ML20126A926

Transcript of 921216 ACNW Working Group Meeting on Performance Assessment for HLW Mgt Program in Bethesda,Md. Pp 1-323

ML20126A926
Person / Time
Issue date:	12/16/1992
From:	NRC ADVISORY COMMITTEE ON NUCLEAR WASTE (ACNW)
To:
References
NACNUCLE-T-0066, NACNUCLE-T-66, NUDOCS 9212210298
Download: ML20126A926 (504)

Text

5 s

@we~s,m s, w w,w w~ wm~ ClGX/A,r, w.,.-% nn nv '

1. 9e hhh $? Y Yb Y "b0 Y N 7%s% M<MOFFICIALTRANSCRIPEGEPROCEEDINGSL

@aq f - . sac a prmue s:

w 2 - -, s n

,g

=in y7.a

, y , .

t 4

+

w. m a rc w w(h. %yy gg.

y% t. ,y' MNa D , , ,q v :yD v \ ;p[s *'q ' ,, yp *a  : m.

n _q' / 8

.  ; 3 g 5 ,

- .MNM h '

• i st # , y *

" I:Ti, ,

~

%,d$.,g%,&e h: V# $ ] %- +

M l <

<:, l, ,

S' '

WUkV g;k1 m ;wn n,,g .sx >n.x  ;

+

h,~2~ .? ', w ' ' > s s

w'

~ ~m I

.L :

+

.n -

. i

..u . ~. ..

- ,  :- . s m p$R .

V AR V^'n: - H. M OFFICE C0%s .EMN Fot. m  %; J.u l m * ?y:

.wn

$J B@yhM%,.+:, ~. N . EM p4 pg e~ .m %.

,pf'NYwe[n $ d' A J p #.' k'

, m

, MTgdg.0FsTH, 4m. c ,

i E CO, ~

MMITffM~- = .

~

i w~

4 i

~

eu w ,

T j ww bh'h(k~ ' , ,3 'gs A

~p; m s

. ,, 3 s

~

j . _. w rpp-

• - mj .

. y.

y oo eg;, w;1 s'~,_3 q 4 q) >

A 3- ' (s , je w s.

+ m

+

e4 r m

l pc ,; t;,y; s d.t. p_~ ht. < =' x4e .3>

-r3 .

v : V i. k = >s; yWW; n  ;

d1

' *r 'b' i

. b, e.2 yy ; Q,Mj ^ 3 t

4 7 t

s t

2 i

1.w ce. m

(( .

s

, G... ,

t

- .. ~

. n a '2 wh4 :4 i ~ '.' ~~ 8 LNuc kar! Regulatory Commission c

b ONl QN;Y wp C+ >

IAdvisorylCommitteeon'NuclearlWaste-

+

= w n+

s QiRM wm:

1 om a , ,

, r

,l q' ; e.  ;,, , My ,+

J 1

.h:W:* A 3 1. o

'4 . N: 3 . ,m /_ a

, s

,r<

@kW.

• W

$ ' M,W' porkinglGroupM6etingon: Performance!

'4

, MM ,j@ # p.~ "TitMjg ,

,s t', s Tass _essmentDforlthe.High-Level;Waate-

. - - ,~

% , y, h

. 4 *, e gb *EM

~

fo -

- w- < , 3Mr sgement:=Programi 4

i s

x si s an gym qs Gy< '

4 -

-7r .. . . . ,, , ,

sN

'%@m;g%,

Qw Q msw C@3% Q R A.f CDocket NoW. -

y epjp gfbg 4-

%gy4,x,r

,n ss w M 4;o

+j g ,j, w v

J y.

e

,g g

~7 ,

u

, . .;n- . T r: : -

c w,- , .., ,:

wwe, s

t r,-? ,J s,c s v 'e ~

L, Q..

>:y(, :g.t .

u8, R .c c:1 a . -

1: 4' --

W

! Z M Lw ',

, Mh, da n%. wn;a. s;g' n.

G % , ;L>i.

+

~

g>

+

n s, ~- yt3

'qPM ,, b y 5

' 4 6^ 1 p +,e < >

~ ' '.rz

- i ,

• m aa R&

%. N gyll , & .l; g[1 W

~'  ? " ' ' ' '

M&ql'.&

. fu.g ,--

qw,K "y ,s~ V% , . > ,

y. a +? -

' ~

. 'm

=-,m g' ,x

,4p 'N -w. '

s..,'

D'2

^

c

?

[pf

, r- 1 ' . ..c; ,,. ., ,

? ..

$j C

^ "

p4Mg'%we.

A >, %.y > ' tOCATIO'E

~ w + ', ,, d.{Q Bethesdaf Marylan,d) s

.c=

s

.s'_ W@%,o*M .e , @q  ; Ms <- w' n ' ;- ^m",

,,~., 3x s

x.

.. n

- pw('p.

1

'h,aof .

m

- ..g;s u

• i 3 .. e ,

, n' 't;s'

)

,\. y

.q <

MM %wn [$,q ,-d. j,v[ n% N, "

WedtNday,dDeced$r)1641992$

. ~ - n . .

M<  : P5GES: W -d23t. , . -

. , ., [m-hkhNh $ '

e ,

" , . ' kW ' , "

h '! s i < NN

"" W, . > V' m e i Rf Qh: 3S ~M m.h :W, o ,W ~ , ,

m:+ n' 3

n n .y

- +x m

a nec am, n u -

a c, p;,n@S e;w s 6 ,- s s i  ;<7' .,;.

cp. ypt

.- sr n 24 w'.

s ,

e1s d- , , a '

M;f w% 404 .-,i y $  ; 4 ,

% w a% , -

.i;3 , t .Ng s 4 x ,s ,.

f- 14 i

• 1' s s

^

+

s n

,g 4

1

>Z ' ' ' '

.%@p$p, W,w ~

?$[m M >

1. W 'y % f,mN T M@, -

• m >

- , m. . a ,- n m% ;o w % n /a M M@VM%D $% s-lQ d? ,s

~

jib W"' <{' '

x qwm. ^-W s l\

s ,

%- W1

' %%M, W ..a ,

e W  %

i 1.7U s 95

?Z gewdyp$ff+& ',-ig

• l WM i f -,

+

v s .

w s .;y

,- 3 ' ,

b.J,)W MM U2 3 \ ,J ,3~'*' ' * ^ - -

M s i Yhv% y

• b ECOMRENN FORi '

NN ^ b0 x v'> wm W n N m;m, bQY%m.s%e# ~' . - es '/ b , M~ ,<n40ETHECOMM.h s " n-ww % ~ s

. o '

a h .eNQ%hCh. i r '/hk= $$ 7 j iN . . .$ 'i M' 'f1 5 (-i e h -

-N$Nk6NMq.,, N]4 YMs

• M Nb qe

^ ^

N a

W '

v,d;  : <

nQ &

> y! h@W.% & w . eD%:&l$- T.i 4 sm:.;;m, a r 5 w-e 39 J ,

• • m.A I

m;.

.s ey ,

W 'W'-+

a3 3 c;

-m 1 w 3 J -, ;

\

j; c'

s 4

s ,a <

$ %y m & y . '

' F' ,

/ ~

h r[ rdh b '

f h

$ik g~ ? l 01126 KSt.NM, Suite 300 41' rh y -

, s@% w , 2M4Th$, qj$WZ  %% y%d; m Me4

~

J n m , uy~ Washington,'D.C[20006 .

'e", 8 *n "

~M '

wg wypW921221029'8 921216  ?(202) 293-3950 "

PDR ADVCM NACNUCLE < < a g@d@!gpp y%gf$ T-OO66 PDR "aM e M~N ' e ' "E

' M v" -i L _;

lO L'&%Mmv

, 4 <

m s .

cJ F gegw%';gwnype .

ype ogg n@n- gm v r w# @7 y - W e,s,= ~

cv++,. -s , -

~

w ,

@g . -

em -RhmmimiMut W mi gN#@g;i m i,.m p vlMiu'WVrgm di u m MII.lk m MpQPM fl II;E m mgB-l u {;_ u41 :- '" { _ '-j;.;;,j- '

g g g ggg a

c -

,  ? -

JRG\'A1 RCJ)JVr-6046 O OFFICIAL TRANSCRIPT OF PROCEEDINGS -

n = m ceny.n a rca .

o E ! EE OF THE COMM! tit 5 4

ggg .

Nuclear Regulatory Commission Advisory Committee on Nuclear Waste Tide. Working Group Meeting on Performance Assessment for the High-Level Wastr Management Program Docket No.

0 1DCA110N: Bethesda, Maryland DAE Wednesday, December 16, 1992 PAGES: 1 - 323 ,

t 9

1 ACNW OFFICE COPY- RETAIN FOR

._ MllF50FTHECOMMITTEE c l:

0'

^g ANN MLEY & ASSOCMTES, LTD.

1612 K St. N.W, Suite 300 ~

1 i

pg. n .,. Whington, D.C. 20006 th _ (\-

9212210298h21216 l- PDR ADVCM NACNUCLE i: T-00bb PDR

^

b . . .

5

r_-.

q 1

{}

\, /

PUBLIC NOTICE BY THE UNITED STATE NUCLEAR REGULATORY COMMISSION'S ADVISORY COMMITTEE ON NUCLEAR WASTE December 16, 1992 DATE:

The contents of this transcript of the proceedings of the United States Nuclear Regulatory Commission's Advisory Committee on Nuclear Waste, (date)

December 16, 1992 , as Reported herein, are a record of the discussions recorded at the meeting held on the above date.

This transcript has not been reviewed, corrected or edited, and it mal contain inaccuracies.

0 ANN RlLEY & ASSOCIATES, Ltd.

l Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

-=

4 1

1 UNITED STATES OF AMERICA 2 NUCLEAR REGULATORY COMMISSION 3 ***

4 ADVISORY COMMITTEE ON NUCLEAR. WASTE j 5 WORKING GROUP MEETING ON PERFORMANCE ASSESSMENT 6 FOR THE HIGH-LEVEL WASTE MANAGEMENT PROGRAM 7 Nuclear Regulatory Commission 8 Room P-110 9 7920 Norfolk Avenue 10 Bethesda, Maryland 11 12 Wednesday, December 16, 1992 13 14 The above-entitled proceedings commenced at 8:30 L

15 o' clock a.m., pursuant to notice, Paul W. Pomeroy, Working 16 Group Chairman, presiding.

4 17 PRESENT FOR THE ACNW SUBCOMMITTEE:

18 Dade W. Moeller, Chairman of ACNW 19 Martin J. Steindler, Vice-Chairman of ACNW 20 William J. Hinze, Member i 21 David Okrent, Consultant l

l' 22 Mick Apted, Consultant 23 Paul Davis, Consultant 4 24 Georgio Gnugnoli, Designated Federal Official 25

' h- ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W.,- Suite 300 Washington, D. C. 20006 (202) 293-3950

2 i -

1 PARTICIPANTS:

2 3 N. Eisenberg M. Federline 4 H. Dockery F. Bingham 5 P. Eslinger J. Boak 6 D. Fehringer R. Wescott 7 R. Codell T. McCartin 8 B. Sager 9

10 11 12 13 O" 15 16 17 18 19 20 21 22 23 24 25 O ^"" ai'ev a ^ssoci^Tes. 'id-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 '

(202) 293-3950

3 1 PROCEEDINGS 2 [8:30 a.m.]

3 MR. POMEROY: The meeting w.ll come to order. My 4 first order of business is to apologize to the audience 5 since our PA system is not working. That is not a 6 performance assessment system; that is the public address 7 system. We have one microphone for the presenter and the 8 rest of us will try to speak in loud enough voices so that we can be heard by the audience. I have had the privilege 9

10 of sitting in the audience for the past two days, and I know 11 how hard it is sometimes to hear what's going on up here at 12 the table. All of the presenters are asked to repeat any Our 13 questions that come up for the benefit of the audience.

transcript is going to work.

() 14 This is the working group meeting of the Advisory 15 16 Committee on Nuclear Waste relating to Performance 17 Assessment in the High Level Waste Program. I am Paul 18 Pomeroy, Chairman of this Working Group and a member of the 19 Advisory Committee on Nuclear Waste. The other membert.

present today are Dade Moeller, Chairman of the ACNW, Martin 20 21 Steindler, Vice Chairman of the ACNW and Bill Hinze. As 22 consultants and experts assisting the Working Group, we have 23 with us today Mick Apted, Intera Sciences, Inc., Paul Davis, 24 Sandia National Laboratories and David Okrent, from the 25 University of California Los Angeles, will be joining us.

Q ANN RILEY & ASSOCIATES, Ltd.

COud Reponers 1612 K. Street, N.W., Suite 300 l Washington, D. C. 20006 (202) 293-3950

4 1- Our subject today is the current status of_the O2 performance assessment, the IPA Phase-2 and the TSPA efforts

-3 within the high level waate' program. In order to' clearly 4 understand how research, laboratory and field work and other 5 analytical processes factor into the methodology of 6 performance assessment, the Committee has requested that

~

7 both NRC and DOE staff present t air respective approaches 8 in integrating modeling and information acquisition. It is 9 also expected that the NRC staff e"6 DOE staff will address 10 how their modeling and PA activitie influence their 11 information acquisition efforts.

