geometrical' field.to a molten zr thing. When you get to the molten thing, you need something 21 about. activity coefficients or components of these 22 constituents.
23 The other, of-course, is the~ inaction of some of 24 .these reactive things.like silver in'an-aerosol with things 2s like' tellurium, s i
?! .
I .
'~
l h,,i r- , I
~
l n 182 l l s1 MR. KRESS: I would like to comment on that, Lloyd. 2- Of' course ~, CORSOR is strictly an empirical model and u
- 3 supposedly.in puricism-covers a multitude of sins.
4 But you,are rignt, the. coefficients should be 5 : important to release. I think there are a lot of unknown or 6 . untreatedlchecmicalLquestions. 7 For _ example , we are still -- 'I think everyone a .isDsitting here buying off on cesium iodide and cesium
'8 hydroxide based on chemical-thermaldynamic calculations as
..m conditonssof interest.
114 'I-do'not know if we really added in all the silver E and the cadmium, and-the boron that one might have into U those to see.if)they make a difference.
- 14 I think when you make a chemical thermaldynamic 2 calculation you need to:have all the species there. And, 16 'of course,1there are chemical reactions of: surfaces, and
- - 17 _- -with-aerosols, and chemical reactions in'the melt, and 8 _ perhaps. chemical reactions on a long-term basis in N- - aqueous conditions and with-production'for the methyl
-8 ~ iodide-and that' sort of. thing.
21 -The question is, how important are these? I 22
. would hope.:that Sandia lor someone would factor this sort of 23 - thing;into;their: uncertainty sensitivity analysis.- 'I really 24 -
had no 'wayL to do. it, = address the importance of them in 8 this study. e b 9
'183 x1' MR. ZUMWALT:' -That'is bringing up one point and 2
I Ldon't'know ifIthis'is supposed to come in here since we But 'a lot' of boron can be
~
areinotitalking.about BWRs. 4
'releasedIfrom. BWRs tand there is a - fair amount of boric acid 5 ~
in.BWR water. 6- - So, cits- release 'should be estimated and. also
-chemica15 consequences. considered. Sandia has made-a good start on this. _It may'have;a very important effect on maybe
' ~
makingia more volatile.~re1 ease of' iodine or holding cesium
.better,~or what[not; a number 5f things.
,11 :
MR. SILBERBERG:' I: guess the question there would-12 be, there may be these, if you will, perhaps second order 18 of reactions, or.in other words some attempt actually made. 14 The question is, under these~ conditions, will they II make a contribution, will they be in the residual part of 16 the reaction..or will they actually come forth as being the
~
' dominant or: major form.
That, I think, is the. issue that is open. I 8- think those are some of the questions that we are really
" ~
asking Sandia in doing their work. 21 You know, if you get one percent of the reaction 22
.because'ofIboron, okay,. fine. . We ought t.o know that.
But 23 - ik the big picture it:may not affect it. It might affect it as'you get'down to resisual release, say, got-rid of 99 percent of the.-stuff, what about the residual 1 material that i e
-= x 184 L1= might be released? Then it would come into play at that 2: ' point,_ having gotten the major release of that out of the way.
3 ,Again, its. impact on risk consequences would again 4' :have;to.be' evaluated.
'5~ :MR.' WALKER: . I have two. One directed to you, Mel, 16- and'-I guess 1the other at Tom.
J7 You called this afcode validation. task and I guess "8 -from what--I hear 1I don't think you are trying to validate 8- -a codefinLthe same sense that one validates it for design or 18 'like one would; validate relap. 11 ' I guess I am not quite sure what it is you are 12 - trying to'do. When.you get through, are these codes going
~
13 to be. validated for use by the industry in picking up source
>14 . terms, or does what you are trying to do have some 18F technical-expert tell you where the major holes are? Or 18 - are you trying to have him tell you that generally he. thinks 17 - ' things arc.okay?
18 .I am not.quite sure what you are.trying to 18 ' accomplish-in'this task. 20 MR. SILBERBERG: The-first thing, in - terms of end
- 21~ -oojective' -- the approach .might, be somewhat similar but 8 the end-objective is clearly not'in the direction of the 28 ' design code, even though one might want'to use the same M . principle.
25 1 ::I would ask Tom or even the people from Sandia to e 0
,f',
=
p - I 185
- c. .
t H yadd to wh'at'.I say here, if they would like to. 1-
- 2' 'MR. WALKER: uJust let me add one thing. A lot of I .
F 3 that depends on how you implement that QA objective we
- 4. thad up;there onothat slide.
5 MR.'SILBERBERG: But_as we see it, the task at a' hand right-now and in the near future is to say, we have 7 .a methodology. 'The methodology, how does it stack? In b .s. other words, to what extent does the methodology reflect 9 data, a data base? .What. supports'the methodology that one is ici now_trying to come up with in a refined way, so-called ( 11; refined ' methodology? i l 12 - What'it is, it is really a fine line between how
- 13 well do we know is'the code validated against supporting 14 data in the context of an uncertainty, an uncertainty range 15 for.the application that is being used.
i 16 Now, let me"try-to explain what I thought I was 17 - telling you. 18 (Laughter) 19 MR;!SILBERBERG: If one is looking at, you know, leL factors in a methodoloty that says, "I am looking for numbers 21 Liike factors of ten, or eight, or six, - or a hundred," or what a
~
have;you, if or.e is going to make a change, what this ! n- methodology is to bring.about,,1f you will, a robust change E -
. Ml in source term of that order for a 'given situation, a 2B 'givenisequence -- I' don't want to talk about source term V
186 1 across the board -- then what is its level of validation that 2 would give you confidence that you can go that next level, 3 as-opposed to worrying about ten percent or five percent, or 4 all the physics. 5 I think that the whole thing gets put into 6 perspective when you take an exercise like the Element 1, 7 the Oak Ridge report that Tom described and the Sandia 8 sensitivity and uncertainty analysis, and basically you put 9 those two things together and say, " I am never going to 10 really be able to test all of these codes at full scale,"
- 11 I niean the whole ensemble of them, particularly, let 's say, 12 MARCH. I hope we never have the opportunity to_really test 13 it unplanned.
14 You have to really ask yourself, given the 15 methodology I have and the limits of validation, and the 16 application I have in mind, how far off might I be in my 17 application to reactor scale? 18 MR. WALKER: Let me make a comment. I think that 19 last slide that Tom had up, the very first bullet he had on 20 it, said that the way to get there le goody validated codes. 21~ That is kind of a nice " motherhood" statement, I 22 agree. But you have to put that subjective judgment in, ZI or you spend all the world's resources in getting there. 24 MR. SILBERBERG: What you need to do is -- and I 25 think Tom knows this, he did not have it on the slide because 9 O
, ;W ' :;' , _ g' ,
s vm
, r- ~
_ 187 u .- ' ' , jj s, % , _ 1: . 1Ifknow:weitallted
~ -.
abontiit,. basically you have to set a
- y. 2 : validation: criteria which says, " Projecting ahead to the'
' ~
^
{3: Euse 'of? this -methodology at full-scale I want to set some b4 ;.
. tkind of a.. confidence level that I know.within some amount I
is; iand.setLa: criteria,!and then judge your validation task." l I
- e ,
1Just to'say, "I want a validated code," validated ' (7- '.forcwhatipurpose and to what,. criterion? It is the purpose 8 of validation criteria Land that: kind of- thing' that we are
,9. LjustEcomingito gripstwith,-it is the document, the Oak Ridge
. 10 ; -documenti.and:theiSandia study that will help us define those 11 : ' criteria,- Ii b'elieve.
12; LMRCWALKER: I agree'with you. I think'that until la you worked the-problem all the way_through and in addition 14 Einclude.some: context:on ther importance which you have to get 18 _. out' o'f f.a . containment , a much better ' containment event tree, le. (you have ,not' got a basis to- make. the kinds of judgments .you - l17f need.to make.
~ 181 One example I saw up here today, just as Tom did, 19 I .he said;we need a better core migration;model, for example.
" ~
t so - ..I think theathing that-has. popped out of.all.your evaluation 1211 is'that indeed that might-be important in the TMLB prime
' 22 ~ ;seqnence.- But -Il think in :the ; rest' of. the sequences it is not m, 38 .very
- important. Maybe TMLB' prime is not too important any
. 36 : more..
~
^
1 35 - iso, I guess.I don' th1nk we,get to a place where yg 's i
.. fe
'4' '
( i 4 >
y,sp, -- -
-r . 7'- - ~ -- - . m n - :c~ y- - - - ,
' - , s
> r q- 7s , ^ -
? T~ >
188
~,
, , -n
- m. ; ^ + ~ ,
C [' '
~ . -
' ;;f 1h-...=y .
- y_. , - . ~31' .wer can make~ those : kind's off j udgments , t pqgp ~
c
~ j7 s -
EMR.'SILBERBERGs 1But-we need to do that.
~
, i2- . -[
p. 1 . ; 6 e , t h : e"- , ~ ~, _ c 3' - ~MR..WALKERiJ yWe . need;to, yes, j ce I
=~, ,
f- 4'- ~I-5think there-are still pieces missing in the 2
,s. . ,
1 , C. t.a l$ + x 8:- ~ ove.rallipicture you are tryingito.put together, particularly
~
.m . . N 4: thefsystemsTi nteraction. kind of= stuff.
That is just not in ;
- J7 there .yet'.
m .i 7
~
', -8i '
, < EMR. SILBERBERG:; , Right.. ; Do you want to add- 4
-a s. .
. .. ~ #
~
J8 :something? n :e , ile - [MR.MIPIN, SKI:- Lipinski, Sandia.
- s.p :
,L "
< 4 11 The' uncertainty. study.can tell'you how good or s
< 12 ' I how bad tihe codes are. , But they cantt give you the l
+ ,
E 18E criterioni:for validation.- 'That has.to.come from decision a '. s t fmakers up abo,ve.: 14 : i % LL L18) -MR.ISILBERBERG': ?You need to: combine them.. Go ahead . i 16 ' MR.iCYBULSKIS: This;is a Jim Gieseke slide, maybe-L , 1 n; V .. , . c17! l:Ir oughtito let him: talk about it..
? t p
- 18 1 (Laughter) ,
h^ n 18 I MR.'CYBULSKIS:- But in.the-final' report,. Jim and
- ILput'together aisort o'f'a=forewordsor introduction that 88 1'
,' * ' 21 f
[ ; spells..outhwhat; wel reallyj thinkjis[ meant iay ' validation. It 88 sinvolves'this! set of cyclical processes.
, - , + n ,
- 3 $.*
7 ,
, 4 - ::
ss ! 'InMthere;is, of course, validation goals. . lThese-L v _ .T
. . . . Jis like, how good 1are py - Se * :have?torbe':quantifiable?, goal's'andiit -
s - 4 35 -
, , :- . :thefoutputof,thesel codes;Jwhat.'isoure confidence' level;in. ,
g Vy 4 ') , t l / ,. f I . A- 1 , ( 4
~
fy., 4 y ? v'
, q 4
? l 'y t
, l , r .y ,,m: r. 9
,,. , d ,.,
M , Z t y.:
>,g,.
7, , .nm , C, _,j '* 7w
189 1 4
! these outputs compared to what we want them to be in terms p.
2 of 'some goal. It has to factor in, of course, experiments ~ 3 and the modeling, and the non-modeling, what is not there, 4~ and-requires a sensitivity uncertainty analysis. 5 It'all has to be eventually based on experience. 6' Those'are'on full-scale grade, but we can't really do full-7 ' scale experiments for most of this stuff. So, we have to 8- factor in the scaling levels and'what that would do to the 9 uncertainties. 10 So, we realize that is a big order and.this little 11 thing we put together is a review of the status of validation. Et 'Such a review should say, how good are the codes in meeting 13 a goal? That is-what you mean by the status of validation. 14 But we don't have a goal, we don't have a 15_ . definition of what validation is. So, you know, it-is an 16 impossible-job. So, we could not.really meet that, all we
- 17. could really do is'say, are these codes based on experiments?
.18 - What are the experiments, -and give you a- subjective opinion N plus, give you a lot of _ help in telling,you~ what is in the 2 codes and[what is.not in_the codes, that sort ' o f .s tuf f.
21 So,-it'is not:a code validation exercise at all,.it et - is a. review of the status and it does.not really have a good-
?