12 We will hear from the NRC staff this morning, and 13 this afternoon from representatives of the Department of

() 14 15 Energy.

Issues important to this working group include the 16 following: a, what is the current status of the PA programs ~

17 -- and to those of you who have sat through the previous two 18 days' technical exchanges, we will apologize in advance for 19 repeating some of that information -- b, how have other 20 efforts-in the high level waste program affected the 21 ovolution of the. IPA and TSPA capabilities. That is, other 22 offorts such as reschrch,-site characterization, et cetera.

23 C, the converse question, how have PA 24 accomplishments affected the other efforts in the program.

25 D, what are the lessons learned in this program up to this ANN- RILEY & ASSOCIATES, Ltd.

.] Coud. Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202)_203 3950

5 r~s >

1 point._ Not.only are there any potentially fatal flaws in

?'

-2 this performance assessment araa,-but are there lesser areas 3 of difference that could result in significant_ delays to the 4 licensing process at a later point in time. Finally, what 5 are the future plans, short term and long term, for 6 performance assessment within both groups.

7 The NRC staff is in Phase 2 of its iterative 8 performance assessment program which is focused on 9 assembling an expanded, more robust total systems code which '

10 includes consequence modules. Dose modules are also being I

11 incorporated. Significant progress has been made by the NRC 12 staff in developing this second phase of the IPA,-and we 13 will hear in essence a progress report at this time. The 14 staff has also been pursuing an approach to deal with and

(}

15 use expert judgment in Phase 2.5.

j 16 Equally, the DOE has made significant progress _in-17 their transition from the performance assessment calculation 18 exercises, PACE, to their TSPA efforts in 1991 and 1992.

19 TSPA efforts at Battelle Pacific Northwest Laboratory and at

! 20 Sandia' National Laboratories have been integrated in l

L 21 developing in TSPA, and results'from both TSPA's will be l

22 . presented this afternoon.

[

23 A few operational remarks. The working group 24 format is informal. Most of you have been here before and -

25 know that. I encourage pertinent questions from everybody, ANN RILEY & ASSOCIATES, Ltd.

'Q Coud Repoders.

1612 K. Street, N.W., Suite 300 Washingtoni D. C. 20006 (202) 293-3950

6 1

1 not only our consultants and experts and Committee members.

2 We don't intend to limit discussion in any way, and at the 3 same time we do have a large body if material to cover 4 before the end of the day. I know everybody will coopercte 1

5 in attempting at least to stay on schedule.

6 I also would request that any of the participants 7 in the working group today recuse themselves from or limit 8 their participation in any portion of the' program where they 9 feel that their opinions or advice would constitute an 10 apparent or real conflict of interest.

11 If time permits this afternoon, we will have a 12 follow up discussion. This discuss will be in the form of a 13 roundtable elicitation where the various participants will have an opportunity to comment as well as to provide

() 14 15 observations and recommendations.

16 The meeting is being conducted in accordance with 17 the provisions of the Federal Advisory Committee Act.

18 Georgio Gnugnoli will serve as the Designated Federal 19 Official. Georgio is on my left. The rules for 20 participation in this meeting have been announced as part of j 21 the notice of this meeting that was published in the Federal l

o -22 Register. We have received no written comments or requests 23 to make oral statements from members of the public regarding 24 this meeting.

25 Hopefully, a transcript of portions of this l

ANN RILEY & ASSOCIATES, Ltd.

]' Coud Repoders l- 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

7 1 meeting will be'kept. Should any member of the public or 2 any organization desire to make any comments relative to the 3 subject of the meeting please make your intentions known to 4 Mr. Gnugnoli, and we should make every effort to fit you  !

5 into the meeting c:hedule 6 Before beginning with the first presentation, I 7 would like to take this opportunity to ask whether any of 8 the ACNU members or our invited consultants, have op9ning 9 remarks to faAe or questions to ask. -

10 (No response.]

11 MR. POMEROY: Hearing none, I would like to turn 12 the meeting over to Margaret Federline, Chief of the 13 Hydrology and Systems Performance Branch at the NRC. I

() 14 would like to say, Margaret, that it's a pleasure to see you 15 back with us.

16 MS. FEDERLINE: It's a pleasure to be back.

17 MR. POMEROY: That microphone does work Margaret, 18 the one that you have.

L 19 (Slides.)

20 MS. FEDERLINE: We appreciate the opportunity to 21 meet with ACNW members and Consultants today, to discuss our 22 Phase 2 activities and progress in that area. We really 23 hoped to be able to come-and discuss our final results with

-24: -you today, but some unexpected complexities in some of the 25 code development as well as some changes to our work ANN RILEY & ASSOCIATES, Ltd.

f] Coud Repoders 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

\

8 1 schedulo as a result of higher priority work has meant that 2 our Phase 2 production runs will be delhyed until January.

3 We will subsequently be reviewing the results, 4 doing an analysis and producing a report later in the 5 spring. We will be happy to come back and meet with the 6 Committee at that point and discuss our final results. We 7 hope the status report will be or interest.

8 Today during the briefing we hope to provide you 9 with an overview of NRC's performance assesement activities 10 during Phase 2, a detailed example of the IPA process, and a 11 specific discipline. Dr. Pomeroy, you had mentioned 12 yesterday the prospect of perhaps limiting that presentation 13 depending upon the time schedule. We will certainly be 14 willing to accommodate whatever, as we go along.

J )

15 MR. POMEROY: I think at this point, with further 16 discussion with the other Committee Members, that we would 17 like to try to include that if we can, Margaret.

18 MS. FEDERLINE: Fine. We are prepared to do that.

19 The use of expert judgment in NRC's licensing process, and 20 considerations for future NRC IPA activities.

21 [ Slides.]

22 MS. FEDERLINE: As we have seen over the past two 23 days, both from DOE and NRC, performance assessment is a 24 valuable tool both to the site investigator and developer 25 and the regulator._ In July, 1989, NRC sent a letter to the 1

ANN RILEY & ' ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 j

w 9 l 1 Department of Energy, transferring our site characterization 2 analysis. In that analysis we emphasized the importancc of 3 an early iterative performance assessment activity to focus 4 site investigations as well as to determine compliance with 5 performance objectives.

6 We were pleased to see and pleased to have the 7 exchange with DOE over the last couple of days, to see their 8 beginning that activity and pursuing a total systems 9 analysis. During the briefing, we hope that you will see 10 the integral role that our iterative performance assessment 11 plays in the regulatory program.

12 IPA is a valuable tool for us because it provides 13 both a quantitative analysis which allows consideration of 14 both magnitude and likelihood, and it is specifically

}

15 directed towards regulatory requirements. In a hierarchical 16 fashion IPA considers both a quantitative estimate of 17 performance and auxillary analysis which allow us to 18 evaluate the assumptions to determine parameter values, 19 through more detailed models which support the simplified 20 performance assessment models. I think we have seen that in 21 detail over the past several days.

22 Although attributes and components are very 23 similar, mission differences cause differences-in the scope-24 of the respective performance assessment programs between 25 NRC and DOE. As noted on this slide, DOE is responsible for ANN RILEY & ASSOCIATES, Ltd.

.@ Coud - Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

10

-1 determining if the site is suitable for licensing. DOE is 3 responsible for also developing _the tools and techniques 3 necessary for demonstrating compliance with NRC regulations. l 4 We do consult during the pre-licensing phase, and 5 I think our technical exchange is a good example of how 6 productive that can be. NRC revievr DOE's performance 7 assessments during pre-licensing consultations, and we are 8 currently involved in that process right now, reviewing the 9 TSPA. NRC will evaluate DOE's demonstration of compliance 10 in the licensing application. Therefore, we need an 11 independent understanding of the repository program. Much 12 of that is provided through the NRC research program.

13 The overall objective of NRC's IPA activities at 14 this point is to develop, maintain and enhance, the NRC 15 staff capability to review effectively performance 16 assessment in the DOE license application for.the high level 17 waste repository. I can't emphasize this enough. I know 18 there were a lot of questions over the last two days. There

19. are still remaining, a lot of difficult technical issues and 20 a lot of tasks that we must confront.

21 one of the important things to remember is that we 22 at the NRC will be faced with conducting this review during 23 the licensing proceeding, and we must have a sufficient 24 staff understanding of performance assessment and capability 25 to conduct this review. So, this is our early focus.

(7 V

ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

11 1 [ Slides.)

2 MS. FEDERLINE: There are some very important 3 additional benefits of IPA; the ongoing NRC evaluation of 4 DOE site characterization. This specifically provides 5 information for NRC commenting on DOE's study plans and 6 DOE's site characterization progress reports. In this way 7 we can impact decisions that DOE is making about site 8 characterization, to ensure that they have an adequate u 9 license application.

10 Another benefit is the ongoing NRC evaluation of 11 high level waste regulations. We have been interacting to a 12 great extent with EPA in the development of the regulation, 13 and we hope to be able to contribute to the National Academy 14 study which will input based on the Energy Policy Act of 15 1992.

16 A third important benefit is technical input to 17 the development of NRC staff positions and other regulatory 18 products. One example that I would like to emphasize there 19 is the integrated perspective that IPA provided which 20 emphasized the need for DOE to consider appropriate coupling 21 of processes and models. This was in the thermal loads 22 technical position that I know the Committee is familiar 23 with.

24 Another benefit is to contribute to the 25 . integration of multi-disciplinary prospective in the NRC Q ANN RILEY &- ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202)_293 3950

12

' .1 -high level waste program activities'and research.. I think 2 our source term model is a good example here, where we 3 integrate hydrology, geochemistry, metallurgy, and processes 4 that must be understood to interpret destructive events.

5 It's a very important aspect, that this is one place where

• 6 all the disciplines can work together and assess the impact 7 of one discipline or another, as we look at a total system 8 performance .

9 MR. POMEROY: Margaret, will we get into that at a 10 later point in a little bit more detail? I would like to 11 know a little more detail about exactly how that integration 12 takes place and, specifically, how the interaction of the 13 work in progress affects research or how research affects i

() 14 15 the work in progress.

MS. FEDERLINE: Yes. We have a presentation that 16 will specifically address that. But if we don't answer your 17 question please feel free, and we will develop it in more 18 detail, 19 MR. POMEROY: Thank you.

20 (Slides.]

21 MS. FEDERLINE: -Another benefit is the 22 identification of technical uncertainties which helps to 23 define research needs and establish prograwmatic priorities.

24 An example here was the IPA source term work indicated the 25 need for more detailed auxillary analysis and impacted the-O ^"" ai'ev ^ Associ^Tes. 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

- =.

13 1

1 direction of the EBSPAC program plan, which is an auxillary 2 analysis which supports the source term work.

3 MR. ICMEROY:- A question -- David.

4 MR. OKRENT: It's really more an observation than 5 a question. It relates to your NRC evaluation of IILW 6 relations. I just-want to note in the law what the Congress 7 asked of the National Academy were findings and.

8 recommendations on reasonable standards for protection of 9 the public health and safety, including A,B,C.

10 It's in what I read that goes from EPA to the 11 Academy or another discussion, somehow that first sentence 12 which is in effect to me the major part of the charge seems 13 to get lost. I hope that the NRC keeps in mind that in fact

, 14 the charge as Congress has given it is really, again, 15 findings and recommendations on reasonable standards for 16 protection of the public health and safety.

17 MS. FEDERLINE: It's a very broad scope of the 18 standard. As a matter of fact, we have discussed tnat with 19 our office director. We are aware of that, and we 20 appreciate-you bringing it to our attention.

21- (Slides.)

2? MS. FEDERLINE: When-we met with ACNW in October 23 of 1991 when we addressed computer capability which had been 24 requested by Commissioner Rogers, we discussed our 25 perfcunance . assessment strategy at that point. We discussed ]

i O ^"" ai'ev: $ ^SSoci^Tes. 'ta-Court Reporters 1612 : K. Street, N.W.,- Suite 300 Washington, D. C. 20006  ;

.(202) 293-3950-

]

l

14-1 with you at that point, our approach to the review of 2 performance assessment in doe's license application 3 submittal.

4 This has not changed at all but I thought it_would 5 be worth just mentioning again today, that we plan to review-6 the entire performance assessment at a broad level. We will 7 conduct more detailed review of specific areas or 8 significant areas. Both IPA and systematic regulatory 9 analysis will indicate areas where verification by 10 independent quantitative analysis is necessary. At least 11 one rough quantitative check on overall performance will be 12 performed.

13 I think as we look at the scope of doe's work and 14 NRC's work, it's important to keep our approach in mind.

}

15 one of the major benefits of IPA is the 16 integration role that it plays among the technical 17 disciplines in the program, as I mentioned earlier in my 18 presentation. To illustrate this, I have included several 4

19 slides here. I don't expect you to pick up on the details 20 of the slides. I think the important thing to notice is, 21 how much involvement there is of research staff, NMSS staff 22 and Center Staff.

23- This truly is a very broad and encompassing 24 activity. Although the resources for IPA are quite small 25 overall, we-do see a level of-participation of a wide number Q ANN RilEN & ASSOCIATES, Ltd.

Coud Repoders -

1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l l

15 o 1- of staff. We feel that this carries over into the

- 2 effectiveness of the rest of the program. As people become 3 involved in this total system analysis they begin to 4 understand how disciplines relate to one another from a 5 systems perspective.

6 You can see this is certainly one of our main 7 improvements over Phase 1, is the incorporation of large 8 number of people in the work and bringing this parspective 9 to the rest of the high level waste mission.

10 With-that, I would like to turn it over to Norman 11 Eisenberg, who will provide an overview of our Phase 2 12 activities.

13 MR. PoMERoY: Thank you, Margaret.

14 [ Slides.)

15 MR. EISENBERG: In preparing for this 16 presentation, we had to walk a line between trying to 17 present a complete story, also recognizing that many of the 18 ACNW members had been present the previous two days. I am 19 going to give an abbreviated version of the overview of the 20 Phase 2 work as it stands right now.

21 -Hargaret has indicated what the overall objective 22' of the staff's iterative performance assessment activity is 23 which is to develop the capability. There is also the 24 knowledge of what doe is doing and has done, e_ e the time

'25 required'for the review of the license application is very-

,] ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300

- Washington, D. C. 20006 (202) 293-3950

._ _ __~ . . - . _ _ _ . _ _ ..

16 1 short. We need to be in a state of readiness as soon as it O 2 comes in.

3 In addition, there are some additional objectives.

4 It helps us to interact in the pre-licensing phase with DOE 5 on the progress on their site characterization program. It 6 helps us in the continuing evnluation of the high level was 7 regulations, both NRC's and EPA's. It provides a basis 8 technical input to various regulatory producte including 9 technical staff positions, other regulatory products such as 10 the license application review plan, and it also brings up 11 issues to be explored by the high level waste research 12 program at NRC.

13 In addition as I mentioned Monday, there were two

() 14 15 special objectives that we had for IPA Phase 2.

that we have had Sandia National Laboratories as our One was is I

16 consultant, doing work in developing our performance 17 assessment methodology for over a decade. We switched the 18 work over to a Federally funded research and development 19 center, the Center for Nuclear Waste Regulatory Analysis.

20 We wanted to use Phase 2 as an opportunity to transfer the 21 technology to the Center from Sandia.

22 It had been the practice since we had neen going E 23 through one site after another and one geologic media after 24 another to do a demonstration or have Sandia do a 25 demonstration of their methodology for each media. There l

O ANN RILEY & ASSOCJATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 L

17 1

=

1 was not time to do that, so we are-using Phase 2 as the  !

_t -

2 opportunity to demonstrate and_ evaluate the Tuff methodology 3 that was prepared by Sandia.

4 Let me just briefly state that we are_ organized 5 with a management oversight board. There is three technical 6 coordinators representing the three organizations involved 7 that are working on Phase 2 which is NMSS, the office of 8 Reseacch, and the Center for Nuclear Waste Regulatory 9 Analysis. The work has been divided into six teams, of six 10 pieces of work, covering different technical aspects.

11 Team one is working on the system code, team two 13 on scenario analysis, the third one is flow and transport, 13 fourth is the source term, fifth is perhaps not properly

() 14 15 namea but disruptive consequences, the sixth is sensitivity and uncertainty analysis. As you can see from the names of 16 the people involved, that this is a collegial operation.

17 People from all three organizations have leadership-18 positions both in the technical coordination and as in the 19 team leads.

20 These are the current 1 cads mentioned here. There 21 has been some history of changes due to personnel changes.

22 MR. HoELLER: Norm, you didn't mention it 23 explicitly but does not this entire effort provide training

'24 and education for your staff, it provides an opportunity for 25 them to make mistakes,.provides an opportunity to develop ANN RILEY & ASSOCIATES, Ltd.

'C Coud Repoders-1612 K. Street, N.W., Suite 300 Washington, D. C. 20006

, (202) 293-3950

18 q- 1 maturity in this methodology, and in turn keeps DOE on-its 2 toes.

3 MR. EISENBERG Certainly, the first three things 4 that you mentioned. I am painfully aware of making 5 mistakes.

6 MR. BOAK: We are painfully aware of his keeping 7 us on our toes.

8 MR. EISENBERG: Good, I don't have to speak to 9 that one.  :

10 MR. OKRENT: Norm, to what extent do you feel 11 teams one, two, three, four and five deeply think about 12 uncertainties? .

I 13 MR. EISENBERG: Let me get this straight from_the

] 14 beginning. When you say uncertainty you mean the uncertain 15 state of knowledge in what the real physical world is like, 16 both in how it's modeled and the data needed to describe it.

17 You do not mean for example, something which is.a 18 considerable concern for which we spend a lot of effort, 19 which is the variability in the natural systela.

20 The repository is a rather unusual system to be 21 placed in the context of risk analysis which performance 22 assessment I believe anyway, is a subset of risk analysis.

23 It contains the natural system, and the natural system has 24 within it, huge heterogeneity _which is often absent.from man 25 made systems. Special attention and focus is placed on that h ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. - Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

19 p 1 particular type of variability.

2 A. lot of effort and a lot of techniques are used 3 .to describe that. If I understand you correctly, you are 4 talking -- you are not.

5 MR. OKRENT: That's also part of the uncertainty.

6 MR. EISENBERG: You are using a very broad 7 definition.

8 MR. OKRENT: I think it really encompasses 9 alternatives.

10 MR. POMEROY: Paul.

11 MR. DAVIS: I guess maybe the question is -- tell 12 me if I am wrong -- is the role -- l 13 MR. POMEROY: Please speak up, Paul.

14 MR. DAVIS: Is the role of team six only to do the 15 mathematical part of uncertainties as defined by the other

~

16 teams. The role of team six is not to say what is the 17 uncertainty in the source term, is it?

18 MS. FEDERLINE: Could you repeat _the question?

19 MR. EISENBERG: The question is, is the role of 20 team six merely to do the mathematical manipulations in 21 order to shed some light on sensitivity and uncertainty, or 22 is it to be concerned with the real uncertainties that are 23- inherent in the-modeling and in the system.

24 There are two different questions. Dr. Okrent's 25 question was, do the people in teams one through five really O;

V ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington,' D. C. 20006 (202) 293-3950

l' 20 j - 1 worry about the uncertainties and how concerned are they 2 with them. I would say that they are quite concerned with 3 them. There is a considerable amount of agonizing before we 4 decide on a strategy for modeling and before we decide on 5 any particular set of data that goes into the model, to try 6 to properly represent the uncertainties to the extent we 7 can.

8 Let me quickly add that the way we have designed -

9 the work, we have the quantitative estimates of performance 10 and the auxillary analyses as Margaret mentioned. The 11 auxillary analyses are to a large extent, the purpose for 12 their being, is to explore the uncertainties. It's to 13 evaluate the validity or the goodness of various assumptions

. 14 and various modeling strategies to go from what might be 15 considered raw data to the synthesized data which are 16 needed, that go into the actual computational models.

17 I would say that we are focused on the uncertainty 18 if nothing else. One of the important things that comes out 19 of the work is a keener appreciation of where the 20 uncertainties lie, and how large they are. I think that 21 virtually everybody that works on this would agree, that the 22 numbers are perhaps the least important thing that comes out 23 of the work. But all the thinking that goes into creating 24 them is what is the real important part.

25 MR. OKRENT: Is there an effort made to document Q ANN RILEY & ASSOCIATES, Ltd.

Court Reporters

-1612- K. Street, N.W., Suite 300 i Washington, D. C. 20006 l (202) 293-3950

T 21 1 at this state of time, one's assessment of the knowledge and

~2 the gaps and the uncertainties? or, is it sort of buried 3 into whatever the input one decides to put into it?

4 MR. EISENBERG: I think we make a good faith-5 effort to try to document and indicate where the 6 uncertainties lie. Certainly, having gone through the Phase 7 1 documentation we do that in a collegial fashion. I think ,

l 8 there are large forces at work that say you describe what ]

9 you did but you also describe what would be alternative 10 models, what could be alternative data, and why.

11 MR. OKRENT: Again, that's a part of looking at 12 uncertainties. If I can press it a minute, you make an 13 effort to see am I able to provide a distribution with 14 parameters for the uncertainties, or do I not know enough to

(}

15 do that with anything but pinning something on the wall and-16 so forth at this stage.

17 MR. EISENBERG: Yes.

18 MR. OKRENT: Yes, what? Something on the wall?

19 MR. EISENBERG: We do make an effort to try to use 20 the most, if you will, refined and quantitative estimate for 21 example, for parameters. But we resort to whatever we need 22 to do in order to get an answer of course, and then-document 23 the need for additional data.

24 For example in the Phase 1 results, we used 25 certain data to describe the retardation and the chemical

-\~'

O ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite -300 Washington, D. C. 20006 (202) 293-3950

,e.-

22 1 behavior of plutonium. At the same time we pointed out that

=2 this was a need for further research and further 1 3 information, not necessarily to be'done by the NRC but in 4 general, that if.we were going to do a better job of 5 modeling we needed to have more information on how to treat 6 that particular aspect of the modeling.

7 I am not sure exactly what you are trying to get 8 to. I think we are all aware of the uncertainties, and we 9 try to articulate what they are as we go along.

10 MR. OKRENT: I will come back to it. We are not 11 quite there.

4 12 MR. POMEROY: Margaret.

13 MS. FEDERLINE: Let me just add one point.

( 14 Another important dimension of our performance assessment 15 input as we review DOE's TSPA, we are going to try and point 16 out where there is not enough information available. That's 17 part of our process of the review. I think as early on as 18 we are in the process, that is an important contribution to 19 looking into uncertainties and pointing uncertainties out to 20 DOE, although they may exist in both DOE's performance 21 assessment and our performance assessment.

22 MR. OKRENT: Let me give an example that I 23 remember from my days in reactors. One day I asked the 24 staff to come in with their estimate of the probability of-  ;

25 the probable maximum flood for sites. I must say very ,

O ^"" ni'ev a Associates. 'ta.

COud RepOders 1612 K. Street, N.W., Suite 300  ;

Washington, D. C. 20006  ?

(202) 293-3950  !

I 1

23

-, I wisely, the man who came in froin the staff said I think my

^

2 best judgment is that they range from ten to the minus three 3 to ten to the minus 11 per year.

4 He was, I think, displaying what he thought was a 5 not too narrow band of knowledge. In fact, they did vary 6 videly around sites. I don't know if they did that much. I 7 think having that estimate, if you will, is really quite 8 useful in helping what you think you need to look at, what 9 you think -- if the range is all on one site and no matter 10 what happens it's not going to affect anything, that's -- in 11 this particular case ten to the minus three is not a happy 12 number. It remained a problem of interest. Thank you.

13 MR. POMEROY: Mick. Speak up, please.

O 14 MR. APTED: A point of interest. Later on in the V

15 agenda there's something on expert judgment. Is that effort 16 part and parcel of your iterative performance assessment, or 17 will we see some sort of link up between what you are 18 talking about and the use of that in this uncertainty?

19 MR. EISENBERG: It is part of the iterative 20 performance assessment. However, Phase 2 -- again, because 21 of the timing and resource limitations -- is not addressing 22 it. We have a Phase 2.5 which we are kind of squeezing in 23 between a couple of iterations, which is attempting to 24 address in a particular area the expert judgment.

25 MR. POMEROY: Margaret, could I ask you just an O ^"" ni'ev & ^SSoci^TES' 'ta-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

24 ge y 1 administrative question about something you brought up. Is

'V

~

2- the staff going to review the two-what I understand are 3 contractor documents that contain TSPA?

4 MS. FEDERLINE: We are~ currently reviewing the 5 Sandia document. We are only budgeted for one review, so we 6 are going to have to make a determination about the PNL 7 review. That's why I was pressing Jerry yesterday about 8 what standing they have and what DOE would like us to do.

9 I am sure for our own purposes, whether -- we will 10 provide formal comments back on the Sandia. Whether we do 11 or not on the PNL, we will certainly review it for our own 12 purposes for insight into our own planning.

13 MR. POMEROY: Jerry, this afternoon maybe you will 14 review that a little bit for the Committee at some point.

(

15 MR. BOAK: What I said yesterday?

16 MR. POMEROY: Right, in essence.

17 MR. BOAK: Yes.

18 MR. POMEROY: Go ahead.

19 (Slides.]

20 MR. EISENBERG: This is a diagram of sort of an 21 information flow chart or structure of the analysis that was q 22 done-in Phase 1. It does not show the auxillary analyses.

23 It follows the format of a typical -- what I would think is 24 typical -- risk analysis. You have a system identification.

25 step. You then split the analysis into two parts which feed O ANN. RlLEY & ASSOCIATES, Ltd.

Coud RepOders

- 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

25 1 back to each other.

2 One part is the front end analysic which 3 determines what the scenarios are and their probabilities.

4 The other is the so-called back end analysis, which gives 5 you the consequences of the'various scenarios. You combina 6 them into a risk curve. You perform some sensitivity and 7 uncertainty analyses on both the overall risk curve and on 8 the particular consequence analyses, and you then you 9 compare results to a regulatory standard.