2- basis on-how to review-the status. So, we did the bet we. 24 could,:really. 25 . . MR. WALKER: Can.'I'just_makeLone other comment?[
r 190 1 I guess that is my perspective that may influence what you 2 do on validation. 3 It seems to me that what we are trying to do is 4 I contain the radionucleids that give release in a core melt . 5 . accident. If we had perfect containments, we would not have 6 to worry about fission product behavior. 7 Generally, what the studies of the industry are 8 showing, what they are telling us is that by and large the 9 containments hold together for a long period of time. If to that is true, the containments get attenuated in a very 11 marked fashion in the containment itself. 12 What that tells me is , the important codes are 13 those that predict the containment attenuation and you can ! 14 almost forget about the rest of them if you 'can keep the 15 containments together for a long time. 16 MR. KRESS : If Mel's Containment Failure Loading 17 Committee comes up with that conclusion, I think you are 18 exactly right. 19 I think it is clear, the containment is the key. If 20 it does not fail early and stays together, why, your source 21 terms are probably going to be low except for those sequences 22 that bypass the containment. I think that is clear. 23 So, I do not.think I would disagree with you at all. 24 MR. LEVY: One part of this process, I think, has 2 been mentioned before. It is not clear to me how you are 1
.Fl'.
191
- 1 going to fit it into!the program. I think it was raised 2
- very early,Leven the;first session, about some comparison 3
to real accidents. 4' I think it is a way.'to get-a feel abo'ut whatever 5 package .you have is an' order of magnitufe of f 'or only .a factor 6 of two'to three.off. ~I am not too sure where that-is going 7
-eo get done..
. 8 I gather there is an. uncertainty analysisi there 8
-is this what I call evaluation.rather than~ validation. I 10 think there comes a time when the containment holds it,-we i
11 don.'t have.a problem. But there is as good a case ~where 12
- the containment held together an'd we can make a good 13 comparison of the fission product release up to that point. I 14 ithink we ought-'to do it.
15 MR. SILBERBERG: Any other' questions for Tom? 16 I now would like to introduce--Terry Butland from 17
- .the United Kingdom Winford establishment who has been kind 18 -
enough to come here and present a summary of acstudy that 19
-he did~on the TRAP MELT code, a sensitivity analysis using
'8 '
the TRAP MELT code and becoming familiar with.it and doing 21 a variety of things with it, I' guess, and to it.
. 22' ~
I think you will find this to be a rather-23 ' interesting study.
" -Mr. Butland: Well, this talk goes-a little bit 26 : farther than-just TRAP MELT in actual:. fact. You will probably
~192 }
l' see precisely how far it does go as I go through the talk. 2 What I put there is the source term plans over and 3 above}where they.were ten months ago,' w hich is where I J 4 personally. started. 5 'Of course, we have source term estimates in the 6 UK-prior to.that. But from ten months ago, our plans wcre
- ^ 7 to acquire relevant U.S. calculated codes and.to implement 8 them in the=UK..
9 So, we had' aiready aquired MARCH and ' that is 4 10 Version 1 of MARCH -- yes, Version.1 essentially. Ab'out
-r 11 ten months ago, we aquired MERGE, CORSOR and TRAP MELT ,
. M And we'have implemented those.. U- Then~our plans were to assess and validate these 14 codes as told in source term predictions.. You notice that 15 . so far-I have not made any; mention there of containment 16 codes, 17 On the assessment to be done by performing what 18 we called a " mimic" of the current, that is what is being 19 discussed here -today, U.S. source - term work with the Surry 1 20 PWR. Now, I know you were talking about many others but we 21 -have a~much more limited amount of effort and so we have 22' concentrated so far here.- That is under way and I want to 23- ' say a little about that today.
- 24 Then, the other aspect of the work is the 25 examination of
- the theoretical modeling incorporated in.the
193 1 codes, and that is also under way. 2 Then, by using the codes to predict relevant 3 experiments, such as MARVICAN. That is, I guess, beginning. 4 Improved basic data such as chemical interaction 5 rates get incorporated in the codes as they become available, 6 and in the UK at the moment we are looking at deposition 7 velocities for particular species. Indeed, one of the things a that is emerging, for example, is that cesium hydroxide 9 does have a temoerature-deoendent deposition velocity which to probably we ought to take account of. 11 Then, our ultimate aim is obviously to improve 12 UK estimates of accident source terms. ' 13 Now, when we acquired these codes. I wish to apply C 14 them to the Surry 1 calculation. We made sone changes which 15 were easy to make at the start as opposed to simply rewriting 16 a code or something like that. So, instead of using ORIGEN 17 for inventory calculations, we used the FISPIN code which 18 has a lot of associated libraries and a lot of people who 19 work on its libraries, the validation of it. 20 This is the MARCH 1 version which we have available 21 to us and I know you are using different versions in the U.S. 22 now, MARCH 2. 23 With regard to the MERGE code which does the 24 thermal hydraulics of the primary system, we made various 25 changes to this code which I will come to in a little while.
194 1 With regard to CORSOR, we have developed that a 2 little. It is not a mechanistic development, it is still a
-3 I bailding in experimental data, but we have tended to rename 4 it MATREL and we will go into that a bit in a minute.
5 Then, TRAP MELT, and we have not yet in our mimic 6 study of Surry 1 gone into the containment. TRAP MELT, we 7 have not changed very much from what we received from the U.S. 8 Now I-want to just say something about inventory. 9 So, the first stage was a comparison. of the inventory for 10 the Surry 1 PWR as calculated by the U.S. ORIGEN and the 11 UK FISPIN route. This was calculated for a mean burn-off of 12 22,000 megawatt days per ton, and I understand that Surry 1 13 is a three-batch fuel cycle. So, this means that one batch 14 at the end of the equilibrium fuel cycle is at 33,000 megawatt 15 days per ton. 16 The U.S. results were taken from BMI-2104. You 17 will notice that there are similarities but also notable is differences. The cesium is very similar, but the iodine is 19 quite a lot different. We are calculating a smaller 20 invento ry . We will come back to that in a minute. 21 The tellurium we are also calculating a smaller 22 inventory, and notably too on the silver, fission product, a that'is. 24 Nou, in BMI-2104 some of the inventory is grouped 25 so another comparison is possible which you cannot do for some o
r -
+
r
, _ 196
'I not contributing to a. radioactive. source, not contributing 2 .to' dose. This is contributing to. dose. So, this discrepancy 3' is not.as worrying as one might think from that point of view.
4 But nevertheless, since iodine - figures in chemical reactions 5- and~ we ' hear about cesium iodides and so forth, :then it is 6 -interesting to see that we get a different mass of iodine
'7-1 through 9.
8 . That is what has come out of the inventory
~8 . investigation so far. We wish to follow that up by looking
~
10 at tellurium and some of the others'later on, but we have a 11' not yet done that . thoroughly up to now. 12 We then moved into the MARCH andI MERGE calculations , 13 ' and I have ust summarized some of the changes that we made.
' 14 Whilst'we are still using the Unit'ed' States codes MARCH and 15 MERGE,' even though we would' feellwe would like to make 16 changes to them in certain areas, what we have done in the 17 mimic study is to examine the data that we are.using 18 . critically to see what impact that might have on the final 19 resul'ts.
8' Tks have changed structural data for the Surry 1 21 PWR,'andiI think this is a significant' point'in' comparison 3 with what was done in VolumeLI of BMI-2104.- We have changed 8 the radio power profile so that it is 'in fact much flatter
~
24 and much more representative'of what would really be the
' 8 case. . It;is my feeling that in' the original document tit was
-.a-. . 1
r;; 1 195 1 of the fission products. individually, notably these h'ere. . 2 I have included that on this' table. Again we get very 3 similar comparisons except for iodine and bromine,1and that-4 difference is due in fact to the iodine effect. - This effect 5 here in Group 5 is the tellurium difference. The others are s e very similar. 7 . What I want to do now is just to follow up on that a for1just one of the nuclides, and this is iodine, and ask 9' ourselves why we got that dif ference for what was probably to regarded as something we should all.know. As-far as I can 11 see it comes like this: 22,000 megawatt days.per tonne. The 12 mass of iodine is made up mainly of,these muclides, primarily 13 the stable 129, 14 If we look at the chain yields, the comparison to 15 what I call C3I there, .which is RUK- dataset on yields -- 16 you will see that as time has passed the two have become 17 farther apart, and FB5 is then moved away'from us. Furthermore , is we-have in the UK, or from the UK, a fairly large examination is of uncertainties associated with yields some one' called 20 Eric Crout and he has one sigma uncertainty for-the chain 21 . yield'129, 23 percent. So, in fact one might argue this 22 dif ference is not totally inconsistent with what he regards 23 as the uncertainty in the data base. 24 Now, the important other point about this table 25 is that: 131 is in much better agreement, -and of course _129 is ' O e
197' 1< far-.to "coson" like -- it was'not exactly coson like but it :
, -t
.was far too much in th'at direction.
2 .- 3- We have made various changes to -- this should 4 really be the fuel. pin melting temperature and not the-5~ f uel: itself,' a conglomeration. l
- v 6 We have adopted some dif ferent water ' reaction 7 data. -This is a fairly important. point.. It seems to me that 8 tne MARCH-code is very. dependent on the. times that one l
L '8 adopts,'the maximum times'that one adopts. We reduced it
?
10 . :byy a factor of ten, and those have-some important numerical 1 11 -facts. l~ u We also did not assume any cooling in molten L D noces,. nor any -- water reaction ~ in molten nodes on the
. 14 -grounds that it did'not seem to me if;the node was not l-15 molten, any steam could get into it.
16 So, those are a few of the changes.we made to the
- l. ,
17 MARCH data. We then also made some. changes to MERGE. .943-l 13 felt the: code was not treating, time-step control properly at l' 8-the time'we received it, such thatifast accidents like the l 20 :A81 Haltleg - =which is the only one I am going to be-talking L 21 ' about here -- was not _ properly calculated. ' \l 22 We were'able to correct that and carry on.. So,.that ' is a brief summary of: the main. changes made at the MARCH-
~
23 24 - MERGE stage.- 25- Now, how do you summarize " the -ef fects of something 4 r, - a n + , n ,-n , ..- ~ g,,---
E . 198 i 41 like'thatfbecause MARCH produces an enormous amount of work? 2 Well,1we are only; interested at thisi point in primary system ' 3 retention, ~ we. have .not gone on to the containment yet. So, 4 one can summarize to some extent because of that. 5 _The only. things that.I am going to present here 6 are two or'three slides. Here is the start of core melt. 7 We are roughly the same. But we predict a much faster a slump. 8 Now, this is still based'upon the old assumption-10 in the work of the early BMI-2104 study, that is that the 11 core slumps when it is 75-percent molten. We did not have U any of this dribbling that we had today. 13 (Laughter) 14 The difference is that it emerges essentially'from 15 the radio power profile being flatter, which enables one to
-18 get to the 75-percent molten position faster. That, I think, 17 is the main reason.
18 In fact, it seems to me that the' radio power 18 -
. profile, what you assume there, is quite an important 20 piece of data for this:particular core slumping model. It is 21 ~
fairly obvious if you think about it as to why that should be. 22 Now,.one'other piece-of data which I wish to put 23 on- the board is not because we have differences but because
'N' of something I want to say about releases. Now, this slide 25 has not'come.out all that well, I am afraid.
,, g ,4~_, .*
1-This is the steam mass in the primary system as . 2 a' function of time. .I should:say that all my times and all i 3 my slides relate'to' start of accident, having been through
-4 .the labor myself of wondering when the start of graphs was,
- , s 5 'that is'what I have adopted.
6 One notices the steam mass drops down quite rapidly. 7 Whetner that is' correct or.not is of course another debate. 8- What I want.to then say is that if we look at the hydrogen i-8 ! mass-there is a different scale'here, but the hydrogen mass
- 10 rises. Now, why.is that important?
Well, in my view it is important because of one 12 l. t og the parameters that is missing from CORSOR, which is [ 13 atmosphere or environment because release would be 14 expected to.ba function of the environment. A_ steam 18 environment is an oxidizing environment, a hydrogen environmen le is a reducing environment, and that has been shown by a-17 paper reported at the RESO conference earlier this year to be is an important consideration in release of fission products. ' 18 One particular mechanism would be with regard to
# tellurium. We have heard today that Tellurium is now being 21 modeled' not siraply as a simple release as it was in the early 8 studies but people are considering the raction of tellurium l}
8' l and zircalloy.- In a hydrogen environment the zircalloy would 8' -not be oxidized so that. reaction'would go on at one 8 particular rate, effecting the release of tellurium into f _ . . _ _ . . _ . . . _ _ . - . _ . _ _ _ _ ___._____._m_-
4. 200 l' .the atmosphere, and'in a steam environment the zircalloy 2- would be oxidized and would go on reaction betwen tellurium 3 and the zire, oxide, going on at a different rate. 4 Now, it was shown in this RESO paper that one 5 can have for some fission products factors of a hundred in a release rates. So, that is.one effect that is being currently 7 not modeled and obviously should be. 8 Now, if we go a little bit farther and ask, well, 9 how do our release rates compare, our mimic release rates -- L - 10 not by any means our final release rates -- compare with i ! 11 the BMI-2104 Volume I? This is one way of summarizing its 12 This is the core inventory released prior to core 13 slump for AB Haltleg UK and U.S. We}l,whatdoesUKmean? i-l 14 It means UK inventory calculated by FIFPIN; UK core i 18 temperatures calculated by the U.S. code MARCH. , 16 UK release rates, and the U.S. means U.S. inventory l 17 calculated by ORIGEN; U.S. temperatures calculated by MARCH 18 1.1, and the U.S. release rates. ! 19 Now, the UK and U.S. release rates essentially 20 come from the same data base. So, we have an. issue of l . 21 interpretation of that data base here. 22 Now, if we look at the differences, I' will. try and 23 say why they come up. There are really two sets of data here. 24 One is the ' percentage of the inventory released and the 25 other is, well what is that amount in kgs? i
f 801 1 Iodine. The difference there is because we have 2 a different iodine inventory. As a fraction of the inventory 3 we both have at the moment the same sort of figure. 4 We look at the tellurium. There is an inventory 5 effect there as we saw earlier on, but it is certainly not 6 as big as this. And there is a difference there in 7 interpolation of release rates. As Tom was saying, I think, 8 earlier on, there is a sparcety of data and one draws curves 9 through it. 10 Well, in CORSOR the assumption is that the release 11 of material per time DT is a fraction of the inventory there 12 and the expression they use is AE to the BT, where A is a 13 constant and B is a constant -- well, A and B are functions 14 of temperature and this is a sort of average for the data. 15 In the UK interpretation of those numbers, we 16 in fact do not have an A and a B, we interpolate directly 17 in the measured data. So, we have a release rate difference. 18 here. 19 liow , this is very apparent with barium and as far 20 as I ca'n tell there, the reason for that is that the barium 21 release, as far as the experimental data is concerned, it 22 does not extend over a very wide temperature range. So, 23 there is an assumption being made here about extrapolating 24 outside the temperature range, is leading to that difference. 25 That again is release rates, and on strontium. Going on to UO2, there is a difference there which comes
202 s. I about because we have at the moment an assumption that 2 when the clad breaks, UOr dust is released. That is not in 3 the present CORSOR. That gives rise to that master difference , 4 That does, in fact, raise one little minor point, 5 minor. quibble that I would have with starting. graphs at s-6 the start of core melt. Of course, clads break before core ; t 7 melt so-you have gap inventory coming out before then.. 8 I have a graph here which goes into the position 9 for cesium in a little more detail. This is a busy graph, to as they say. So, we have this symbol which is all U.S. 11 data which I think is running up here, and then the old UK 12 data is running up here. ! s t l u But progressively we have changed the temperatures, i 14 the inventory, and the release rates. One sees in fact 1 15 ' that for cesium in terms of the final mass emitted, by the le time of core slump it probably does not matter too much. But 17 of . course, that is not an exhaustive, by any means, uncertainty l 18 study. - is I have done the same here for iodine, and the spread 20 is a bit greater. The old U.S. is up here, and the old UK 21 is down here. So what do we have, a spread of about three or 22 four in ten which is, I suppose, quite large. 23 So, I think there are one or two interesting 24 observations there about releases. Let's now just go on to l 3B primary system thermal hydraulics and what do we come up with l
~
b 203 t 6
-1 there so far. This 'is unfortunately _ another relatively l.
2 " busy" graph', but we-have steam temperatures here in the i.
,3 primary system as a function of time.
4 .The U.S. calculations are-these dotted lines. I t 5 The UK calculations are the solid lines. Now, this line here 4- .is'the U.S.-calculation for Volume II in the BMI Volume I ,
'7 study. Volume II>is the upper grid plate, so that is the
! .8 gas temperature around the upper grid plate. The U.S. - V ' 8 calculation is predicting a lower temperature than the UK 10 calculation, using the same code, essentially, but with 11 -the change in the time.-- control. U Now, this does not mean that this is what I
/ 2 would' finally say-is the answer to this temperature because 14 we are still evaluating MERGE and we have heard about lots of -
i is things today,.what is: wrong with.that. But this is just the 18 change of two things that have caused this -- three things, i 17 I think. , 18 One is the different way of running MARCH'. The is second is a time stat control,'and the third is where do i ! 80 you start a MERGE calculation. ! 21 The MERGE calculations, as far as I understand them, l t 22 in the past have not been started at the beginning.of the l , i 28- accident'but some way into it, usually at the start of core l t 24 melt or something like that, f 28 Well, I would contend that is not a sensible thing P p P S 4
+
304 L .g to do because ~ you have to feed in some intial conditions,
.f you like,'for the structures at the 2 some temperatures, i 3 start. That is not a trivial matter.
4 So, we have tended to start MERGE'as far back as 5 Possible in time when the temperatures have not started to shoot up. That is, I think, one of the main reasons why our l g _7 temperatures are higher for steam early on. 8 MR. LEVY: Excuse me, are you sure your power
, distribution would not account for some of that?
10 MR. BUTLAND: I think chat would account for some 11 of it'as well, yes. But I think this, you are seeing most l l gg of the changes in these temperatures are early on. I think l 13 myself that that is partly due to the initial condition issue. g4 Now, we can take this a little bit further and look g at.tnestructuretemperatures,andthesameobservation,really ,
,14 emerges because we are supposed to be comparing here this
! 17 dotted line with this dotted line -- that's right. So, our i is structure temperatures are higher. l , j gg ~ Now, what happens when you get to TRAP MELT,'where se do we get? Well, coming ouf of this we have another interestin g et observation. First off, we have just there cesium iodide, a cesium hydroxide. We have this' assumption that seems to be 3 ' growing up that cesium iodine is'the okay thing. Se We have heard a little bit about boric acid here i , 25 a few minutes ago. Six months ago, probably we were not too ; [ P _ _ , _ . _ - - ^ - - ' ' ' ^ ~ ~ - -
'k
. 205 1 . worried about that but-these days I think we ought to be a bit 2 more worried because we have done some experiments which show 3 that boric' acid breaks down cesium iodide to form a borate, -
4 cesium borate and free iodine. 5 So,'the; assumption of cesium iodide is, I think, 8 open to sorme question again. That is another 'one of the un-7 knowns that has to be fed into the Sandia study.
~8 'Now, the other interesting observation, once again l-l 9 I taust say that these UK figures are by no means final to retentions because we are still assessing these codes and l 11 we have not done everything we wish to them.
2 But what we are seeing here are fractional j u retentions, so that these two numbers here in each case should i f 14 add to one or there abouts. The reason that they do not add -- - 3 m let me just rephrase that. '
.18 Those three numbers should add to one. We are 17 seeing that the retention, calculated with the code so far, 18 the U.S. and UK, in fact looks very similar. But why does 18 that retention come about?
l 30 We_go back here. We are saying that cesium. iodide 1 L 21 is retained by being on particles which are on the wall ' as primarily. The U.S. is saying that it is primarily condensed _ 1 2B onto the wall, the same for'oesium hydroxide. Nothing ; se very different' there with : regard to " tellurium because SS tellurium seems to be dominated by chemical reaction. Nothing i
206 1 else much matters. 2 VOICE: Would you repeat that, please? 3 MR. BUTLAND: The tellurium is dominated by the 4 l cnemical reaction with the walls. 5 MR. ROWE: Go back a second. 6 MR. BUTLAND: You mean to this? Yes. 7 Well, the UK calculations have shown that most of 8 the cesium iodine that is retained is on particles which 9 deposit on the wall because we have .24 there while the U.S. 10 had .11. The U.S. showed that the majority of the retained 11 cesium iodide at this time in their work was by condensation 12 onto the walls directly, rather than via particles. 13 The same thing there. Now, why is that? There is 14 one curiosity, I suppose, about it -- probably two. The 15 first one is, we are using essentially the same code that 16 we got this before. Why does it not have very much material 17 effect down here? 18 Now, this seems to be a very interesting function 19 of condensation mechanisms in TRAP MELT. It arises because 20 of the different thermal hydraulics, answers for the thermal 21 hydraulics that we have obtained from the ones the U.S. have 22 obtained. 23 Essentially, the ccndensation on the particle is 24 controlled by the temperature of the particle -- pardon me -- 8 and the temperature of the particle is assumed in TRAP MELT
(_-_ 207 1 to be the temperature of the gas. Condensation on the 2 wall is controlled by the temperature of the wall. 3 We have got a certain difference between us in 4 the degree of heat transfer and for other reasons in the 5 temperature of the wall and the temperature of the particle. 6 We have essentially the same amount of heat in there but with 7 a certain change in the division. That change leads to 8 a change in the condensation on the particles which go to 8 the wall, or else on the wall. It is a fairly sensitive 10 point that we have here. 11 This seems to me to be quite an important issue 12 L cause condensation directly on the wall would seem on the 13 face of it to be more useful than condensation on a particle 14 which then deposits on the wall. That particle may go on 15 downstream. 16 I believe that Jim did tell me sometime in the 17 past that they have seen something similar to that when 18 they have done some uncertainty work which has led them to 19 change their temperatures here and there, and they have seen 20 some of this flip-over. 21 Now, what conclusions have we reached about where 22 the deposition is? Well, the cesium iodide and the cesium 23 hydroxide, most of the deposition in our mimic study has 24 occurred in the upper plenum by condensation on aerosols
# which then deposit.
208 1
.The tellurium, most of the deposition is in the 2 upper plenum by chemisorption on the walls. On the aerosol 3
particles, most of the deposition is again on the upper 4 plenum for this AB Haltleg via gravitational settling. 5 So, the upper plenum is important'and yet, we do 6 not seem to know toc much about it. For two reasons we 7 do not know too much about its thermal hydraulics and we need 8 to know more about its available surfaces for deposition. 8 Again, I think we have touched on that point earlier , 10 This is just illustrating that point. Mass of 11 cesium iodide in the upper part of the upper plenum is a 12 function of time, and we see it rising rapidly in the state 13 of deposited on particles which are on the wall. 14 This one here is on particles which are in the 15 steam, and so the gravity falls off within the upper plenum 16 in that state. 17 MR. ROWE: I have a question. 18 Why are the particles cold enough to have 18 condensation on them, because they are on a cold wall or 20 is there something else that is happening? 21 MR. BUTLAND: Why are they cold enough? 22 MR. ROWE: Why are they cold enough to have
- condensation.
24 MR. BUTLAND: Well, presumably because the temperatu re,
- they have the temperature of the steam and the temperature of
9 209 1 the steam in the upper part of the upper plenum is low enough 2 for this to happen. That is the only answer I can give.
.3 Mike is nodding h'is head, .he hasLthe same answer.
! 4: But I do not wish to be categorical in this. I am pointing 5 up here'an issue which I do not think has been pointed out e before which is that there is some debate between whether.
- 7 we condense on the particles or directly on the wall.
8 This does not seem to me to be a trivial point 9 because particles can be swept on. 10 MR. ROWE: What is the wall temperature compared i 11 to the particle temperature? 3 MR. BUTLAND: Well, you have to realize that in 13 TRAP MELT,.the model for condensation on the particles and I 14 on the walls is a function of the concentration -- well, let's 2 just take the one for condensation on the particles. to It is a function of the concentration of the i 17 cesium iodide in the gas, the concentration of the cesium
- 18 iodide at the temperature of the particle , and it has 19 a multiplying factor which is in TRAP MELT.
30 That factor is very large. So, essentially one 21 could say that the TRAP MELT model 'for rate of deposition El down onto a particle is a factor times the difference of El two concentrations. One is the concentration at the particle Se temperature and the other is the concentration actually in l 25 the gas.
7
~
211 1 straight forward. 2 MR. DUNBAR: ' Ian Dunbar, UKAEA. 3 I have two questions, one for Terry. By wall, do 4 you include _ floor, i.e., horizontal surfaces? And if so, l
- l. 5 do you include gravitational sedimentation?
-8 Do you take account of whether that floor is hot 7 enough to revaporize cesium iodide?