10 I bring this up to be somewhat in contrast. One 11 of the charges for the talk this morning was, what is 12 different in Phase 2. Without going through in detail --

13 MR. POMEROY: I'vla on for a moment, will you. The 14 Reporter is having a problem with her machine.

15 MR. EISENBERG: This is very similar, in that you 16 have a system identification step, a front end analysis, a 17 back end analysis, calculation of a risk Leasure, and 18 comparison to a regulatory standard supported by sensitivity 19 and uncertainty analyses. One of the important changes in 20 Phase 2 is that we added the capability in a limited sense, 21 to look at dose.

22 So Ke have a biosphere transport step and a dose 23 command step. This required, of course, that the system 24 description step be expanded to include the biosphere. This 25 chart explicitly shows that we are-doing parameter sampling ANN RlLEY & ASSOCIATES, Ltd.

j -Q COud RepOders l 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 h

L 26 1 which, of course, we did in Phase 1 also.

gS

\~ 2 Another important aspect is that we are pulling 3 off what I had called surrogates for some of the other 1

4 performance measures of interest including the three 5 quantitative performance measures for the NRC regulation, 6 travel time, waste package lifetime and the fract2cr31 7 release rate. It's important to note that I call thom 8 surrogates number one, because there is not right no'r a 9 clear definition of what they are for regulatory purposes.

10 Humber two, some of the ways that we have 11 calculated these were sort of -- they were targets of 12 opportunity. We used whatever looked like a simple 13 effective way to calculate them without spending a whole lot

(} 14 15 of time. That was not the main focus of the effort, but it was something that if we could extract from the analysis we 16 unnted to do so. That's why they are called surrogates.

17 The aux 111ary analysis step is very important. I 18 can't emphasize it enov 1. It doesn't look as comprehensive 19 and as important as all the other steps but I think it is 20 quite impo-tant, because it sheds light on the models that 21 are used and the codes that are used to do the computations.

22 It also develtro parameters that are used in the parameter 23 sampling routine.

24 I should add, that there was significant 25 improvement in the source term thle time by adding more O ^"" ni'ev 5 ^ssoci^Tes' 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

27 l

1 direct releace pathways and adding gas source and gas te 2 transport explicitly. These were addressed in Phase 1, but 3 we have added them to the computation in Phase 2.

4 (Sliden.)

5 MR. EISENDERG: The aurillary analyses, lot me 6 just briefly step through them. In Phase 2, it was 7 evaluations of the Sandia developed flow and transport 8 codes. These are codes that treat fracture flow and 9 transport and interactions between the fractures and the 10 matrix. The offects of layering, faulting and dipping on 11 fluid flow. There was a review of the USGS regional 12 modeling approach. There was a separate activity to do 13 regional modeling.

() 14 15 For example, for the regional modeling activitit e, we get inputs to some of the parameters describing the flow 16 system that we use in the actual computation. There was an 17 evaluation of the effectiveness of using Kd's to describe 18 transport. This is an issue that geochemists talk about all 19 the time, and it's an issue that needs further work on.

20 There was an analysis looking at retardation of carbon 14 in 21 tha geosphere, moving as a gas. This was a new activity 22 that was then incorporated. The results of that were then 23 incorporated into the transport modeling in the actual 24 calculation.

25 There was an evaluation of the formation of O ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D, C. 20006 (202) 293-3950

l l

l 28 j I colloids in the field near the wasta packages, an evaluation 2 of source terms for magmatic events that might be different 3 from the kinds of source terms that you would see under 4 normal repository conditions because of the additional 5 heating from the igneous activity. There was an extensive 6 body of work done comparing various sensitivity and- ,

.J 7 uncertainty methods. j 8 MR. HINZE: Norm, you have used the term 9 " evaluation" over and over again. Do I read that as 10 verification, validation? What does the word " evaluation" 11 mean?

12 MR. EISENBERG Let me start off by saying what I, 13 understand to-be verification and validation. To me, 14 verification is certifying that a computer code does the 15 mathematical operation that it was intended to do correctly, f 16 Validation, to me, mesns that a computer code or Fodel, 17 conceptual model, describe a physical situation --  :

-E 18 MR. HINZE: A natural system. .

19 MR. EIDENBERG -- or an unnatural-system -- to 20 the degree necessary for the purpose the model was intended. -

l 21 MR.-HINZE What did you do in your evaluation? ,

1 22 MR. EISENBERG: Generally, we did not get closeIto 23 validation because.in most cases the data-are so sparse.

24. What these evaluations were, were-in some cases inter- .;

25 -comparison between computer codes which is not exactly L

Q ANN RILEY & -ASSOCIATES, Ltd. j Court Reporters

- 1612 K.- Street, N.W., Suite .300. 3 Washington,- D. C. 20006 l

(202) 293-3950 I:.

>. . . . - -.-.._..-..--...-,...;..-.-.aa.

- .-,.x-_...- - - - . . .. - ,- , - - - _ ,

t 29 1 verification -- it's benchmarking, another term. It in one 2 aspect of verification.

3 There was the use of models to explore other 4 aspects of the system. For example, this study looked at 5 other aspects of the system that were not included in the 6 codes that were used to predict performance, to determine 7 how important those other aspects were such as the formation 8 of faults, and how that might affect the flow that could 9 affect the performance.

10 So, these were, if you will, one step removed 11 types of analyses to determine how importa:st certain factors 12 were.

13 MR. HINZE: That's all dependent upon having some

( 14 good data to use in those models.

}

15 MR. EISENBERG: Of course. As I understand it, 16 the regional data are somewhat more extensive than the 17 specific site-specific data for Yucca Mountain.

18 MR. HINZE: Do you plans for --

19 MR. EISENBERG: If I could, let me just talk about 20 this one, which is another kind of activity that is 21 important to do. A more precise computer code that included 22 -- I hope I am not getting too far afield from what I know -

23 - that included chemical effects that the Kd approximation 24 does not include, was used to model the movement of a 25 species front in a column of water in a porous medium.

O ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 1 l

(202) 293-3950

30 1 That more complex code result was compared to the O 2 simple result, to see how far off it would be. This is the 3 kind of -- it's too bad Dr. okrent isn't here right now --

4 it tells you if you have some confidence in the more complex 5 code. It tells yoi.i how good the approximations you are 6 making and the ass.umptions you are making to do the 7 estimates, how good they are.

8 MR. HINZE: Is this part and parcel of your 9 sensitivity analysis?

10 MR. EISENBERG: No. The sensitivity and 11 uncertainty analysis is much more of a formalized 12 quantitative approach to varying parameters and seeing the 13 effects that they have on performance. This is, if you

() 14 15 will, a kind of conceptual sensitivity analysis, where we are addrcssing the conceptual models.

16 How that I have the floor. This is something that 17 I think is hard to communicate about the repository system.

18 I think for a lot of engineering systems people are used to 19 looking at models and computer codes which are very good, 20 which have a long history of predicting things and being 21 used effectively in all sorts of context for designs and 22 evaluations of safety, and you think about getting data and 23 putting them into the computer code and getting a result.

24 I think for the repository system which contains 25 so much of the natural system, the data also change your b

Q ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

31 sr. 1 model. They change your idea of how the system is put

$k 2 tugether and how it works. So, some of these evaluations are 3 really looking at and trying to address the conceptual 4 models, which is a much more difficult but a very important 5 aspect of uncertainty for the repository system.

6 MR. HINZE If I may follow up. At what stage 7 will you be getting into the validation of these models?

8 Where is that in the feedback loop in the performance 9 assessment?

10 MR. EISENBERG: I think we will talk a little bit 11 about that later. We have several activities ongoing to 12 look at validation. First of all, let me say my view, which 13 is that validation in the classic sense of comparing the

() 14 15 predictions of the models and codes to actual performance are by in large impossible for the systems we are talking 16 about because of the long performance periods. I can't 17 wait 10,000 years.

18 Wo are talking about a degree of validation or 19 some type of partial validation. What we can do is use the 20 same models and codes to make predictions on shorter 21 timescales, on shorter spatial scales, to gain some 22 confidence in them. We can use natural analogs.

23 (Slides.]

24 MR. EISENbERG: This is what Part 60 talks about 25 in terms of support for the models that are used for O ^"" ni'ev a ^ssoc'^Tes. 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

32 1 performance. It says models will be supported by an 2 appropriate combinations of such measures of field tests, in 3 situ tests, laboratory tests, and monitoring data in natural 4 analog studies. I think we are constrained to do the same 5 things.

6 We participate in a number of Jnternational 7 studies such as Interval, whose goal is to do this partial 8 validation and also to develop methods of validation. We 9 have an activity with the Swedish Nuclear Power Inspector, 10 to develop a strategy for validation jointly. And, the 11 office of Research has, I would say, some extensive work 12 both at the center and at the University of Arizona which is 13 intended to provide the field data and the laboratory data 14 by which we can do these partial validation activities.

15 MR. HINZE: That's great, but where are they fed 16 in and at what point is this done in the performance 17 assessment as you are now carrying it out?

18 MR. EISENDERG: I think Rex is going to talk about 19 that. If you could just hold off until --

20 MR. HINZE: Sure enough. No problem.

21 MR. EISENBERG: -- until he gets up and talks 22 about it.

23 MR. POMEROY: Mick.

24 MR. APTED: Norm, the word " applicability" of Kd l 25 is particularly -- I guess I am playing se:rantics this O ^"" ni'ev a Associ^res. 'ta.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C 20006 (202) 293-3950 ,

4

- - - . . . _ . , . , _c ,. . , , , -- r~-

33 7s 1 morning with you. I don't mean to. I guess I need to know 2 where you are going with that. Everyone acknowledges the 3 weakness of the Kd. Yet, it's way ahead of what's in second 4 place, certainly internationally and in the states as a 5 concept.

6 When you compare it with these other modelr. which 7 I assume might be isotherm models or some sort of non-linear l 8 sorption, are you looking for -- the Kd model is 9 conservative and hence applicable, or what do you mean by 10 applicable? Are you going to rule it out?

11 MR. EISENDERG: Once again, I think we want to 1 12 know the direction in which it might be off and the 13 conditions under which Kd's may or may not be applicable.

[^)

V 14 MR. APTED: If they are off higher -- if they 15 predict a higher release rate and then a more detailed model 16 which might capture more of nature, does that mean they are 17 inapplicable for performance assessment?

18 MR. EISENBERG: No. We need to know that this is 19 part of the comparison to the regulatory standard. You want 20 to know whether your models are conservative, and if they 21 are to what degree.

22 MR. APTED: Don't we know that almost by 23 inspection of the mathematical models? I mean, if we are 24 going to put in a more detailed model, take more credit for 25 things -- it's a bit of a strawman, isn't it?

O ^"" ai'ev & Associates' 'id-Coud Repoders 1

1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

34 1 MR. EISENBERG: Maybe for you it is Nick, but I

'- 2 just can't do it by inspection.

3 MR. APTED: I mean, again, what I have seen of 4 international reviews of Kd is that acknowledging all the 5 weaknesses, it still ends up being a rather conservative 6 approach to the concept of some sort of physical chemical 7 interaction during migration.

8 You mentioned that you had some results in Phase 2 9 that are going to be coming up to compare?

10 MR. EISENBERG: The auxillary analysis looks at 11 that.

12 MR. APTED: Do you have those results, I mean, 13 today? Can you spill the --

I' 14 MR. EISENBERG: Tim, did you want to speak to O) 15 this?

16 MR. MCCARTIN: I think part of what he's saying is 17 correct; that --

18 MR. POMEROY: Please come forward and speak up, l

19 Tim. l 20 MR. MCCARTIN: Part of the applicability is 21 related to what you were suggesting; that indeed, could you 22 pick a conservativo value for the Kd in the situation you 23 were trying to simulate, and in some cases depending on the 24 detailed modeling you could clearly show that you indeed 25 -could pick a conservative value for the Kd that would be O ^"" nev a ^ssoci^Tes. 'id-Coud Repoders i 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

35 1 applicable across the spectrum of your conditions.

• 2 For other situations, indood, you could not pick a 3 conservative value unless you wanted to go to un-rotarded 4 completely. It was more along the lines, how well could you 5 defend that conservative value as truly being conservative.

6 MR. APTED: My concern -- this Kd is an example.

7 We say applicable, I have a feeling that one is driving 8 towards trying to capture more and more physical reality 9 which means in a sense open onded research, open onded 10 modeling, open onded overything, to try to get more.

11 At what stato does applicability say it's not 12 perfect but it is suitable or conservative, or defensible 13 for intent of performance assessment.

14 MR. MCCARTIN: It's more or less -- the

}

15 applicability aspect -- was more or less yes, indeed, wo 16 could pick a value that we felt was conservative.

17 MS. BRADLEY: Let me point out though, --

18 MR. POMEROY: Please state your name, and speak 19 up, please.

20 MS. BRADLEY: I am Jennie Colton Bradley, one of 21 John Bradbury's counterpart, one of the geochemists at the 22 NRC. John's work was showing that Kd is not constant. So, 23 it varies with conditions, it varies with the kinds of 24 substraits that we are talking about.