8 MR. BUTLAND: Well, in principle you can. You.put 8 in an area for the deposition onto the wall, so that wall 10 can be whatever you want it to be. 11 You do account for revaporization from the wall. M MR. DUNBAR: I'think my second question is probably l 13 directed to the Battelle people. l 14 Is it known that the amount of' latent heat or 18 at .least by condensation onto the particles, is it known i 18 that does not affectively increase the temperature of the 17 particles? I 18 MR. KUHLMAN: It is assumed that it does not. The 18 relaxation time -- well, maybe you have looked at this more 38 than we. But I think we have looked at the heat-up of 21 particles, for instance, by decay heating. The relaxation as time is very fast'for temperature equilibration. El There may be some minor temperature differences, 88 but it does not seem to me like they could be very as significant, if at all significant. m_ ._. _ _ _ _ . _ _ _ _ _ _ _ . _ _ . _ _ _ __ _ . _ _ _ _.____.._____.______._____.____m.__________.___________...__-m_.____.____.___.__.-_._____._m _ _ _ _ _ _ _ . _ _ _ _ . . - _ _ _ _ . _ . _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ . _ . _ _ . _ _ . _ _ . _ _ _
r q 212 1 MR.'BUTLAND: There-is one point I'just make to 2' one of the gentlemen in the front row who asked the question 3 .just now. 4 ! Whilst we have-the difference between the ,UK and 5 l U.S. results here, you asked if we are still in fact.getting l 4 condensation onto the particles but not as much. Was not ! 7 that where they had zero and we had none? 8 Yes,. Michael, i: 8 MR. KUHLMAN: Mike KuhIman-from Battelle. 10 I am wondering whether the geometry used for the 11 upper plenum was any different from what we had used. 12 MR. BUTLAND: The answer to that is, no. 13 i., (Laughter) 14 MR. REYNOLDS: May I ask a question? Why did you I 15 say that the latent heat-does not -- I am asking the question 14 over here. II ! MR. BUTLAND: Sorry, i 18 l MR. REYNOLDS: Why did you say the latent I l 19 heat vaporization does not go into the liquid? Where does 8 it go, does it go into the gas? . l 21 MR. KUHLMAN: I did not say it did not go into j 8 the liquid, it will go into the liquid. But I think that the 8-l heat transfer away from the particles because they are so j 24 much small and so much surface area to match, that the
# relaxation time for the equilibration of temperatures is s
a
313 1 very short. I do not think it would maintain a very large 2 temperature difference. There may be some. It is probably 3 something that should be calculated. It .is calculable. 4 MR. BUTLAND: There is of course one other point 5 about the particles and that is that the particles are 6 structural materials with fission products, and the fission 7 products are giving off decay heat. That is not modeled. 8 So, the modeling of decay heat is a complicated 9 issue. Not only is it going to be depositing heat on the 10 walls but we have to consider what happens to temperatures 11 or particles which are being heated up because of decay 12 heat. 13 I just wanted to put out these little graphs just 14 to show one other thing. We talk a lot about deposition 2 velocities and the main deposition mechanisms for aerosol 16 particles in the present TRAP MELT model turn out to be 17 thermophoresis and gravitational settling. 18 The deposition velocity expressed here is in 18 centimeters per second. You see sort of figures like .1, and 20 taen it falls away fairly rapidly with time. 21 If you look at the gravitational settling, the 22 previous one was enormophoresis. If you look at the 23 gravitational settling one, you have figures which are an 24 order of magnitude dif ferent af ter we have had suf ficient 26 time to glomerate. So, at the moment we are showing that the
gp- .i.. 1 - s I 214
-1_ . gravitational 1 settling is clearly the dominant one.
2 -Now, we have been thinking a little bit about 3 , uncertainties in this code-setup,.and we have heard a little 4 but about them'today. I_just wanted to.say'one critwo words r s before I put' up a few slides on the topic which are. by no 6 means' complete. 7 First of all, we have uncertainties in the i . . l 3 - inventory. They are not' going-to.be significant, I don't 9 think,-and they don't affect retentions because retentions - to -are not a function.of -- in terms of fractional retentions,
- 11 , - that is. But of course they will just multiply all the l
! 12 way through. 13 But it does seem to me that it might: benefit us l l
- 14 if we were able to do some better. control benchmark compari-15 sons of UK FISPIN and the ORIGEN code than 'I have been able to do, to try and fully understand some of the dif'ference s 16 i
17 we are seeing. 18 When it comes to MARCH, how do you look at the is uncertainties that arise from a large code like MARCH 7 It 20 is not my contention that one gets hold of MARCH and twiddles
' 21 . every number that goes into the front'end because you will 22 spend your lifetime, probably, doing that.
- 23 You have to ask yourself what really affects the l
24 ptoblem at hand. 35 Now, we are looking here just at retention in r I > .
'e
9 l < 210' u i
- 1 JNow, if you'thenLget that1 difference in those-two concentrations ' as . positive, ?it is very rapidly . multiplied
~
2 i up'byDthis factor. .So,Jyou get a massive _ sweep-down onto
~
3 L 4 'tdie particle very . rapidly. :That.is the model in TRAP MELT. ' 8 What'is the model for the condensation ento the a wall?' ~It Lis essentially (dominated' by the concentration of
~
L 6
- 7 the cesium iodide" vapor.in the gas, multipl'ied by-another
- - 81 ifactorEwhich turns out to be much smaller.
t 8- lSo,1the rate of. going down onto the wall' turns 1EL- out to be slower tha'n the' rate of going ~down onto the
-11 particles if'there is~a' positive rateLto go down_onto the
~
-M particles. You could.have a negative rate to come off the is - particles.
14 So, it seems to be.'not dominatedLby this numerical
. situation. Now', whether-this is right requires looking at.
~
18 16 MR. ROWE:. It is-concentration driven rather than 17 temperature driven.. 18 MR. BUTLAND t. Do you want to'say anything on this, , .18 Jim? l 20 MR. GIESEKE: Well,_the concentration is-the vapor 21 pressure or partial pressure driven. It is a standard mass-
# -transfer problem.. The multiplying factors are the mass as transfer rates to the: particles and the surface area of the se particles, similarly for_the wall, the surface area of the 25 wall.and the mass transfer rate to the wall. It is ve ry g
l 1 t 4 t
215
-I primary system in my little talk.
, There are one or two 2
effects, one or two conclusions you can draw from that. The 3- core-temperatures will.obviously a'ffect_the release.and'so 4-
- the-core temperatures are critical, and the radio power---
5 l . profile therefore - figures .as to 'other things . 6 The core exit gas temperature drives the thermal 7 hydraulics in the primary _ system. So, that is therefore
.a critical.
The.;modeling of core slumping adopted is important, 10 as we have heard today, -and the question of whether -MARCH ' II
'is essentially a 1D'model'in terms of -- because of the fact 12 that steam going up through-the core can't'come back down 13 l
.again.
14 Are there any recirculating-flow issues? Those 15 are-some thoughts on that. , 16 On MERGE the upper plenum structure is clearly a II ! key area, it would seem to be from my short study but it 18 emerges from others. . So, we have issues like masses of 18
, . material, head transfer coef ficients , circulating flows.
1 8' You can look at-little at circulating flows by 9 21 saying'that if the flows are circulating, presumably that, 8 ; gas is giving up more temperature, more heat, to the structure s. , 8 So,-maybe one can just have a. quick first stab at it by- 4 24 changing things like heat transfer coefficients so as to get 8 more heat in. y y - ,
..r-
w 216 1 But that won't do everything because recirculating 2 flows will give you more time for a glomeration. So, 3 presumably it would enhance gravitational settling. 4 In the early studies by Battelle and also in 5 this little mimic that we have done, the upper grid plate 6 structure melts. It therefore is not a depositoin area for 7 aerosol deposition from above. It may also block flows 8 coming out from below. 8 But then, of course, the calculation does not 10 emerge, does not model radiative heat transfer from the 11 upper grid plade to structures above. So, in fact the 12 upper grid plate may not in fact melt. 13 So, that seems to be something to think about. 14 Then there is the question of fission product 15 heat-up which we have yet to look at thoroughly. I do not 16 know whether this is in effect, but all the codes assume 17 one number for the specific heat, where as in fact we know 18 it to be temperature-dependent. Now, that probably needs to be 19 improved. 20 In the CORSOR area, the release rate recommendations 21 are obviously-areas of large uncertainty. 22 Then, in the TRAP MELT area the chemical species 23 itself, and I mentioned something about that, the effect of 24 boric acid on that. Also, in the early studies by Battelle 25 the absorber material was not modeled as forming an aerosol.
217 1 'I-now understand that they have done that. So, it will be 2 . interesting to compare notes on what the result of that is 3 .because I have done a little on that. 4 Now, coming to some of the ways we have looked 5 at' uncertainties, I have this table here for cesium iodine. 6f I have a set 'of base case numbers for' retention -- that is 7- the fractional retention-- of released material; the
's. fraction in the containment and the' fraction suspended in 9 the primary system. Those three numbers'should add to 10 one in each case.
11 Then, the retention is made up as we discussed,
.l M by those mechanisms. l 13 - Now,:this base case that I did was in fact a 14 two-volume upper plenum ~ structure arrangement. There is
'15 -an1 1ssue of how many~ control volumes you should have, also 16 as an uncertainty issue. e 17 So,'we have just made a little stab in-that 18 direction by having:one upper plenum volume and four upper N plenum volumes to see how this affectsithings. It does 20 not-h' ave much effect on the' retention-but it does start 21 - playing around with the amount of material in the various.
22 retained states,. particles on-the wall, condensed directly 28 to-the wall. As you can see here, one plenum enhances that 1W effect. 25 Then we'also had an effect on the -- we also. L =
J 218 1~ changed the heat transfer coefficient to the wall as a way [ 2 of -possibly . thinking about recirculation because the wall l' 3- _might heatlup more. I made this gross, probably gross change, 4L .of a factor of ten. -That heated the wall up to the extent 5 that noth'ing went onto the wall and it all went onto the i 8' particle. E 7: -Then, the control rod effect, once again we end 8' up withia'very similar retention even there. But then the 8 -controlirod effect, the absorber rod effect, if you looked at 10 tne' recommendations from the Germans on the mass. of absorber.
'11 that gight get released, they were saying that about 12
'M ' percent of the absorber would be-released as aerosol. So, that.
D'
~
in this particular case works out to increasing the~ aerosol 14 mass by a f actor of two. 15 If you assume that 'all the absorber is released, 16 then thatLincreases it by a factor of ften and 'the cesium-
- 17 iodide retentions:then'startfgoing up to halfiof the cesium L l- -2 iodide being retained in this AB Haltleg sequence which of N -course is-a short route to the containment.
2 .That is in; fact' changing the chemisorption 21 Ldeposition_ velocity for. tellurium,:and.the only reason I 6 have included it-is to make:sure there is not some.' sort of
~
y
'E
. feedback mechanism.affecting cesium. iodine,in the code.-
88 :Now, in. terms;of retention,':you'end up with the *
+
5 same observations for' cesium hydroxide as for cesium iodide at 7
~.
4 -@ T # 4 -
s 219
- 1. .the moment, but that is probably not quite correct in fact 2 because we believe that the chemisorption deposition 3~ velocity of' cesium hydroxide is higher than is in the code 4 presently.
5- 'Now, this table is a similar table for tellurium 6 where-we are getting very high retentions of 96 percent. 7 So, playing around with volumes, control volumes that you 8 use and with heat transfer coefficients in fact has.very 9 -little effect on that because tellurium as presently-10 modeled is dominated by chemisorption, and changing the 11 aerosol mass has very little ef fect, either. M But if we just move the. tellurium deposition 13 velocity down by a factor of ten, which is within the range 14 .of the experimental data from Sandia, so moving it from 15 1 to-.1 you send tellurium retention'down from almost le everything to .70 percent retained. 17 iso, it is quite important to get that deposition 18 velocity known to a'better extent. 2 .Now, the last-one relates to the aerosol. Everythin g 2 other than the cesium iodine. and -tellurium is in the aerosol,. 21 and our baseLcase was showing 38 percent retained in this 22 AB'Haltleg. 5 Playing around-with the upper plenum control 24 volumes does'not have,a marked effect, neither does the N . heat transfer. But of course, the aerosol mass does y . y - - - _ ,. -. _ _ . . -
220 1 because you got a great deal more agglomeration. By having 2 all of the absorber released, you.get 73 percent-retention,
.3 but 73 percent of a much higher mass.
4 So, in fact you end up with a lot.more out in 5 the' containment than here. 6 Well, that is-as far as we have gone so far with our studies.in the UK, and remember that what we are doing
~
7 8 at this point in time is. evaluating the codes that we have, 9 not coming up;with a new source term recommendation. 10 We have yet to.go to considering the position of 11' deposition in the containment for some of these mimic studies. Et We have of course done a lot of containment work,.and my 13 colleage here today, Ian Dunbar, has done a lot in that area. 14 But we have not yet~ applied any containment codes 15 to any _ of these sequences which ~ you are studing here in the 16 U.S. 17 Thank you. 2- -MR. SILBERBERG: Any questions ' of Terry? 2 MR. WALKER: Just a' comment. It seems to me that 2 -the numbers you have generated still indicate for sequences
~
21 'like AD-there is not much-retention in the reactor' system, 22 [7.is --
.5 .MR. .BUTLAND: AB, of course. Did you say AD?
~
.24 MR. .. WALKER: 'Yes, AB.- lIt seems to me like whether
^ M .7 comes out of 1.'0:comes out is not very different.- So, it
&, o __ . . , , . - .- . _ _ ._ , . . __
4 ,. ,
^# %..
w 221
!seems 'like the numbers that got used in; that' SRDR-256
~
^
; - 1 j- 2= ge'nerally confirm -- it l is_~ good enougli. ,
-3 MR. BUTLAND: We.have the author here. Do you 1.4 -want to.say anything, Ian, or'not?
I
'5 ' MR. DUNBAR: -'Yes.
- -I am notsthe primary selected' '
6- -expert-on 256, but I.believe that nextLto no retention ' 7- was claimed for a hot leg break, . and that all the retention
- a. that.we claimed:was in; cold. leg breaks.
.9 - MR. SILBERBERG: Thank you very much, Terry._ It
~
10 was a very interesting presentation, I'a'ppreciate the effort. 1 11: - MR. KASTENBERG: Are'we going.to get copies of 12 ;these? 13 MR. SILBERBE,RG: Yes, probably tomorrow,-in 14 tomorrow's~round.
~
18
. I have tried--to consider what is~next on the 16 agenda and the hour, and'the-fact that we.always try to ?I 17 -- strive'for. excellence:here.
l p
'Is The best= estimate of?my judgmen't says that we-
- ought to continue'with Surry, and we will'try to maybe.go
~
19
- 20 to'something'like'a-6 o' clock time limit.or ajfall, whichever -
~ -comes first.
t
)
- 22 -
(Laughter)
-23 :
. MR. SILBERBERG:. I?thinkLit would be interesting
- 24i to ' hearf theTSurry . presentation .'because' IL think: it - helps ' us~-
" - ^ 25 - 7 pat l at' this. point; the --Sequoyah prer.entation t into perspective h 4 ,
, .w-, m 3 y , n - + . +
-#,f- , - , . -y. y , ,., wnA. , mo . . , . 3 -s,- - -
s . + .
- p. - - .
i - u= -
- 222
~
Jalso.because a: lot:of questions.were coming up about, how 1
~
does this compare back to Surry.
~
2 I think to the extent that we could use the tine 3 4 .in suchL a way -- I know you may not bef ablel to do all of it,
,j
~5- ;but at-the expense of maybe leaving something out, is there 6 a 'ponsibility .of getting -to the _. key points on each- one 7J .of the presentations that will bring us to where you are, 8 .you know, your final wrap-up?:
9 'MR. CYBULSKIS: Can we have a five-minute-break 10 so I can make a phone call? 11 - MR.-SILBERBERG: Yes, we will declare a five-minute
~
Ut . b'reak . .. u- We have all of the material for today, it!is
~
14 now in the adjoining . room- .- They are laid out in order, so-2
'if you; just want' to go through and pick them upL in ~ order 16 -
-you each can get-a copy._
J 171 (Whereupon,<at 5:20 p.m. a-fiv'-minutee recess^ N was taken.)' 19 - MR.-SILBERBERG:: LThe meeting will please come 2 t'o order.
~
21 I MR. GIESEKE:= ' Good evening, welcome.to the ! g ' 22 J evening session.- 5 . (Laughter) - 1 i 24 MR. GIESEKE:
~I ths .are' going _ tol go -on: and talk about 25 :
A -
7 323 1 Surry. It will be;the same actors as before. Peter-
'2 Cybulskis will start off with the thermal- hydraulics.
3 MR. CYBULSKIS: The. sequences that we investigated' 4 for-Surry are essentially the same sequences that were done
~5 in'the-initial; study. They are AB, a large break, loss of 6
coolant accident with loss of all electric power, which 7 1means loss of all active engineered -safety features, 8 inc1'u ding core injection and the containment safety system. 9 TMLB prime, which is a classic transient sequence 10 with all the containment safeguards failed and no makeup II to the primary system. u The S2D' sequence, which is a small break loss of 13 coolant accident with failure of the emergency core coolant 14 injection system,.but.with th'e containment _ systems operating, ! 15 and the V-sequence, which is the interfacing. systems LOCA 16 which results in loss of coolant outside the containment. 17
- At the same time defeats the active core safety systems.
18 - The key differences in-terms of what we did some 19 months ago'and.what we did this time is, one, switch from 20 MARCH 1.1 to MARCH'l.2L-- did I say 1.'2? It'must'be late:in the~ day, MARCH 2.0.
~
21 . 22 As I keep emphdsizing, there.have been some rather 2 23 E significant. changes in irlput, having to do with the (apper 24 ~ core structures, courtesh*.ofWestinghouse. Also sor 3
-s 25
. significant changes in input 11n the- lower core suppc{rt
. . .r-- _ _ . ...-- . - . .
A - e x- 224
> l'
~ structures,~and ILhave~to agree with_a couple of points that
- 2. Terry. Butland made in tierms ' of core . power" distributions.
I 3: Lthink'it'became pretty clear'in'some of our discussions on l. 4' ithe.PeachfBottom-GrandLGulf-cases'that-core power distribution
~..
s 5
'c'an have'a significant effect on your predictions.
6' l :In the~ calculations'that we are. talking about now 3 {- L 7-
~weLhave a m'uch flatter power. profile which we believe to x
8 beJm'ch more realistic..
~
f u
' 18 I' guess I would'beiremiss not saying againt that 10 ,
7
- inl terms of~theiswitch from MARCH'l.~1 to MARCH'2 we not
( lli
.only' changed the code but we als'o changed the choice of L
12 core options.' In the Surry' cases we are using what I call , 13
.the; gradual slump option ~as op~ posed to.the coherent slump 14 4
~
option;in~' earlier cases. I 15
'Let me talk about the AB sequence --
16 MR. ZUMWALTi. I .wonde'r 'if ; I.- can j ust ask a quick 17 :
-question.
8
-Looking over the numbers,. comparing Volume!I and
~
l , L b I! it looks.like the coreLinventory of iron had changed'by-
~
18 I .V, E' aJfactor'of eight'.
~
Does-that'sounduright to you?
'21
. MR. CYBULSKIS- No,Jitldoes not.'
i.
.- f 22 ' . MR. ZUMWALT: IThatDis wha.t the numbers say-there.
I
~
'E MR. CYBULSKIS:-..I am-not: exactly sure,-do you
~ .
24- - mean,the core inventory-or'the: primary. system inventory? l LILdon'tL think the core inventory-of--iron which really. r .
% ~ -
v- g=.-.e - e e-- y 9 - . - - .e,., y ,r.e , p- c-* r -r-w e-1 , -, e-- -w r V -r-w--
s, t '225 l' represents miscell'aneous metalfin:the core, I' don't believe 2 chang'ed appreciably. 3 The' amount of total metal in the" reactor vessel 4 . changed'significantly,.but I: don't recall.the exact numbers.
- 51 I don't'believe<it was'anything like a ' factor of eight.-
6 MR. ZUMWALT: I got these'from tables where I 7 ;get . inventory both of the fission products .and the non-fission 8 products. Of course,!in V'you also list.the chromium and 8- nickel,'but'just comparing the iron, there is a rather large !~ 10 factor. I- II -MR. CYBULSKIS: Are you talking'about the inventories 12- of)the~ fission products or the inventories of the structural ! 13 ' materials,'or some combination'.thereof? ! 14' MR. ZUMWALT: I am referring specifically to 15 Table-6.8 in Volume V, and the' corresponding-table in Volume I . L 16 I don't-know what that~is. s. 17 MR. CYBULSKIS: Dr. Kuhlman' tells me' that there 118 is something clearly wrong with. that table-and'the numbers - 1 18 should not be taken at their face ~value. Apparently,
'#' something j got-lost.in the translation.
- 21; There is'a: comment from'the back that apparently i
22 L lthe Krypton inventory is.~ wrong, too. Did~I~ hear correctly? 23 In the-firstLstudy[it.was wrong. .In this one,-.isfit right? 24 :
. Clearly, the're'- are some problems with that' _ table.
25 lThis.-is.just a' cartoon;or schematic in the. i:
m.c - ,
^
g 226 1 _primaryfsystem. In the AB sequence'we assume the break 2 right.about in here, so'that.the gaseous fission products
.3 come out;of_the core,: swirl.around the upper plenum or 4 whatever they do out-the hot-leg, infthe containment. A 5 very sho'rt path.
- 6. Just to give you some perspective on time scale, 7 when things happen I will: just pick-one of these and not go 8 through'all of them. But basically,-the core uncovers during 9 the initial blow-down and' they started to melt. It is 10 somewhat sensitive, to some' extent disturbingly sensitive 11 to the conditions in the containment.
12 That is because if' primary system pressure for a-13 - .large break LOCA is assumed to be equal to the containment 14 pressure which changes the density-of the. steam, which affects 15 the metal lottery action._ They start a melt and there are 16 : some'diffe'rences in the slump. But the general picture is 17 the same.
' 18 - Early start on melt, early core _ collapse. When 18L ihe_ core _goes into the bottom head, you evaporate the M residual water relatively quickly. Since there is-noi M~ ~
' pressure difference to load the head,'it. takes somewhat
~
l8 L1ongerito-fail the head than we typically see.in the 23 transient sequences. 24 LIn this~ case the cavity:isidhy, so you' start the
. 25 - attack'of the1 concrete and'then-the failure mode -- we looked q
r 227 1 at a number of failure modes, in this case we looked at the 2 AB beta sequence which is an isolation failure and the AB 3 gamma sequence, which is a hydrogen burning sequence. In 4 the early study, the earlier study, we had called it an 5 AB delta, inferring that the failure was a long-term over-6 pressure failure with the changes in the codes, changes in 7 the input in terms of how we described the reactor. 8 One other significant change, we tried to use the 9 actual concrete composition as we understood it in Surry, 10 instead of assuming it. What we are predicting is more 11 downward penetration in the concrete. E! So, what used to be the AB delta case has now 13 turned into AB epsilon, meaning that it appears based on 14 the calculations that melt-through probably will occur 15 before-over-pressure failure. 16 MR. FULLER: Ed Fuller, IDCOR. 17 What is the concrete for Surry, is it a basaltic 18 concrete? 19 MR. CYBULSKIS: ft is basically a basaltic K) concrete, as I understand it. We have a composition from -- 21 I am not exactly sure what the source of the composition was. Zt I believe it is the specimens taken out of the Surry plant 23 itself and it turns out to be basically a basaltic-type M concrete with very little carbonate in it. Mi I won't dwell too much on the core temperatures
+
7 --
-, ,g
?<
\,
- i. ,
~ ,.._ A 228 S
~ '
U _1 and-th'e5 MERGE-predicted temperatures, they are ve.ry similar 2' 'to what,we saw before for the sequoyah plant. Again, since
.3. nthe - upper plenum is the only part that is;available during
;- 41 the. retention 1of fission products, let me point out that
; . 5 we'used:acsingle volume for the upper plenum. In_this 6 .particular case:we'used three dis' tinct structures in the 7; upper plenum. 'In Sequoyah, I believe, we-had'used four'.
But basically, the .way they were . bro' ken down here
~
.s:
9 were, the first structure was-the-upper core plate and the
- 10 associated hardware.- The~secondistructure.~ consisted of the 11 forest of control rod, guide tubes and: suppott columns, and
' 12 : =tne. third structure was.the upper support structure as'well m asl the . core ' barrel._ : Earlier, Lwe had seen relatively o little 14 . heating of these fixed structures,.so we lumped them here 2 and then'they go directly'to the containment.
lif The technicolor :s'lides arelcourtesy of' my 17 . secretary, she-_got tired-of making'th'emlall the same color. 8 Moving'on,'th'is is the pressure time history in a~ .the: containment-for the beta case isolation failure-case. We 20' treated the isolation failure as-a.two-volume case.- The
- n; ifirstLvolume, dark line,= represents the. reactor containment agt;. 23- lasisuch; the dotted line represents the cafeguards. building iss- <intow'hIch"theisolationfailure-isassumedto_~takeplace'.