25 One of the things we are trying to evaluate is O ^"" aev 5 ^SSoc'^TES. 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

36 1 what kind of variabilities are we seeing, what kinds of 2 sensitivities does Kd suffer, and what kinds of constraints 3 can we put on use of that value in modeling situations. The 4 model that John was used was for a very simple system, in 5 looking at potassium versus sodium absorption, which is a 6 very simplistic model. It's a way of addressing how Kd 7 varies and what the isotherms look like.

8 So, it's starting our approach to looking at Ed's 9 with more complicated models which have different sorts of 10 speciation. It looks to us at this point, that it is not a 11 simple question and is not going to be a simple answer. The 12 Kd is going to be variable, depending upon the conditions 13 and the species that you are talking about.

(} 14 MR. POMEROY: Norm, continue.

15 (Slides.)

16 MR. EISENBERG: Let me briefly try to go through 17 some of the enhancements that were added in Phase 2. We 18 developed an internal modo system code. This is a little 19 bit of jargon. In Phase 1, we generated files that 20 described the consequences of various scenarios and then 21 manipulated them with a system code.

22 In Phase 2, we have actually developed a system l

23 code that calls these computational modules and steps 24 through the various scenarios more or less automatically.

l 25 One of the driving forces behind all this was to make sure O ^"" nev $ ^ssoci^Tes' 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006

-(202) 293 3950

37 1 that the data that were used in the various computational 2 modules, some of which even though they were simplified 3 models, are quite complex -- was to assure that the data 4 used in different modules weare consistent, and that we could 5 improve the confidence that we had in the results that we 6 got.

7 We added the dose assessment capability, a limited 8 does assessment capability. We evaluated the Sandia 9 scenario methodology that had been developed for the NRC.

10 As I think we pointed out yesterday or the day before, wo 11 had a group at Sandia

• hat worked in the regulatory part of 12 the laboratory that did nothing but NRC work, developing our 13 methodology. There are other groups at Sandia working in 14 high level waste and still are working in high level waste.

15 We evaluated the methodology and found out that wo 16 had to modify it to some degree in order to get it to work 17 for our particular application.

18 MR. HINZE: Before you go ahead, could you just 19 expand a little bit in terms of what you are doing and what 20 your dose assessment model, that capability actually does?

21 MR. EISENBERGt Okay. It uses the Diddy Code.

22 The objective is to get a collective or population dose. It 23 goes through at least three pathways to do the dose 24 essessment. It looks at air dispersion, it looks at liquid 25 pathways and it looks at the chain as pathways by which O ^"" aev a Assoc'^Tes. 'id-Court Reporters 1612 K. Street, N.W.,- Suite 300 Washington, D. C. 20006 (202) 293-3950

_ ._._.___.._.m_ ._. _ _ _ ...._ _ _

l

)

38 l 1 people can be exposed to radiation. I guess it looks at 2 direct exposures also.

3 I am not sure what more you_want me to say about 4 it. That's kind of its -- I guess one. thing that I should 5 add is that, again, partially because of limited resources  ;

6 and partially because we felt that the focus of this work 7 should be on the repository system we did not do any work on 8 varying the parameters that affected the biosphere 9 transport. There are some studies that look at that. There 10 is an international study that looks at that.- That, we 11 felt, should not be the focus at this time of our work.

12 There could be a considerable amount of-13 variability or uncertainty involved in the dose modeling, 14 which we have not captured in Phase 2.

( }L 15 MR. POMEROY: You haven't tried to treat the 16 problem of individual dose.

17 MR. EISENBERG: We are going to try to come up 18 with some crude estimate. But I suspect that nobody is 19 going to like it. It's just=that, it's a surrogate for 20 doing it the right way. We didn't spend a whole' lot of '

21 effort. As I said, it's kind of a target of opportunity.

22 MS. FEDERLINE:' It will1 be an average individual 23 dose from the population dose calculation, not maximally 24 exposed individual.

-25 MR. POMEROY: Norm is-probably right.

h ANN RILEY & ASSOCIATES, Ltd.

Coud: RepOders 1612 K Street, N.W., Suite 300 Washington, D. C. 20006 '.

(202) 293-3950

39 1 HR. OKRENT: On that last question from Dr.

O 2 Pomeroy, in view of the Congressional action, isn't there 3 some pressure on the NRC to assess individual doses at --

4 let's take a specific site -- Yucca Mountain, because that's 5 what is in the act -- and to decide how well you can do it 6 and what the uncertainties are and so forth. In other words 7 it's not just something you would extract in --

8 MS. FEDERLINE: I will be addressing that this 9 afternoon in my strategic planning remarks. That's one 10 thing responding to the National Academy in changes in the 11 modeling, that will be a high priority in our planning 12 efforts. I will be addressing that.

13 MR. OKRENT: Do you think you have the tools now 14 that would let you do a defensible calculation?

}

15 MS. FEDERLINE: We will have to incorporate the 16 Genio Code into our performance assessment. That should not 17 be an impossible task. We believe we have the capability to la do that.

19 MR. POMEROY: Go ahead, Norm.

20 MR. EISENBERG: Iet me just speak out of turn and 21 say yes, I think there are tools available. Remember that 22 if we are talking about doses over 10,000 years it involves 23 what I would consider to be a vast extrapolation of human 24 behavior and practices far into the future which are going 25 to affect the dose calculation.

] ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 L

40 1 MR. APTED: Norm, 10,000 dose cut off, that's 2 interesting. I don't think anyone else is doing dose 3 modeling cuts off at 10,000 years.

4 MR. EISENBERG: You mean, around the world?

5 MR. APTED: Do you know something that I don't 6 know about 10,000 years?

7 MR. EISENBERG: I am sorry. I would think that 8 would be a minimum time period --

9 MR. APTED: I thiink that would be a laughable 10 minimum. I mean, 100,000 might be uore of a minimum.

11 MR. OKRENT: I think the last words are not in on 12 this subject. I think if and when one can get EPA to look 13 at other risks in society, you may in fact find -- or anyono 14 who is talking about times beyond 10,000 years -- to look at 15 risks in society. They may have to dig up some very strong 16 arguments they now don't have, for why one particular waste 17 does this and no others and so forth. RCRA is a 100 year 18 sort of thing.

19 I would be a little bit cautious. I know what the 20 international community has been doing. In a sense, it 21 seems to me they must have been influenced by the way the 22 first EPA standard came out and the way the first 23 discussions were. If you look in the introduction to some 24 of the international documents in fact, you will find one or 25 two paragraphs that say at some point we need to somehow Q ANN RlLEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l 41 1 harmonize with how chemical wastes are being handled. That 3 is already to be found in one or two documents.

3 I think we need to see where the future goes. An 4 issue that the NRC is certainly going to have to think about 5 is do they recommend that if one is to calculate out to 6 10,000 years, does one recommend that it's impossible to 7 predict what will occur and therefore if the calculation is 8 to be done we should have a standard biosphere. Otherwise, s

9 the problem becomes intractable.

10 It seems to me that you have to decide is, is that 11 your recommendation or not.

12 MS. FEDERLINE: We were asked by the Commission to 13 provide a paper, some background historical perspective, on

(} 14 15 the National Academy issues. We are in the process of doing that right now, and we would be happy to share our views 16 with the ACNW.

17 MR. PoMEROY: Thank you. I think that would be 18 useful, Margaret. Dade.

19 MR. MOELLER: Another key factor in all of this is 20 whether you address the maximally exposed individual or the 21 average member of the critigal group. Unfortunately or 22 maybe that's not the proper word -- both the EPA and the NRC 23 and the EPA in its standards and regulations and NRC in its 24 regulations, only mention the maximally exposed individual.

25 Internationally, the ICRP stresses the critical-O ^"" ai'ev ^ Associates' 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

42 1 group approach. The NCRP, unfortunately, doesn't say ono 2 way or the other. This Committoc did got the NRC RES staff 3 to, in one of the regulatory guidos supporting Part 20 that 4 we reviewod a month or so ago, they did incorporate the 5 concept of the critical group. This is another factor that 1

6 has to be addressed.

7 MR. OKRENT: Dado, could I add one brief comment.

O In the safoty goals instond of a fence post man -- what is 9 used as an integration over the first milo -- that avorage 10 is what is considered as the doso.

11 MR. POMEROY: Paul.

12 MR. DAVIS: Isn't your role within the IPA to tell 13 how difficult this is to do and then maybe mako

( 14 recommendations to the licensing effort. Is it your role in 15 this intoraction that you gave us earlier on the purpose of 16 the IPA to receive regulatory foodback, in to say if you do 17 it hero's the headachos in doing it.

18 MS. FEDERLINE: Yes. Our focus and our 19 interactions with EPA have boon to comment on the 20 implomontability of the standard which is NRC's 21 responsibility and not the basic environmental staadards.

22 MR. DAVIS: Those recommendations would be on 23 that.

24 MS. FEDERLINE: Yes.

25 MR. EISENDERGi I didn't mean to indicato that IPA O ^"" ai'ev a ^ssoci^Tes' 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950 l

J

43 1 was developing recommendations. We hope that by trying to 2 do some of the assessment itself, we can learn a lot and 3 shed some insight on how difficult it is to do various 4 regulatory strategies.,

5 [Slidos.)

6 MR. EISENBERG I think I got up to 16 scenario 7 classes. In Phase 1 wo had four, and at this time we have 8 16. Wo have a substantially more refined treatment of ficw 9 and transport in unsaturated rock than in Phase 1. We havo 10 added the gaseous transport pathway to the calculation. As 11 you will see on the next flow chart, you will soo wo have 12 also addod a source term to go with it. We have added 13 saturated zone transport. In Phase I we assumed the

() 14 15 accessible environment with the water table.

Thoro has boon a vast improvement in the source 16 term in Phase 2. We have added mechanistic treatment of 17 wasto packago failure, of substantially improved treatment 18 of wasto dissolution, and transport from the EDS. As I said 19 before, a gaseous source term, carbon 14. We have,'in 20 addition to expanding the probabilistic analysis of 21 scenarios, we have had to add the consequence modeling to go 22 with them. We have added that capability.

23 The sensitivity and uncertainty analysis, we hopo 24 is now a far better treatment in terms of being a turnkey 25 system, once we got the output from the production runs of O ^"" ni'ev $ Associates. 'ta-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950-

44 1 the system code we hope to be able to just dump the data O 2 into a program software package and get results which we can 3 then interpret.  !

4 Those of you that might remember Phase 1, we had 5 the repository device divided into kind of a linear segment 6 with four regions. We have now seven regions which are no 7 longer linear but that give some aerial definition to the 6 repository.

9 MR. PoMEROY: Martin.

10 MR. STEINDLER: Norm, what do you mean by an  !

11 improved waste dissolution model? What is improved?

12 MR. EISENDERG: last time we were looking at --

13 let me see, you can correct me if I don't get this right --

14 last time we used a solubility limit and essentially an 15 advoctive transport from the EBS that was raised to the 16 solubility limit if there was enough waste in the solution 17 chamber to come up to the solubility limit. If it was less 18 than the solubility limit, then you got whatever the 19 concentration was coming out at the advective velocity.

20 This time there is a more detailed treatment using 21 a stir tank reactor which has radioactive waste in it, which 22 is decaying and having in growth of daughters as well as a 23 compartment for the waste that comes out but cannot 24 dissolve, but will re-precipitate because it can't go above ,

25 the solubility limit.

] ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l 1

l 45 1 1 We have both advoction and diffusion taking wasto no* 2 out of the container. There was one other thing I was going 3 to say. Of courae, the concern has boon with the oxidation 4 of the wasto form, as a producessor to being able to 5 ciascl /o that part of the calculation that is still in 5 there. Old I loavo anything out, Dick?

' MR. CODELL Except the gas phase source term, it C v4sn't prenant in the Phase i study at all.

9 MR. HINZE While you have that up thoro Norm, 10 this highor spatial resolution. Has that provided you with 11 a decreano in the uncertainty in the analysis? What has 12 this done for you?

13 MR. EISENBERG: That's a hard question. I would 14 say for some of the uncertainty, absolutoly not. But what

[)

15 it may enable us to do -- remember, we haven't gotton the 16 respits yet, so we are not sure. It probably may enable us 17 te got a better quantification of how big the uncertainty 18 is.

19 MR. HINZE: If I could follow that up. In your 20 answer to Dr. Okront earlier in your discussion, you talked 21 about one of the aspects of uncertainty is the spatial 22 variability, temporal variability and interactivo 23 variability and so forth.

l 24 I would assume that what you are trying to do is l 25 decrease the uncertainties with this higher spatial Q ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006

(202) 293-3950

i 46 1 resolution.

2 MR. EISENBERG: That's correct. What it does is, 3 it gives -- you are moving away from using average values 4 and let's you use local values for regions of space. So 5 that, the different regions of the repository now have 6 different waste loadings in them, so that the temperature 7 history does not have to be the same across the repository.

8 They have a different flow coming into the different 9 regions, so you have that kind of variability across the 10 repository which we fell more correctly reflects the actual 11 conditions.

12 Now, I think it reduces the uncertainty because we 13 are -- instead of using an average we are using the local 14 variables. Whether the spread of the results will be any

(

15 different, I don't know. But I think it gives us a feel for 16 how important that variability is.

17 MR. HINZE: What is controlling the spatial 18 resolution; is it the data that you have, is it the codon 19 that you have, is it the time that you have or all of those?