J
^
What IL'see-here'is[ initial pressurization of the-
~
24 25 - 'containmentadue to1 the' blow-down.; The pressure drops. very; 7 N
229
-1 quickly due to heat losses'as well as a leakage out of the
~2 : isolation failure. There is an increase in the pressure 3
- when the core slumps into the vessel head, gives you another 4 puff of steam.
'5 Then, once'the steam stops it slows down. I believe 1
6 we go-through.the vessel at about.this point. Since the 7 cavity is dry, we don't see much influence on the pressure. 8 It' continues to decay. 9 At this point the containment atmosphere becomes 10 flammable. There is a fairly sizeable hydrogen burn which
' 11 takes the pressure down~in the decays. We see very little 12 effect'on the pressure'in the~ safeguards building, there is -
13 a little blip here; it stays fairly constant. l-14 MR. FULLER: o In that last pressure pulse fr'm-the 2 hydrogen burn, does most of the hydrogen come from core 16 debris concrete attack? 17 MR. CYBULSKIS: I-believe that most of the hydrogen
' 18 actually.comes from the in-vessel phase of it. But the l 19 reason why it did not burn' earlier was because the steam 20 - concentration was too high at the, time of -- earlier in the
. 21 ' accident,.I should say,'not in any particular point in time.
- 31 As' time goes on, there is'some additional hydrogen s-
~B' .being added to>the containment butfnot at a very large rate.
24 At the~same' time, the steam concentration is decreasing due
.5, to continuing condensation on the walls.
i , e
,, - - - --w .a- -- ,-mn -
230 i 1 . The timing in this particular case is controlled 2 strictly by when the atmosphere becomes flammable. An , 3- ignition source'is obviously assumed-to be present.
-4 ~In this' case,'one-can make a pretty good case for.
5 the ' hot debris . on the . floor: being the ignition source. ,
'6 -MR. KASTENBERG: I have a question.
7' - MR.' CYBULSKIS : Certainly. 8 MR. KASTENBERG: -Pete, as you go along, could you
~
'8 -help us' refer back to your original MARCH 1 calculations and 1 10 point.out-the differences? Because we are not approaching 11 .this as though' it is the first time we looked at Surry.
U Mk. CYBULSKIS : That is a good point, Bill. 13 In'the case of the AB sequence, I-guess I alluded
'14 to-them without pointing out, the first time we did the AB 2 sequence we'did not. consider the-secondary; building or the -
16 safeguards building in_the. process. 17 So, we assumed that the leakage went directly to 18 tae outside. As I pointed ~out, we used a-~different' kind of 19 concrete in the calculation. I~ alluded.to some of.-the so : differences in the structural input. 21- - In termslof.some of the differences in the behavior from my;. thermal hydraulics viewpoint, of course there'is 23 -the different mode of core slumping that I think we' talked
~
24 about at some length. Th'e other key difference is'the timing of this
~
25 L -m___
,A ,
. 231 11 hydrogen burn. I believe that in the earlier studies'we
~
2 predicted that the hydrogen.woul'd burn at the time of 3 head failure. In this particular study the head ' failure is 4 here. Tha do not predict burning-until somewhat later.
- 5. This is a direct effect of.the changes in the 6 burning model, hydrogen burning model that have been 7
incorporated in MARCH 2 as opposed to' MARCH 1. 8 MR. REYNOLDS: Peter, I have a question. .Reynolds. 8 LMR. CYBULSKIS: Yes, i. 10 MR. REYNOLDS: I.am noticing that the differences 11 - seem to be the retention in the primary system. You stated 12 a lot of differences outside the primary system. We are not 13 to TRAP MELT yet, I guess. 14' But if I take~the releases from-the primary 15 l system'and multiply them by your old numbers, you get your 16 new numbers almost. 17 MR. CYBULSKIS: I.will let Dr. Kuhlman talk about 18 the fission-product-releases-from the primary system. 19 MR. REYNOLDS: 'My feeling right now from looking 20 - at these numbers is,-the big difference is in the primary 21 . system, whether it is the. thermal hydraulics that are doing 22 ' it or TRAP MELT is-doing it, I don't know-yet. 23 - MR. CYBULSKIS: Well,-I think_you are getting a 24 little-bit ahead of me. I.am not-the'one to answer those 8 questions. I am sure Mike will-be happy to. comment on that
232 1 point. He looks happy over there. 2 (Laughter) i 3 MR. CYBULSKIS: That was the AB beta case which 4 is the containment isolation failure, and I think I will 5 stop at that point. 6 Basically, there is nothing -- well, I think the 7 key point being the timing of the burn. 8 If we take the same sequence, going on to the 9 next, if we take the same basic sequence but assume that 10 the containment is now intact and look at what the hydrogen 11 burning would do to an intact containment, this is the type 12 of results you observe. 13 I think the time scale here is large, so you lose 14 some of the detail. Let's see, this thing was corrected. 15 Some versions of it had the wrong time scales. 16 We get the same large burn or a similar large 17 burn as we saw in the previous case, except in this case 18 the containment is intact. You have not lost any of the 19 hydrogen due to leakage. You have not lost any oxygen. 20 You get a burn that is large enough to fail containment 21 under the assumed conditions. 22 Again, with regard to differences between what 23 we saw before as opposed to what we are seeing now, we 24 saw somewhat, very similar behavior except for the timing 25 of the burn in the previous study. In the previous study, we m
n 233 1: p_redicted the burn as taking place right.at. head failure. 2' It is delayed in timecin-this particular case. 3
.The next vari'ation in the containment failure 14- : modes that we considered for.AB is theLmelt-through case, 5 melt-through and/or long-term over-pressure. To get there s basically'would' imply that either the hydrogen does not 7 ' burn or if it burns, it burns in such a fashion that you do 8 not 'get large pressure peaks which imply a number of small 9 burns, i
10 ' Then the containment pressurizes gradually and it l' l 11 is in general competition between the rate of the attack of-U the concrete and the pressurization of the containment. L 13 In this.particular setlof calculations, what I am l 14 . showing here is depressurization at-the tis.e when we would 2 . predict that you have melted through the bottom head. So, 16 this is melt-through, andsyou cansee the pressure is still 17 relatively modest. But we-have already penetrated through
- 18 a substantial amount of the concrete foundation.
- 19 This is significantly different'from what you 2 saw in the' earlier studies. For this particular composition l
21 of concrete and the input-conditions that come out of the H , analysis of.the accident sequence, we are seeing a substantial ly n. 23 'more rapid downward penetration in the concrete than we have M typically seen in the past studies. 25 :MR. FULLER: How thick is.the base mat, how-large
r - ;- ,
'4~~
~
234 1- -a melt-through before you penetrate it?
~ ^ ~2-
~
MR. CYBULSKIS: In the Surry reactor, as I understanc
~
~
3 the design,_the base mattimmediately under the vessel is
- 4 -
ten. feet'offconcrete.. In other areas it is twelve feet. 8 Outside'the. cavity, I-think there is two feet of concrete 8- above the steel liner, if I remember correctly.
~
-_. 7 So, this represents -- actually, what I used, I 8/ -taink,'are 300 cm of concrete penetration.
l 9' MR. DUNBAR: Ian Dunbar, UKAEA. 10 IsLit realistic'to have~an instantaneous de-L 11 - pressurization through the' base mat? l 12 MR. CYBULSKIS:..That is not an instantaneous l , 13 depressurization.. If you plot these things, there.is sub- !~ l . . l 14 .stantial slope to that line. But if you plot it on.the time. 2 scale that we plotted it,=you cannot really see it. 18 I believe that for. purposes of the calculation we f-17 . used the seven-square' foot hole-that we had used in the l 18' ~ previous' cases. But as :I indicated, it is not an overwhelmingl y r No significant thing. j N' one thing that we did not include in this-L ?E calculation that smight be properly included would be the 22 back pressure due to whatever static head, equivalent static
= se ~ head'that.the, ground or.the groundwater, or what.have you, o so would'present. In' fact,~it may not.depressurize-through
. as atmospheric, if you took that into account.
235 1 The other thing that is associated, obviously, 2 with the depressurization into the ground as opposed to the 3 depressurization to the atmosphere is a significant potential 4 for some decontamination factor in going through the ground. 5 But that is maybe for somebody else to discuss. 6 I will stop on the AB sequence at this point, as 7 far as the thermal hydraulics are concerned and go on to the 8 next sequence that we looked at, which is the TMLB prime 9 sequence. 10 Again, just to refresh everybody's memory, I do 11 not have the dark lines shown here, but in the TMLB prime 12 sequence we have loss of makeup to the primary system, no I 13 containment safety features, loss of makeup to the secondary 14 steam generator. 15 The first thing that happens is the steam 16 generator will boil dry. When the steam generator boils 17 dry, the primary system will start to heat up. It will heat 18 up to the set point of the relief or safety valves, boil the 19 inventory off through the presurizer into the quench tank M which looks like a beaker here, and eventually the core will 21 uncover and melt. 22 The core, as the sequence has been traditionally 23 defined is assumed to uncover and melt with the relief 24 valve continually cycling as required to hold the pressure. M Melt-down occurs under high-pressure conditions. As has been
236 r l' pointed out~at various times, including by Dr. Bari earlier 1 2 .this morning, there.are alternate scenarios.possible. There 3 could.be-the buck-seal failure possibility that could change i-4 the scenario,.or you could argue that the relief valve may 5- .or may not survive that long. 6 -I'think one must recognize those possibilities. E 7 For purposes of these calculations, we analyzed the sequence
~
a as it has been traditionally analyzed.
'8 Again, I wont dwell'on the core temperature
- 10 MERGE histories. We talked about them at some length this 11 morning. The results that we see here are very similar as 12 I believe wa described in the results.
13 I will again use my very brief schematic just to 14 _ point out for purposes of the fission product calculations,_ M this is a kind of a crude representation of what the fission 16 products see on their way out of the primary system. 17- The gaseous fission products are leaving the i 18 core, again pass through the upperfplenum which in this - h 18 case was a single structure, with the same three structures 20 that I talked about before, the core plate, the control rod 21 . guide tube forest and the fixed structures consisting of 22 the upper support' structure and core barrel. 23 -Then they pass through the piping and the 24 pressurizer which is treated as two separate structures, and 35 on'into-the containment.
237
- 1. For the TMLB' prime we are looking at two. .
.2 . containment. failure modes, an early failure and a. late failure .
-3 The-pressure temperature history in the containment that 4- LI-show on this slide is'again for delayed failure which 5 'similar to the AB sequence t'urne'd out to be, as the 6- calculations predict, .a melt-through case rather than a long-7 . term over-pressure' case.
l 8. .In this particular case, there is no containment L l 9 response until about the time of' the ~ steam generator dry-10 out,..at which'-time you' start <to lift the pressure relief 11 valve, pressurize the containment;. core melting starts some-l 12 where in here, and then the steamfrate decreases. l u Reactor vessel. failure'is at this point. You 14 release the high-pressure gases. FromLthe primary system ( 2 you also get a cavity / interaction withith'e debris. We l l 16 used some fairly nominal parametersLin this particular 17 calculation so the pressure rises.
- UB The fact that the pressure is fairly low and i
le the rise is'not excessive, youLare at about the. design ' 20 . . pressure at this point. So,-this-type of pressure pulse 21 would~be.very unlikely to" fail the containment.. L 22 Then theresis a decrease in the pressura due to-
.. # ' condensation, and-then there--is a very slow' increase as M the core-concrete interaction takes over.
~
25 'Again, at this point in time we are predicting b
t 238 11 'thatfit' penetrated through 300 cm of concrete and would c
~
2- ;de' pressurize;-- again the point that was raised earlier. If 3 i
.you take into account the back pressure and possibly 4 pressure: drop, the depressurization could be somewhat 5 different.
.6- MR. LIPINSKI:- .Lipinski from Sandia.
Pete? 7 MR. CYBULSKIS: .Yes, Ron. 8 L MR..- LIPINSKI: L It-is not on this viewgraph, but it. 9 might'have been on the viewgraph for^the timing of the l 110 sequence. I noted in the tables that the corez uncovery ? started an hour and-a-half earlier for Volume V than it did 12 fin.. Volume I.
. 13 Can you explain:where this hour and-a-half came l - 14 ' from?
2 -MR. CYBULSKIS: ~t he exact l- .I do not recollect 18 numbers. I guess I should have reviewed'. Volume I before
- 17 - coming over here. But there are'several differences 18 associated with the MARCH 2 calculations versus MARCH 1.1.
i N I am caught again'by not pointing them out. 2 Let me mention them.asithey come to mind.- Number EH .one, the decay heat correlations in MARCH 2 are updated over
~
23 those in MARCH 1.1. The-net result is of the order of a 2 percent increase in decay heat at any point in t.ime. 24 - So, .the . time scales ~for - things like stere; generator
=M- dry-out are shortened proportionately, if you will.