20 MR. EISENBERG: Some of all of that.

21 MR. HINZE: Could you explain that a little?

22 MR. EISENBERG: Certainly, one of the things -- I 23 will get to this in the next talk if this one ever ends.

24 MR. POMEROY: It may not.

25 MR. EIEENBERG: We are pushing the limits of our Q ANN RlLEY & ASSOCIATES, Ltd.

COud Repaders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

47 1 computers. We are trying to work on a Cray, an advanced 2 acientific computer. We are having difficulty in run timo.

3 Essentially what you do when you separate the repository 4 into those regions is, you have a nimilar computational 5 modulo for the source term and the transport. The more 6 regiona you have the more timos you have to run thone, and 7 the run timos becomo great.

8 Another problem, of courne for tho source term, in 9 that the repository design in in the very preliminary stage.

10 One could question the advisability of going to a very 11 refined spatial definition when tomorrow we could got 12 another report from DOE that says we are not going to do 13 vertical emplacement, we are going to do horizontal

/O 14 emplacement.

V 15 MR. IIINZ E: I guous what I am going to got at if 16 you are going to discuss this again is, whether you should 17 be doing greator apatial resolution or whether you are at 18 that point and will be for como timo.

19 HR. EISENI1 ERG: I don't think we have decided on 20 that yet becauso we haven't gotten the results. I think a 21 lot will depend on how big an effect we soo.

22 MR. PoMEROY: Paul.

23 MR. DAVIS: I have a question in the process of 24 going from IPA-1 to IPA-2, and then three or four, or 25 whatever you are going to. In going from one to two, you

'<~g ANN RILEY & ASSOCIATES, Ltd.

t) Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 l (202) 293 3950

48 1 chose these things to change. Is there a systematic and 2 documented approach to why I change these things, and will 3 the same systematic approach be used going from two to 4 three.

5 MR. EISENBERG: We went through -- first of all in 6 our documentation for Phaso 1, we came up with a list of 7 recommendations for what to do in the future. The 8 recommendations included not just how to improve the 9 modeling but also information needs and research work that 10 needed to be done, without specifying whether it was NRC or 11 DOE or whoever needed to get the information. But that, 12 additional work needed to be done.

13 MR. DAVIS: What is the basis for saying this-14 model versus this model needs improvement, and what does 15 that mean, improvement? You don't believe it's 16 conservative?

17 MR. EISENBERG: I think it's a non-formal 18 elicitation process by the people who did the work, 19 expressing their concerns about the assumptions they were 20 forced to make and the limitations they felt were present in 21 the models that they used.

22 MS. FEDERLINE: Let me just say that --

23 MR. EISENBERG: I didn't finish. After we came up 24 with the recommendations we went through a formal planning 25 process in which we came up with a program plan for Phase 2.

O ^"" ai'ev

• Associ^Tes' 'id-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950-

49 1 That had to be revisited because of some of the things that 2 Margaret talked about earlier, that some of the things were 3 proving to be more difficult and more time consuming than 4 the resources allowed, and because other work came in and 5 took resources away.

6 We have done formal planning exercises in the 7 past. We are sensitive to all the advice that we get from 8 overywhere. For the next phase we are going to go through a '

9 planning exercise for Phase 3, which will be done in 10 association with a developing and fleshing out a long term 11 strategy for performance assessment.

12 . MS. FEDERLINE: I just wanted to comment that this 13 perspective, this strategic plan that ACNW has urged us to 14 do will give us some perspective on the priorities, and how 15 critical it will be to make an improvement in a particular 16 code and what that improvement is needed for. That's the 17 connection that we are going to be making in the strategic 18 plan, identifying key technical issues and saying that this 19 model is needed to address this regulatory issue.

20 MR. DAVIS: That strategic plan was not in place 21 between one and two.

22 MS. FEDERLINE: We had a more philosophical 23 performance assessment strategy which listed some general 24 goals, but we are going to take one step further and refine 25 that to go from goals to technical issues, to tell how-we Q ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington,' D. C. 20006 (202) 293-3950

50 1 are going to deal with those technical issues in performance 2 assessment.

3 MR. DAVIS: Let me ask it a different way. Sorry, 4 Norm, I wasn't real happy with the answer. If I took your 5 first results it shows the CCDF that violates. Did you as a 6 team sit down and say should we look at things that are too 7 conservative and let's move it to the left, or should we 8 look at processes we didn't include that could possibly move 9 it to the right, or did that CCDP play any role in your 10 deciding?

11 MR. EISENBERG: Let me just say what is my 12 personal approach to this.

13 MR. POMEROY: Norm, would you say for the audience 14 the question.

15 MR. EISENBERG: Paul wanted to know what kind of 16 was our guiding philosophy in looking at the Phase 1 results 17 and moving onto what we needed to do in Phase 2. Were we 18 trying to remove conservatisms or to address things that 19 might be non-conservative so that we could get a more 20 realistic approach, or that might make the assessment of the 21 repository to be not in compliance with the regulation.

22 First of all, the job of the NRC is not to 23 demonstrate the licenseability of the repository. That's 24 DOE's job. In that sense I think they have an advantage in 25 that, if they can demonstrate a conservative approach to O ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

51 1 modeling that demonstrates compliance with the regulation, 2 that's all they have to do.

3 The NRC staff has a sowswhat, I think, more 4 difficult task. The more difflialt task is to evaluate what 5 DOE in saying, and to try to pass somo judgment as to when 6 they say it's conservativo, is it really. I think we have 7 to huo closer to the line of realism, and try to got closer 8 to what actually might be the performance. I don't think so 9 should take as many liberties or have the advantage to tako 10 as many liberties, as DOE might have.

11 I think we are trying to Lind more realism. Of 12 course, always trying to look at areas that might be 13 important problems for demonstrating compliance, whether

(} 14 15 they chango the performance or whether they change the confidence we have in the estimaton of performance. I think 16 those are the kinds of things that we are focused on.

17 I think we definitely have a differont focus. I 18 think wo might do things differontly if we woro interostod 19 in demonstrating compliance. But we are not. We aro 20 interested in getting the capability to evaluato a 21 demonstration of compliance.

22 MR. POMEROY: Norm, I think you.are about ono-23 quarter of the way through the total number of-slidos that 24 we have' hero. We are running a little bit behind, I 25 suspect.

- .O ^ " " rii' e v $ ^ s s o ci^ T e s' 'id-Coud Repoders -

1612 K. Street, N.W., Suite 300 -

Washington, D. C. 20006 (202) 293-3950

52 g~ 1 MR. EISENBERG: You know how it is when you get up

( 2 here, you like to talk.

3 MR. POMEROY I know it, and we love it. Please 4 go ahead.

5 (Slides.]

6 MR. EISENBERG Very quickly. We plan to produce 7 results from Phase 2 in four different areas. We will have 8 a main report describing Phaso 2. We have in the process 9 already, several NUREG CR's that will describe auxillary 10 analysec. We have already available, some professional 11 papers and more are on the way. Of course, we will 12 participate in meetings such as this one, to try to present 13 the results and also to get peer review and Zeedback.

To complete Phase 2, we have to complete the

() 14 15 testing of all the modules and the overall total system 16 performance assessment code. We have to continue the 17 testing of the sensitivity and uncertainty analysis. Turnkey 18 system, we need to complete the documentation of some of the 19 auxillary analyses, do the production runs, do the 20 sensitivity and uncertainty analysis, and go through the 21 painful process of documenting the results.

22 Then we will be done, and I am done. I guess I'm 23 not.

24 MR. HINZEt No, you are not. I have one more 25 question, because I didn't understand your answer. You may O ^"" ni'ev a Associates. 'ta.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

53 1 have answered it and, if so, I apologize.

2 Paul asked a question about the role of team six, 3 the sensitivity and uncertainty analysis. I don't know that 4 you gave an answer, whether that is a mathematically 5 computer modeling oriented task group or if not, what else?

6 MR. EISENBERG: A large part of the work they are 7 addressing is the mathematical computations that provide the 8 basis for determining sensitivities and what we call an 9 uncertainty analysis moaning, which variables or input 10 paramotors, which uncertainty in those parametera produces 11 the largest uncertainty in the output results.

12 A large focus is on that. But I would say that 13 the individuals involved not only do they have their own

( 14 keen appreciation of the physical systein and what in

)

15 important and what to look out for but they are working 16 closely with the module developers to look for the 17 conceptual uncertaintics also.

18 It sounds like kind of an equivocal answer, but 19 they are really doing both. The focus is primarily on the 20 mathematica2. The main product out of task six will be the 21 mathematical analysis. Task six, you have to remember, is 22 also the group that is looking at methodology, looking at 23 the methodologies for sensitivity and uncertainty analysis.

24 MR. HINZE: Very good. At the risk of asking 25 another question because I see you haven't really filled out O ^"" ni'ev a Associ^Tes' 'id-Coud Repoders 1612 K. Street, N.W., Suite 300-Washington, D. C. 20006 (202) 293-3950

54 f 1 your time --

2 MS. FEDERLINE: We have three more presentations.

3 MR. HINZE: I am going to ask another question 4 anyhow, if I might, please. When one performs modeling, one 5 of the principal decisions that you have to make is 6 something about spatial and temporal resolution, and the 7 size of the model. We have heard a lot in this room about 8 the size of the region that should be investigated and 9 that's important and so forth.

10 Are you in any part of this analysis, looking at 11 those kinds of problems in the various items that need to be la taken into account?

13 MR. EISENBERG: I would say yes. I can give, I OV 14 think, two examples. One would be that the individual 15 module developers are looking at for example the 16 discretization in time to determine whether they can get 17 away with larger time steps and still get reasonable 18 estimates of the performance for that particular computer 19 module. This, for example, wcs done in the source term code 20 as a way of saving time in the computation.

21 We are definitely looking at that kind of 22 discretization. I would also say that the regional flow 23 studies give an idea of how much of the geography has to be 24 incorporated into the model cf flow in the unsaturated and 25 saturated zones in order to determine how far away O ^"" ai'av a ^ssoci^Tes. 'ta-Court Reporters 1612 K. Street, N.W., Suite 300 Washington D. C. 20006 (202) 293-3950

55 1 perturbations to that may have an effect on performance. I i 2 think we are looking at the spatial discretization, if you 3 will, or the spatial extent of the modeling.

4 I think there probably are a whole lot of other 5 examples. I think Tim McCartin might be able to speak about 6 the spatial resolution and some of the codea used for flow 7 and transport. I think this was going on all along, if I 8 understand your question correctly. l l

I 9 MR. HINZE: It's an important thing to keep in 10 mind. I assume that the technical coordination group is 11 certainly on top of that, meaning you and your colleagues.

12 MR. EISENBERG: We would hope so.

13 MR. POMEROY: Okay, Norm, thank you very much.

14 Margaret.

15 MS, FEDERLINE: Next, we will hear from Rex 16 Wescott. He will be discussing changes in the IPA ,

17 methodology resulting from research and other high level 18 waste investigations. Then, he will discuss changes in the 19 high level waste program that have resulted from IPA.

20 [ Slides.)

21 MR. WESCOTT: I think that you will note that I

22 added contributions to the title. I think I did that, 23 because I want to emphasize that this is basically an 24 expanding program. I think much of the research has been-l 25 more in the area of contributions rather than changes to a O ^"" ai'ev a ^SSoci^TES. d-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 1

_ _ . _ . . _ , . , - . m- -. , ,,

56

- rg i fixed program.

2 I think as Norm has pointed out, we have gotten ,

3 some very significant contributions from Sandia that we have 4 used.in this Phase of the IPA such as DCM-D3, NEFTRAN 2, l b both of which are computer codes and the' basic scenario 6 methodology which we modified somewhat from what was  !

7 de191oped at Sandia. Also, some other major contributor was 8 the Center for Nuclear Waste Regulatory Analysis. They have 9 four major projects that have been contributing to IPA; the 10 geochemistry research project, integrated waste package 11 experiment, the stochastic research project, and the 12 performance assessment research project.

13 More indirect contributors have been development 14 to TOUGH from Lawrence Berkeley, and University of Arizona

}

l 15 Apache Leap TUFF studies, which I think Norm mentioned that 16 I should speak to in a little more detail.

17 In regard to Sandia -- and I know these have been 18 discussed before so I will go through them real briefly --

l 19 discussed yesterday. DCM 3D is a dual continuum 3D 20 groundwater flow for unsaturated fracture porous media. It 21 represents fractures and matrix as two separate continue, l

22 interconnected to a transfer term. It was used in IPA Phase 23 2, to determine matrix and fracture flow in the unsaturated 24 zone by a series of simulations which actually were used to 1'

l 25 set up a table as input for NEFTRAN 2. It was also used in O ^"" ai'ev a Associates. 'ta.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

I 57 q 1 auxillary analysis to examine fracture and matrix flow in U 2 more detail.

3 NEFTRAN 2 is an improvement over NEFTRAN 1, which 4 is used in the first phase of IPA. It stands for Network 5 Flow and Transport in time dependent velocity fielda. It is 6 a one-dimensional network model. It looks at transport of 7 decay chains of radionuclides, as one of the main features 8 it adds -- one of the main features it has, I should say.

9 It also can look at -- NEFTRAN 2 can look at time varying 10 flow fields which is an improvement over NEFTRAN 1 or the 11 first NEFTRAN.

12 It was used in IPA Phase 2 for transport of 13 radionuclides uaing velocities calculated in DCM 3D, 14 actually input from a table as I said, and for calculating

% ))

15 time integrated discharge which gave us release and 16 concentrations which are used for the dose calculations.

17 [ Slides.)

18 MR. WESCOTT: That's regarding the Sandia scenario 19 methodology which was a modified -- we used a modified 20 version in IPA Phase 2. It looked at screening of events 21 and processes, scenario construction methodology and 22 methodology for estimation of scenario probabilities. I 23 think Norm had the same basic items in a chart where he was 24 showing how it contributed to the program.

25 In regard to projects from the Center, we have the C) ANN RlLEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 i _ - _ - _ _ _ _ - - - - _ _ _ - _ - _ _ _ _ - _ - _ - _ - _ _ - - - - - - - - _ - _ _ - - - - - - - - - _ - - - - - - - - _ _ - _ - _ _ _ _ _ - - - - - _ _ -

58 73 1 geochemistry research project which is actually the

('") 2 unsaturated mass transport research project, where we are 3 looking at fluid rock interactions. It provided insights 4 used in auxillary analysis really to the carbon 14 5 transports, specifically in regard to reactive transport in 6 carbonate system three phases.

7 We had the integrated waste -- and that should be 8 project -- waste package experiment. That was looking at 9 basically corrosion, stress corrosion, cracking, material 10 thermal stability and microbiologically influenced 11 corrosion. Experimental results from this experiment were 12 used for the crevice corrosion algorithm in SOTEC, which is 13 the IPA source term module.

[d

\

14 MR. APTED: My understanding is that the final 15 container here is not particularly selected. Was this on 16 zircalloy or was this on one of these austenetic materials.

17 MR. WESCOTT: The question was, since the '

18 container material is not selected was the experiment done 19 on zircalloy or another material. I can't answer that 20 Perhaps Bootie Sager or Dick Codell might be able to.

21 MR. APTED: Doesn't this point out part of the 22 problem. I understand the need for NRC and the Center to do 23 research, but to do specific materials research that haven't 24 been decided on, you are going to end up with a very 25 wonderful crevice corrosion model on material thar. may never (nj ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l 59 1 end up in the repository.

2 MR. SAGER: That is correct.

3 MR. APTED: Why not focus on the things that will 4 go into the repository, such as high level waste forms and 5 zircalloy or something so the ground even if it shifts in a 6 later design -- which, if one thing that history has told 7 us, is that things continue to change.

8 MR. SAGER: There are two things here. One is to 9 develop -- if Dick comes up to speak I think he will explain 10 the method which we have opted that came out of these. One 11 of the objective in the IPA 2 was to look at the lifetime of 12 the containment. It's not that we wanted to change -- we 13 wanted to define a method out of that, that could be used to j

() 14 15 meet the objective of IPA in Phase 2.

out.

That's all that came 16 MR. POMEROY: Go ahead, Rex.

17 [ Slides.]

18 MR. WESCOTT: Then, there's the stochastic 19 research project. This was primarily looking at stochastic 20 analysis of large scale flows and transport through 21 unsaturated fractured rock. This resulted in the refinement 22 of the big flow code which had been developed earlier but 23 which was refined in this particular project. It was used 24 in auxillary analysis related to the effects of tilting i

25 stratigraphy and presence of a fault zone.

O ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

I 60

- 1 The other research project is the performance 2 assessment rescarch project. This resulted in the 3 refinement of the PORFLOW. code. This was used in auxillary 4 -- go ahead, Paul.

5 MR. DAVIS: What we heard about over the last two 6 days was the deterministic analysis of flow on the tilting.

7 Can you tell me what the stochastic aspect is to that 8 research and how it fits in?

9 MR. WESCOTT: I will let Dick go ahead.

10 MR. POMEROY: Can you speak up, so that the 11 members of the audience can hear you as-well.

12 MR. SAGER: I will try. The stochastic project, 13 the flow is a product that came out of that. The analysis that we did assumes the distribution of the spatial and so (di 14 15 on, to study the effect of the Ghost Dance Fault was the 16 main objective.

17 This is not-part of the total system code. This 18 is part of the auxillary analysis which was done in 19 stochastic --

20- MR. POMEROY: Please speak up.

21 MR. DAVIS: Was it presented over the last two 22 days, or did I miss something. I heard the deterministic 23 analysis of the tilting and things. I didn't hear the 24 stochastic analysis. -What were the results and how did 25 they feed into your IPA.

-(] ANN RILEY & ASSOCIATES, Ltd.

COud RepOders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

61 4 fs 1 MR. MCCARTIN: I did not go into the distribution

(_ 2 of the variables that went into the model. It was, as 3 Bootie says, there was a correlation and there was some 4 stochastic variability put into each of those layers. I did 5 not -- just in the interent of time, I did not go into that.

6 (Slides.)

7 MR. WESCOTT: Here are some indirect contributions 8 over here but are very important contributions, was the 9 TOUGII code developed at Lawrence Berkeley Laboratory. This 10 stands for code for transport of unsaturated groundwater and 11 heat. It was originally developed under funds from DOS

)

12 geothermal as opposed to the waste program. The users 13 manual was funded by NRC.

/~^h 14 It was first used by NRC to model a University of V

15 Arizona core heating experiment, the results of that 16 experiment. That was INTRAVAL. One of the first uses in 17 terms of IPA was it provided limited confirmation of some of 18 our gas flow modeling assumptions. It wasn't our use, but 19 the paper by Sang seemed to confirm many of the things that 20 we liked about the gas flow model developed by Ben Ross, 21 and that gave us some added confidence in using it, one of 22 the reasons for selecting it.

23 We made a_ limited attempt to use it in IPA Phase 24 2. It would analyze heat induced moisture redistribution.

25 We wanted to get an idea of how important this might be to O ^"" ni'ev a ^Ssoc'^Tes. 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

')

62-1 gas flow. We expect.to use in the future IPA phases txt 0 2 analyze near fiel'd-drying and re-saturation and verify other d 3- aspects'of our-gas transport model.

4 In' regard to the Apache Leap TUFF _ studies, the IPA: ,

l5 contribution -- it-hasn't made a contribution yet -- we'are 6 expecting one from data collection from the inclined 7 borehole experiment and pneumatic studies in the multiple 8 borehole field site. This-contribution, we expect to be the 9 use of the data by the Center for development of alternative 10 conceptual models and investigation of' validation

-11 strategies. .

12 Norm said I would'get into a little more detail 13 here. From my recollection -- and I invite anyone-who knows

() 14 more about this project, to correct me if-I am wrong -- they _

15 inclined borehole studies are basically going to provide-16 geochemical data for looking at_ groundwater travel-times and

< 17 groundwater flux movement. The pneumatic borehole is going 18 to look at-primarily scale _ effects on modeling and also 19 perform some cross borehole experiments.in terms of the_-

l 20 averaging effects.

21' I think Dr. Hinze,_you asked a question on 22 scaling. I know that Professor Noyman right now is using the-23- results from the pneumatic borehole experiments, to_look at-24 changes in permeability at small intervals as compared to' 25 larger intervals, and trying to-develop relationships O ^"" ai'ev a Associates. 'id.

Coud Repoders 1612 K. Street, N.W., Suite 300 '

Washington, D. C. 20006 (202) 293-3950

63

_?-

L1' between the. pneumatic. properties measured at different i

4-

~

2- intervals so.that we.can get an idea of what-the effects ofi ,

3 scale'are.

~4 If anyone has some significant-things to add {to 5 that, please do.

6 MR. POMEROY: Rex, how transferable do you think ,

7 those studies are' going to be?

8 MR. WESCOTT: The question is, how transferable do 9 I think the studies are going to be. Admittedly, we have 10 had some concern because much-of the work that was being 11 done had to' deal with saturated flow.- I:believe right now 12 Dr. Noyman is directing.his efforts toward unsaturated flow.

13 If it turns out they should be transferable, I haven't seen-14- any initial results of his work on unsaturated flow._ That's 15 our-desire.

16 MR. OKRENT: What does the term validation 17 strategies _mean,-and do you'really-think one can validate 18 these-codes?

I '19 MR. WESCOTT: 'I will tall _you what-validation l

20 strategy means to me ,Land I will probablyz be corrected.

i 21 What validation strategy means to nelis the development-of

22 ' methodologies.as-to how you validate'different models based-2 3 .. upon the type of data you have collected.: I-think it'was-24 pointed out,by Norm, that-different models being-:used:for s l 25- -different things-require:different degrees-of_ validation.

O ^"" ai'ev * ^SSoc'^TeSd-Court Reporters .

1612 K. Street, N.W., Suite 300-

' Washington, .D. ' O. 20006 (202).'293-3950-u __ ._ . _

l l

1 l

'64 e- 1- I think validation strategy is a question of both 2 methodology and degree. Hcn1 we actually intend to validate 3 models, I think the only thing I can say is that from the 4 type of data that is going to be collected I think it would 5 be useful for validating both stochastic statistical models 6 as well as mechanistic models. I tnink we will have two 7 types of validation data that will be gained, and we will 8 probably be developing strategies addressing both types of 9 models.

10 I don't believe that fully answered your question, 11 but I don't know if I --

12 MR. MCCARTIN: Part of the validation strategy 13 work is being funded under the office of Research PA work at 14 the Center. I think Norm, years ago -- at least a few years

(}

15 ago -- coined the term iterative validation procedures.

16 Along the lines of iterative performance assessment, I think 17 validation is iterative also, in that you will design an 18 experiment that you think will get you the right answer.

19 You go out and test it, and then you come back and 20 look at the modeling results and you will find some 21 parameter you haven't evaluated, something that was left 22 untested that you want to go back again and do another 23 experiment for or repeat the experiment, whatever.

24 I think the strategy connected with the Arizona 25 werk is along those lines. Arizona is conducting some Q ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

65 gg 1 experimental work on making measurements in unsaturated 2 tuff.

3 If we can connect the performance assessment 4 people at the csnter with that group so that they try to 5 validate some of our concepts _and what strategy do you need 6 to -- what is the strategy between experimental work and 7 modeling. I think that iterative -- and it will'be an 8 ongoing procedure. As they do experimental work the Center 9 would conduct some simulations, provide feedback to the 10 experimentalists as, could you go back and test for this.

11 I believe it's the iterative nature between 12 experimentalists working and the modelers working together, 13 going back and forth to get to some type of strategy as how 14 best to as one would say partially validate the tools that

] }

15 we are using.

16 MR. WESCOTT: I was sort of thinking about Dr.

17 okrent's question as Tim was talking. Maybe one thing I can 18 say as far as direct methodology, the pneumatic borehole 19 experiment is going to be focusing _on geochemistry. I think 20 the use of geochemistry to determine groundwater travel 21 times and fluxos -- that may be where a definite methodology 22 may come out as to how one uses geochemistry, at least in 23 this particular situation as a validation tool for 24 deterministic or for more hydraulic estimates of groundwater 25 travel time and flux.

O' ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

i 66 MR. OKRENT:

p)'

'l 2

Maybe I should ask, what is your definition of the term validation, and what do you mean if 3 you say you have validated a code or nearly validated a code 4 and a set of parameters, 5 MR. MCCARTIN: I would think we would use the 6 slide that Norm put up from the regulation, As much as 7 people complain about the regulation I think this is an 8 example of a fairly concise statement of the things yot 're 9 going to have to do to get some confidence in the models 10 that you are using. I think the Center in their validation 11 strategy as well as'the laboratory and field tests, they are 12 looking at analog studies that are being conducted at the i

13 center.

l l

() 14 15 I think all of that is wrapped up into the package of what should we reasonably expect the DOE to do, to base 16 the validity of their models on. I don't know if you want i

l 17 to add to that.

18 MR. OKRENT: That's my point.

19 MR. MCCARTIN: -Clearly, the word validation is not 20 used there as a word. I think the process there -- we have l 21 led them down the path.

22 MR. DAVIS: This is your definition, you are 23 saying, of validation.

24 MR. MCCARTIN: I dor.'t see how we would interpret 25 -- I would interpret that as validation if ycu did all those O ^"" ni'ev a Associates. 'id.

Coud Repoders 1612_ K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

i 67 L n 1 steps. -It's the-best you could do in validating'the models 3

. we have. Why we don't use the words - I am sure there was

~

2  :)

o 3 -some regulatory or legal reason why they didn't want to say 4 -you'will validate your-models -- by.- But I don't think-l 5 anyone would expect the DOE or the NRC to validate their- 1 6 models with anymore thorough information than what is up 7 there.

8 MR. POMEROY: I think we could spend a lot of time 9 on this. I would like to let Marty ask a question and the 10 Paul Davis ask a question, before we go on.

11 -)UR. STEINDLER: You indicated that there were a 12 number of contributions.to the.second phase, and you ,

13 identified and walked you way through a fair number of.

L 14 Are you planning to devise some quantitative measure

(} these.

l 15 of how useful those improvements were?

16 MR. WESCOTT: The question is, are we going to try, 17 to develop a quantitative measure of,how useful the 18 improvements were.