# 1 e r c - { -
F r 239
~l' Another. difference betweenithe MARCH 2 and
- 2. MARCH l'.1 calculations,.there is a closer coupling between 3 the primary ' system and the steam generator in terms of heat 4 -transfer coefficients.
i 5 We make.better use of the steam generator faster, 6 so that -you tend to dry .out. the steam generator quicker. 7 dowever, as long as you use the same inventory, that should 8 -really have no particular effect- on the calculation.
'9 The third difference that comes to mind -- and 10 this one I will.have to double check, I hope you will ask 11 the question again. But the MARCH boil-off calculations
- u. require aus an input break elevation or elevation of the L .
13 safety valve vis-a-vis the' elevation of-the core. ! 14 1If the break elevation is high in the system as i M MARCH defined it.-- and'the elevation is.not real, it is 16 a bookkeeping elevation -- then you will boil off predominantl y c 17 steam.- If the elevation is lower in the system, you will
. 18 boil exhaust -- or two-face liquid through the valve or ,
19 : break until.the valve or break is uncovered. l 20 In particular in the MARCH 2 calculations, what 21 we have been trying to do typically is use an elevation ' Et. which will let water flow out of the relief valve until. El- such time as the equivalent water level in the primary 24 system drops to this level, basically. So, it is-a line 25 uncovery. l l L t
240 1 That has a - tendency, ' by ' squirting liquid out the s 2 break as opposed to having to literally convert the steam and-3 boil it off, you tend to shorten the time to core uncovery 4 compared-to an elevated break which is steam. L 5- MR. RAHN: Question. 6 MR. CYBULSKIS: Yes.. . L 7 MR. RAHN:- Frank Rahn, EPRI. 8- You mentioned that the MARCH 2.0 version has 9 . decay-heat about 20 percent higher than Version 1. It l 10 seems to me if anything,.that the recent ANS standard is 11 -20 percent lower'than the previous one. 12 MR. CYBULSKIS: The recent ANS standard is precisely
, 13 what is programmed in MARCH 2.
14 MR.'RAHN: Well, the new standard, the ANS 5'1 15 standard, is lower than the previous standard by about ten 16 percent on the average, and some places it is 20 percent. 17 MR. CYBULSKIS: As I said before, the ANS 5.1 -- 18
.I forget the exact nomenclature -- is precisely what is 19 modeled in MARCH 2. A very simplified correlation for decay 20 heat was included in MARCH 1.1, which did not include heavy 21 element decay and some^other things.
22 MR. RIZTMAN: Ritzman. I think I want to make
~
23 a comment here'just for information input. We have, as 24 part of our work for EPRI, been using MARCH 2.0 for Surry 25 calculations also, and specifically the sequence.
r - - 241 1 We put the break,high in the system, and we get 2' . core uncovery at about'130 minutes, 30'to 40 minutes down-3 -stream a year. That may be part of the reason -- since you ' 4 talked about your break elevation, I-think that is probably 5 the reason for-that difference. I was puzzled by that. 6 As far as the containment pressure trace with time, 7 our'results look similar. We do not get quite as high a ' Is -pressure spike at vessel melt-through as your calculat' ion. 8' We,are somewhat over'40 psi. 10
-I think that may be differences in assumptions
-11 about water in the cavity, or an accumulato.r dumping rate, 12 I don't know.
U MR.'CYBULSKIS: Well, I think what happens at 14 vessel. failure is a function, there are two things that 15 ' ha ppen .- 16 one, you release the inventory.of high-pressure 17 steams. I believe that-gives you the initial spike, and 18 then-you get the interaction of the debris with water. That 18 tends to be a function of the choice of models that you happen
~
20 ~ to use, and.there are a number of them in. MARCH. 21 MR..RITZMAN: Right. E MR.iCYBULSKIS: Whether you use a simple quench 23 model:or some combination of the particulate quench model 24 -in the' debris bed, you can even bypass them directly and 25 _ eliminate the spike, t
a . 4 242 y 1 So, that is a very_ sensitive assumption of 2" input, parameters in your choice of'models. I q,3 'MR..RITZMAN: Right. And going on, I can't 4 remember'theLprecise' time but. base mat penetration, using.
- ?
is
. again.this Surry-type concrete, our times were-comparable .
6 for those, within 30 minutes, again. 1 7
- That 30-minute difference. evidently is this 8
assumption of_the elevation of the break in the input to 8
' boil. 7 ,
10 MR. CYBULSKIS': A lot of our insights -- if I can 11
. ' use that term'-- or a lot of our choice of assumptions 12' about the elevation of the break is based on a number of 13
. studies that were.done related to the TMI as well as 14 subsequently with codes like RELAP and TRAC, which basically M
p - say that for some period'of_ time you will get liquid flow 16 or two-faced flow out _of? the break, and then they will 17 switch to the steam flow when the surge line uncovers. That i j 18 ~ is basically what we are trying to incorporate in the I 8 MARCH . codes . i 20 I don't happen to have a graph for the early 21 failure case but, as I pointed out a moment ago in response 22 to Bob 's question, you can get dif ferent magnitude' of i 23 - spikes, and in particular.if you go to smaller particles se sizes, this number can go up. as The pressures that we are-typically see'ing in.the ________m . -._m _s_ . - - _ ____m._-__._____.___-____.______.__-__m_ _ _ _ _ m_ _ _ _ _ _ _ ________m..___.-__--____.-___.__.__...m.___-_____m.._m____ A___--_.__ m - - _ _ _ _ _.
-243
'1- ' current set of calculation for.the magnitude of this spike 2- ,are not-as large as we sawLin the earlier set of calculations.
3 The.principalJdifference'for that is not.a modeling. 4 'enange-but it is'a change in input. Based 'on our discussions 5 with the reactor vendor -- and I mn not? totally sure - that 6 we understand the design as well" as we might -- we do not 7- 'have. anywhere near as much structural material below the
~
8 core as we had assumed in some of our earlier calculations.
'8 So that' when' the reactor vessel- f ails , there is
. 10 not as-much' material interacting 1with the. core debris and 11-the-temperature of the material is roughly the same as it s
12 was before~. 13 So, the, pressure spikes diat one sees are not-14
.quite as large as one saw'before.
15 MR. GIESEKE: Pete, you mean stainless steel. 16
, MR. CYBULSKIS: Stainless-. steel structural material, 17 as well as the head.
18 MR. BARY: Bob Bari, Brookhaven. 19 Correct me if I am wrong, Pete, but the fact that 8
, you consider - the early over-pressurization -for Surry, 21 TMLB prime delta and.not for Zion which you consider in anothe r 8 volume -- I guess that will be discussed tomorrow -- one 23 should not draw any . inferences, as I understand it, about 8' tMe relatively' likelihoods of the early over-pressurization 8 containment l failure, dion verus surry.
k
~
244
'l That is, for-the purposes of this activity you 2
could have looked-at.an earler pressurization failure of
'3 Ziona, rather than Surry; is that correct?
4 MR. CYBULSKIS: That is more or less correct, but L :5 not completely. 6-MR. BARI: Tell me where it is wrong. 7 MR. CYBULSKIS: Well, it is more a matter of 8 degree than-substance. 8 6 p.m. ~ In the reactor safety study, we had identified 10 TMLB prime delta -- actually back then it was gamma / delta, II I believe -- as one of the key sequences that turned out 12 to be one of the dominant sequences for the Surry reactor. 13 That conclusion was reached on the basis of two I' things. A hundred psi failure pressure with some distribution is about it and the type of modelings that are used here. 16 We, for the purposes of.this calculation, we are 17 still using a hundred psi ~ failure pressure, subject to 18 l previously-mentioned activities which may change that. , 18 So, we are more or less using that same framework
" as a point of reference.
21 ' In the. case of the Zion plant, we are using the
" -149 psi failure pressure that the Zion probalistic assessment E
~ used at face vaLue. In that context, our pressures are 24 not'anywhere near as close to the failure pressure as they
" are'here. So, the two situations-are not.quite comparable.
~ - 245
'l IBut.you are right in the sense that we really are
.2 not'saying that early containment pressure will take place or
~
E
.will not_take place, Basically, what we are saying is that 4
'there may..be some likelihood,-based.on past studies and'if 5- itIdoes, here is_what~happens.
~
6 Did I.'a'nswer the question? 7 MR. BARI: Well, -I guess it is a question of doing 8 the' containment analyses on a constitent basis,.which.is
'8
.~ advertised in the.early' sections of this volume.
-q.
10 f I guess what you are saying.is that the containment 11 - strength analysis so far has'not been done on a consistent
^
= 12 : : basis, but takenat face value from different studies at 13 different' times.
14
.MR. CYBULSKIS: That is certainly true,-the 15 :
containment---strength analyses have not been done on a 16 consistent ~ basis, andfit is not part of the scope of what 17 ~ we are doing'as part of this exercise to re-do the containment 1' 8 strength analyses.
-18
-There is'a group that is working on it, but I think.to some extent we,are striving in the case of the things 21; ~
.t: hat--I am t lking'about here, we'are striving'to maintain.
22-
.the comparison with what we did earlier.
23 - yg.I did not ! talk'about early containment failure
'N in-TMLB delta, I suspect.that I;would get a question or two 25 I saying,j" Hey,}what happened?" So, for purposes of comparison
_ms
L G ,4. 2 246
'l
;I-have:kept:the'early1 failure mode as,a sample calculation
- 2. in the Surry calculation.
!3" ~
Let.me.go on to the next sequence that we looked
-4 .at Surry, and that'is the V sequence... I do not happen'to 5 have a' cartoon that shows what happens, for which I must
~
~
apologize.. But let-me try to describe this. 7: The V sequence,f.the.:so-called interfacing LOCA. 8 sequence, assumes failure of the check valves that separate
'9 'theflow pressure emergency' core cooling system from the 10 high. pressure' primary system.
11 - If'that.happens, it is assumed lthat the low 12 pressure emergency. core cooling system cannot take the 13 - high pressure and it ruptures.- 14 The rupture, while it'cenceptually,-I think, can M ,take place inside the-containment astwell as outside the 16 containment, it has been typically-assumed to be outside 17 the containment, which r'esults not only in containment-18-bypassoas far as the blow-down'and. fission product release lit is concerned, but also defeating the emergency core cooling 20 . injection system-. 21 so, you-get blow-down outside the containment. In 22 'this particular case accumulators were dumped,.their 23 inventory covered the core, but there is no active emergency-24
' core coolant injection, so the core uncovers and eventually 25 melts.
247-1 -The~ steam fission products goithrough the piping
-- 2 outside the containment, are released to the safeguards 3 -- building and' eventually to the' environment.
~
14- Correct me iflI am worng here, Jim, but my 5 .. memory failsfor the_ moment. We did'look at the safeguards 6 building;the.first-time around'for the V sequence for Surry.
- 7 :Though infthis'particular' calculation I..think we have
- 8 .tried, based on conversations _in particular with Stone &
9- 1 Webster, I~think we have.tried to hopefully model the 10 ' structures.in'the buildings better than we did the first 11 time,'but. basically the-analysis is very similar. 12 The blow-down'and the melt release take place 13 directly to the safeguards building.. The core melts,
- 14. fission products are released through that piping into ul .the safeguards bui1 ding. Then,-when the head melts through, 16 - -the fission products are released to-the containment.
17 ' Again'back to my schematic cartoons as far as 18 the primary system, for the V sequence we go from the core 19 to'the upper plenum. The same~modeling of the upper plenum 20 that -I - talked about with : the same three structures . 21 - The piping leads from' the primary. system outside 22 the-containment. Actually, we have~run two separate cases. 23 The base case ~or the reference case is one where this long run of pipe is modeled as a single volume; in another case
~
~ 24 25 - thatEwefare not quite done with-yet, we have broken up that
~ ~
Rg - mye. ~T ' + --- , 248 4 ix . fil : pipe into'fourivolumes connected ~in series, and'then the-
; . 2.
release goe's toLthe safeguards-building, and from the air
-8
~
- , isafeguards> building..to the environment.
@n ,
;4 I
/ MR. DUNBAR: Ian Dunbar,;UKAEA.
-5 Is: this -~now bypassing , the ' steam generated ,_ and is -
8 [
- that afdifference'from Volume.I?
'7
'MR'.' CYBULSKIS: :That is a-good question.
8 Ildo not.'believe we used the steam; generators in n 8- the/line'because the: connection of the piping--is between
'II the.ste'am generators.'and.the reactor.
~ II.' Rich,- L do . you recallhow we created that in the . 121
'first volume?
13 MR. DENNING: ' There are possibilities'for inter-
' 14 -
ifacing LOCAs at different places. But in the first volume is we; definitely had it downstream of'the steam generators. So, i- 16'
.it did go through the steam generator.
[ : 17 I am not sure whether this:is~ incorrect or whether ils _ :we analyzed the'different one'here. 18 ' MR.-CYBULSKIS: I have been informed that one
# has an option of either going:through;the steam' generators 2r l =or not' going through;the; steam'. generators,~ depending on the 22 assumptions you'make,.I believe.
23
'MR.-RITZMAN: : Peter,:Ritzman'over here.
.24 Now, in this sequence wefdolhave a major difference M '
Lbecause when we analyzed..this.' sequence, we-.have HPI t_~
y ,.- ; ; 249 4-1 injection and core uncovery not occurring until like two 2
--hours downstream.
3 The low pressure pumps are assumed to actuate 4 and start and help pump the RWST"into the. safeguards area. 5 But depending on what assumptions you make about pumping 6
..capac ities, the RWST'will last from a half hour to'an hour.
'7 --l Then maybe it is differences-in break characteristics, 8 But our calculation here, we do not have I don't know what.
8 co're uncovery occurring about 140 minutes.
'10 MR. CYBULSKIS: I think you pointed to the key 11 : difference in assumptions, Bob. We did not' consider the 12
. actuation of the high pressure systems in our analysis, and 13 just the dump of the accumulators.
14 I believe as you apparently.did, you made.the is assumption that the HPI system is available~and in fact is 18
' actuated --
17 MR. RITZMAN:. Why would'it not be? 18 MR.ECYBULSKIS: There may be others better able 18 to comment, but what actually.be your HPI? 3 MR. RITZMAN: Low primary system pressure. Loss . 21
.of level in-the pressurizer.
22 MR. CYBULSKIS: Well, I guess I am really asking 23 the question, I don't really know what the correct answer is. 24 Is that sufficient? 25 ~ Perhaps the appropriate thing is to consider the _. . -.~ -
~
~
x + - - f-
,;7 3: ,
; , .;) ..
+
?., _
~
( -250
~
'* i-
, ~; t ; actuation of.the-HPI. We1did not include that in the.analysic. ..
' L.{ -
- 2' MR. RITZMAN: I^think we also' assumed'you would 3' -have. power availablefin the sequence', and you would also be
~
4 'ablee to maintain.'feedwater flow?to the steam generators,.too.
~
~5. EMR. CYBULSKIS: I guess by_ definition the power is- -
~
' 8 -' D availatble . Now much could the steam' generators do'you is
~
7- a'-function'of~the break sizesyou associate with it.
~
8; ILthink.'for ansix-inch-break this primary system 4 9- :would depressurizefso quickly'that the. steam' generators would not'do you any-_ good.
~
'10 5 11 .Iflyou cut back on the break ~ size,-turn it into-12 something like a small break;-- the point thatLwas' raise'd, t
13 - . I believe,- -. earlier 'in the L day !as'a possible option -- then ;
~
14 you would-hold the primary system pressure up and-in fact-15 - the' steam generators could'do you-;a lot of-good. 16 - .MR. RITZMAN: - I don't know!what you mean, a small
. 17 break'. We tried toiestimate an effective. break size by
~
18 -taking'into account! friction losses along that long pipe.
- 19 I-think we ended.up using'a break size something like'four
' 20 .and-a-half inch' equivalent rather than the full'.six' inches.
- 21 That helps a little , _ maybe ~. .
-22 'MR. CYBULSKIS:-. I'think the key thing, at-least. -
;28- 'as ~ far :as c the steam generator is_ how-: fast the system
- 24 : depressurized.- -
- 26 The more important ' thing and, perhaps an ~ oversight L
i a . . - - - _ _ _ _ _ - _ . . = - . - - - - . = ---------l-- ------~.-.-.--..-------------------.-----------------------------------------------------"--------
^
rv ~
%- =- -
-251 1 on-our part is the possible effect of the HPI system. If 2- .you haveLthe.HPI system operating.at full bore, I would
- 3 expect-that it~would' delay the onset of core uncovery and
- 4. Emeltingosignificantly.
.5 I think that is perhaps something that might be.
6- worth rethinking before we wrap _this thing up. 7 Insofar as the fission product analyses have a not been completed-'for.this sequence, that might be something 9 .that might be worth iterating on before we wrap it up. 1(t Let me go on to the last sequence.that we looked J 11 - at with Surry, and that is S2D -- if-I can find my.right ' ut- ~ cartoon again, schematic of-the primary system. ul Basically, we: assumed'the break in here, so that 14 when you get.to the point on core uncovery, the fission 15 products have to'go from the: core into the upper plenum, ni .through the piping, through-the steam. generator, and out-the
~ 17' break.
ul For the modeling of the. details of the primary
.19 system as.far as the MERGE TRAP MELT calculations, again 20- core, upper plenum, the same as I. described before several 21 l times,; piping, steam generator,-~and then on into the 22 ' : containment.
~
23 -Now, in the-S2D,.let-me throw these on, this'is
. 24 - -somewhat different.than some of the things we have seen.
15 ~This is more or-less typical of what we are seeing for the l
l
. i
_ 252-
~
- ~ small break cases. I. won't dwell on the details. .The only
; thing;to point out here is, this is aEsmall break' loss of' 21 -
!- 3 'Jcoolant accident. You have-continuous blow-down from the
- 4. _ primary) system. So,ethe pressure:is continually' dropping.
v 5 _If the core uncovers, it starts to heat up. At 5 6-- :some' point;the pressure ~ drops' low-enough that-the accumulators 7 discharge, generate more steam.and repressurize'. z8 LYou go:through this: cycle a couple.of times.until-9- the accumulators run out of. water and:you go on in the core L10' : melting.
'I:think again, to-touch on the differences between-
!11 12 3 ' MARCH 2 and MARCH 1.1, there is a change in the primary-
- 2 . system breakEflow model that has'been incorporated in MARCH 2, 14- which'has the net effect of passing more materia 1'through
- 15. - the same break ize compared to what I used in MARCH 1.1.
16 ' Typically, in aHMARCH 1.1 for a two-i'nch break 17 which'is'this, the pressures would stay,.the primary system 18 pressures would stay above the' accumulator set point and
'N we would not get this type' of behavior.,
;E
~
.Using MARCH:1.2 for the same break size but
-M' 'usinglthe new break. flow models, in fact we get into the 22' accumulator discharge type of. situation,which-delays the
'M : onset of core melting.
24 " MR~. ROWE: Question. 25 . MR. CYBULSKIS: Yes.' 6
253 1 lYou show'the core. temperature' dropping
~
MR.;ROWE:
~
2' rapidly when the accumulator water is injected. Is that
?3 realistic'given the temperatures that you have?
.-4' Itriooks- like you have to go through a cool-down 3L on the quench.
6" MR. CYBULSKIS: -There is a cool-down model. Again, 7
.the difference if I did the same calculation with MARCH 1.1 8 ~
f it would be a very abrupt drop in temperature. 8
'There is a cool-down model and there is more 10 than -- again, it is a. fairly long time' scale. There-is II'
- suostantial slope associated with that cool-down. It does
. 12 take time and there are models that try-to take into account 13 the heat transfer regimes that may be appropriate for that 14
[ cool-down. 15
.Again, I won't say that this particular' rate !
. I 16 is sacred, but if the core uncovers, unless you have large I
melt fractions, I think it is reasonable to assume that is
- the stuff can be quenched.
18 But it is.a time-dependent calculation that-
. includes considerations of surface heat transfer coefficients ,
MR. ROWE: So, you have a film -- feature of the regime? _MR. CYBULSKIS: Plus radiation, I'believe that 24' cdiculates it. And then,-in some of these cases you get 25 the core ~ completely covered so that unless you really lose-
254 1- geometry, fit is difficult to see that you would not 2 2 quench. 3 The pressure time history in the containment for 41 .the S2D sequence, there is nothing overly dramatic here. I 5 might point-out -- if it needs pointing out -- that the s' -containment sprays are in fact oiperating in this. sequence 7 which keeps the pressures'relatively low. I believe this 8 Lis the time of head failure-where you get -- I think there 9- 'is a hydrogen burn thrown.in with the discharge of the steam-10 - here.. 11 So, you get a pressure excursion but it is well ut b elow the ' containment' failure pressure or well below the 13 - containment design pressure,.I am sor f. So, there would 14 ' not be:much. question of failure at this point in time. 15 We.are looking at another case which is very 16 similar to what we had done in the earlier study, where we 17- assumed a large hydrogen burn. Again, that is a matter of
.18 when does the ignition ~take place and what has happened 19 since then.
20 You can get significantly higher burn pressures 21 : than is indicated there. n I believe that I will stop at this point and see i 25 if there are any questions on- the thermal. hydraulics. i ! 24 MR. FULLER: Just one on'the S2D. You label it l 25 ' S2D/ epsilon, but it does not show any indication in this u
255 1: figure of any base mat melt-through. 2 Does this in fact happen? Do you.have coolable 3 debris or~what? 4 MR. CYBULSKIS: I am sorry. For purposes of
-5' trying not to keep everybody.up too late, I went perhaps s :a little faster than'I should.
7 For the S2D case in.Surry, as I said, the sprays 8 are on.. The sprays contain the containment cooler so there
.9 is~a continuing heat source.
10
'Again, as I understand the Surry layout, some
-11
.of the spray water can find its way into the reactor U
. cavity, even though the cavity and the sump are normally M' changed. But the spray water can tricale down_past the 14 vessel.
15 If in fact you get a significant amount of water is accumulated in the reactor ~ cavity and it keeps coming back 17 in, there is a distinct possibility that'you may reach a I
.18 benign ending to the accident-for this sequence. , 19 To go back to my containment and event trees, 20 there are various combinations that will lead that. The l 21 epsilon case is predicated on not'having a coolable bed, t
22 so that you would get attack on the concrete and eventually 23 . melt-through. 24 But again, there is a possibility that you could as have a coolable bed, and that would be a termination of the t'
f. 356 1 accidpt. 2 MR. RITZMAN: Just one question, Peter. 3 MR. CYBULSKIS: Yes, Bob, 4 MR. RITZMAN: On these rod temperature plots, 5 could you explain to me what that notation means, " Rod 6 6,1?" 7 MR. CYBULSKSI: Rod? I am sorry. I should have 8 " Rod 6,1," the first number is the axial node number, the 9 second number is the radial node number 6,1 means the sixth 10 node in the first region. 11 MR. RITZMAN: From the bottom. 12 MR. CYBULSKIS: From the bottom. 13 Sometimes it is diffiruit to decide a priori which 14 node you want to save on the plot tapes. A lot of information 15 could be saved and you can't save everything. Sometimes 16 we make better guesses as to what to save than other times. 17 Sometimes the several temperatures are very similar, which 18 may or may not tell you much. 19 But that is the notation. 20 Everybody must be getting tired. 21 MR. SILBERBERG: Thank you, Pete. H Before adjourning, I have one announcement and 23 a request. 24 I understand that we can leave material here 25 tonight, that this room is apparently safeguarded in some
- r. - -
3 q , 357
;1. .with,a lock.- So, the hotel assures us that we'could leave ;
-2 thatihere.-
?-
.3 -Secondly, considering where we are right now
'4 and.:in our. review process, what we have heard today and T5 of course projecting what we will here tomorrow, and l
GL knowing that we have-the report'from Oak Ridge to deal i
'7- with when it comes in, and the fact that we have had some 8L 'new results today that are very fresh and need some 28 ' digestion; considering.the varieties of these things, I 10 would like you to.give some consideration to where we are L 11 in-the review process.
N We had intended to make this the wrap-up review N meeting. 'I would like to get some advice from'the Review 14 Group tomorrow -- you might think about it tonight -- as 15 to the advisability. and the benefit to ourselves and to you l 16 - for-the. possibility of another meeting. 17 I just wanted you to give some consideration to
~
N
~
it and perhaps I could get a sense of the Review Group W tomorrow when I bring the question up. 2 Thank you. Good night. We will see you tomorrow 21 ~at 8:30. 21 (Whereupon, at 6:25 p.m., the meeting of the Peer E Review Committee was adjourned, to reconvene at~8:30 a.m., se . Thursday, October 13, 1983.) 5
!}}