19 MS. FEDERLINE: That will be evaluated in our L 20 results, in our final report on. Phase 2. That was what we- _

21 did in Phase 1. We looked at the research that had been 22 used and provided feedback as to whether =research needed -to 23 be changed-or whether there were additional things that 24 needed to be done.- I would say our process, documented 25 process for doing'that_is in,our report and our analysis of-h- ANN RILEY & ASSOCIATES, Ltd.-

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

.. . . . . . = _ _ . . .

68 1 results.

2 MR. EISENBERG: I think in terms of a quantitative 3 measure, I don't think so. I am not even sure how one-would i 4 come about defining a quantitative measure.

5 MR. WESCOTT: Did you have in mind, say, reduction 6 of uncertainties, something like that.

l 7 MR. STEINDLER: For example, you can pick any one 8 that you want. Let me tell you what I am looking at. For 9 example, the Center is listed as having made refinements to 10 the big flow code and refinements to the PORFLOW code.

11 My guestion is, did they make any difference. Do 12 you really care whether or not these are refined or not.

13 MR. EISENBERG: Yes, I think they made a 14 difference but I am not sure if you can quantify it. Maybe

. 15 Bootie can speak to this. They can model additional 16 situations now that they could not previously.

17 MR. STEINDLER:

That's not the kind of difference 18 that I had in mind.

19 MR. WESCOTT: I can think of an example, but I 20 think it shows a problem in answering the question. For 21 example in our gas flow model, if-we had been able to do 22 what we wanted to do with TUFF, I think we could have

23. changed the permeabilities. We would probably have reduced 24 them to make up for the moisture content.

25 We would have had better performance out of the Q ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 -

(202) 293-3950

l' 69 1 repository. In other words we wouldn't have had so much gas

-.O 2 flux going to the surface.

3 In that regard we might be able to quantify that.

4 The problem is, we could have gone to another model. In 5 other words, abandon our TGIF model all together and go on 6 to a more complete model, and perhaps even gain better 7 performance. I think in that regard it would be kind of 8 hard to say -- we would have to say that the use of the TUFF _

9 model helped us improve our results over what we were using 10 with Ross' model without the influence of TUFF results. At 11 the same time, that may not be as good as using a completely 12 different model all together.

13 MR. POMEROY: Paul?

14 MR. DAVIS: It's on validation, so I will skip it.

15 MR. POMEROY: Let's go forward, Rex.

16 (Slides.]

17 MR. WESCOTT: My last slide which constitutes the 18 second part of my talk, are changes in the high level waste 19 program resulting from the IPA process. We have some 20 examples.- I think Margaret mentioned the kind of things 21 that we were seeing from IPA.

22 The first thing is the development of SOTEC and 23 IPA Phase 2, led to significant changes in the development-24 of EBSPAC which is basically the performance assessments 25 codes which will be used to support SOTEC. They will supply Q ANN RlLEY & ASSOCIATES, Ltd.

Court Reporters --

1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

. _______________.__--_____.__________________.___________a

i 70 1 the analyses, which will be the basic underpinnings for-7-

'~# 2 .anything we do in the source term module.

3 Development in SOTEC also led to changes in the 4 IWPE research program in regard to modeling, pitting and 5 crevice corrosion. I think we already mentioned that, going 6 the other way. Based on IPA activities, it was decided that 7 global climate research should have a lower priority then 8 research in volcanism and tectonics. I think I should -

9 mention here that even though this is probably the kind of 10 concrete result I think you were looking for, it wasn't 11 based purely on the results of analysis. It was also based 12 on understanding of the processas as well.

13 I think any time cne takes results and uses it for 14 prioritization one has to recognize the uncertainty in the

[)

15 results at this time. We have to be very careful that we 16 don't wind up doing something that we hope to undo a year or 17 two later, 18 The carbon 14 transport analyses provided insights 19 which were useful in interacting with EPA and DOE. Dr.

20 Codell made-a presentation before the Science Advisory-Board 21 at EPA that allowed us *.o get some of our input into their 22 evaluation of the current carbon 14 transport. Of course, we 23 have also participated in technical exchanges with DOE on 24 this subject.

25 Problems in modeling corption provided incentives Q ANN RlLEY & ASSOCIATES, Ltd.

Coud RepOners 1612 K. Street, N.'N., Suite 300 Washington, D. C. 20006 (202) 293-3950 j

3 71

-1 for the-sorption research project. The'need to define 0 2 realistic failure modes of packages under seismic loading 3 led to the inclusion of structural analysis under repeated 4 loadings in the repository design construction operation 5 element task three, which is basically analysis codes and 6 methods of the Center's technical assistance project.

7 Difficulties in running the total systems 8 performance code has led to research on improving the 9 efficiency of computational models under the PA research 10 project. I believe this will include research in 11 abstraction as well as to how to go from this pyramid of 12 complex models into models that are going to run efficiently 13 in a source term code.

14 As you can see most of our results really come

(

15 from identification in the IPA process of gaps and knowledge 16 rather than actual physical results at this point, although 17 we hope that this will change in the near future.

18 MR. POMERoY: By in large, you are talking about 19 somehow problems with the code resulting in a need for 20 additional code or different code. Let me take for a minute 21 one case where you are not, namely your third bullet where 22- you decided on priority for global climate researca.

23 Could you briefly explain to me what happens as a 12 4 result of -- somehow the IPA group decided that global 25 climate research should-have a lower priority. How does O ^"" ai'ev a ^SSoci^TES. 'id-Coud Repoders l 1612' K. Street, N.W., Suite 300 L Washington, D. C. 20006 (202) 293-3950

72 1 that-get transmitted to, I presume Me's branch, so that they 2 are aware of what your current priorities are?

3 MR. WESCOIT: You are talking about the process 4 rather than results.

5 MR. POMEROY: That's correct.

6 HR. KISCOTT: It was my recollection -~ and I 7 think Mr. Linoitan and Mr. Silverberg will probably have to 8 speak to this. We went through a staff exercise et 9 prioritization. This was back in April of last year, if I 10 am correct.

11 MR. POMEROY: This is before you were very far 12 into Phase 2.

13 MR. WESCOTT: That's correct. This was based --

14 first of all it was based on auxillary analysis rather than

(

15 the actual performance assessment.

16 MR. STEINDLER: Was it based upon modeling?

17 MR. WESCOTT: In part. I will let Bootie Sager 18 speak to this. It was my recollection -- it was based upon 19 an understanding that I think in part, that precipitation 20 alone wasn't the major controlling factor in terms of water 21 rise and flux. Also, it was the tydrologic properties-of 22 - the materials. It also had a limiting effect.

23 Therefore, even though we may not have much

-24 certainty about how much rainfall we may get, we could be-25 pretty certain that the flux was not going to increase

(] ANN RILEY & ASSOCIATES, Ltd.

Coud RepOders 1812 K. Street, N.\V., Suhe 300 1 Washington, D. C. 20006  ;

-(202) 293-3950

73 i

1 tremendously because of that, nor was the water table going 2 to rise purely on a basis of just increased rainfall 3 tremendously, 4 Bootie is shaking his head, so I guess I have 5 agreement there.

6 MR. SAGER: I agree with Rex. I think we got a 7 statement of work at the Center to do a global climate. The 8 question we hed in our hands was did we have support to do 9 that kind of task. Secondly, would it contribute -- how 10 would it contribute to the overall effort that we were 11 making, inc.luding the IPA.

12 With that increase we said the maximum increase in 13 infiltration could be -- we did some modeling with that

.14

(} changes as you saw in the last two days, and came to the 15 conclusion that indeed we see rising water table but not in 16 a flood -- we didn't see that in the analysis. We did all 17 the assumptions and all the caveats. That's when we started 18 getting with the NRC staff and arguing whether we should do 19 this project now or wait and do something else first, and 20 then maybe tackle that. That's the logic that we followed.

21 MR. POMEROY: JLt that point then-, you had a 22 discussion among yourselves and established these priorities 23 and communicated them to Mel's office?

24 MR. WESCOTT: That's correct. I guess it was a 25 joint decision between Mr. Linehan and Mr. Silverberg as I Q ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1619 K. Street, N.W., Suite 300 tVashington, D. C. 20006 (202) 293-3950

. 1

74  ?

1 recall, to delay the project. I think it was approximately a $'

,3 a 2 year.

3 MR. EISENBERG: Can I add something.

4 MR. POMEROY: Please.

5 MR. EISENBERG: The same people that work on the 6 iterative performance assessment are the same staff that 7 interact with the Office of Research and develop the 8 research need letters that go from NMSS to the Office of 9 Research and that do the regulatory products and all the 10 other activities including the review of material from DOE.

11 So, this is both an advantage and a detriment to 12 iterative performance assessment in the sense that it's 13 desirable to the organization to have these people gain the 14 insights of doing the iterative performance assessment and (G'h 15 then use that information to influence the other activities.

16 Mr. Bernero has told us that's one of the things that he 17 wants out of iterative performance assessment, is to feed 18 the rest of the program.

19 The downside of it is that it takes time to do '

20 those other things. So, we cannot be as focused on doing 21 the work as perhaps we would like to be. We have to walk 22 that line. I think that is an important part of the process 23 that you are asking about, how it goes on. It is 24 essentially the same people in both roles. It's not 25 necessarily that there is -- although we will have planning

/O

\J ANN RlLEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 I

l

I 75

,, s 1 exercises and recommendations -- I think much more of this 2 informal kind of input is perhaps the larger part of the 3 interaction.

4 MR. POMEROY: I would say that IPA is only one 5 part of that input.

6 MR. EISENBERG: Right.

7 MR. POMEROY: For example, my personal conclusion 8 after listening to our working group on climate change, -

9 might be different than that one. Other people's might not.

10 Getting a large outside group of people is also very 11 important part of making such a prioritized decision.

12 MR. OKRENT: Two brief things.

13 MR. POMEROY: Speak up, please.

14 MR. OKRENT: My recollection of the climate

(

15 workshop conducted by Dr. Hinze was that at that time it 16 seemed like there did not exist what I would call a robust 17 method for translating from precipitation to recharge. Has 18 that changed in the interim?

19 MR. WESCOTT: No.

20 MR. OKRENT: Okay, I will just let it at that. A 21 different question completely. You have a-lot of changes 22 shown here but I haven't seen suggested changes 11n what 23 should be done in characterization of the site. Did any 24 ideas of that kind evolve out of IPA 2 up to now?

25 MR. WESCOTT:- I think not. Basically, because we O

ANN RlLEY & ASSOCIATES, Ltd.

COud RepOders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 1 (202) 293-3950 l l

__.m. _ . _ _ _ _ _ . . - _

^

l 76 1 haven't gotten any real results yet. I think that type of 73 2 thing.is going to come out of the sensitivity and 3 uncertainty studies. Of course, we need the data to do 4 that.

5 MS. FEDERLINE: Rex, just let me comment on that.

6 I can't speak to the specifics, but there was a case in 7 regional hydrologic monitoring where a particular individual 8 who was working on an auxillary analysis gained some 9 insights. He actually worked with Czarnecki's model and 10 noted some changes that needed to be made to a study plan.

11 He subsequently made those recommendations in the 12 study plan based on the work that he had done in the 13 auxillary analysis. I'm sorry, I am not able to tell you

(~T 14 the details. But we do have a couple of examples where --

l \/

15 because aux 111ary analysis went on independently we were 16 able to get some insights almost immediately.

17 The Phase 2 where it all has to come together, we 18 are still waiting for results, and we will be able to 19 provide those as feedback into study plans. Study plans and 20 I would say the site characterization progress report is 21 where we are seeing these analysts using their experience 22 and providing feedback to DDE. So, I have only been able to 23 give you one example. I am sure we could cite others.

24 MR. OKRENT: You may recall when I asked DOE 25 whether their TSPA -- if that's the right initial -- had led O' ANN RlLEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006

-(202) 293-3950 L

77 1 to proposed changes in the study plan, at least none were 2 identified off the bat. In fact, I asked was there even a 3 need for the study plan if you recall, in view of the TSPA.

4 It seems to me that question warrants some deep 5 thought, and it's related in my mind at least, to the ,

6 uncertainties that may exist in your ability to model or to 7 put parameters into the model or what might occur with other 8 versions of sort of the same model but nevertheless a 9 different stratigraphy. It is just that we have not seen any j 10 signs so far, is why I raise the question.

11 MR. HINZE: If I may, your third bullet, I want to 12 go back to that for just a moment. I wonder if there's a 13 possibility that that may be somewhat misleading. The first 14 bullet in the attributes of performance assessment that Norm 15 gave us was that the results are quantitative. Certainly, 16 that's one of the major advantages of it.

17 I am wondering if what I heard from Bootie, if 18 bullet three is really based upon quantitative analysis 19 other than that simple modeling which was quite simple. I 20 am wondering if that isn't misleading to point out that IPA 21 analysis led to this decision that global climate research 22 should be de-emphasized relatively.

23 MR. WESCOTT: I agree with your point. I think 24 that's why I tried to explain it a little bit as I was going  ;

25 through it. It definitely would be misleading to say that j O ^"" ai'ev $ ^ssoci^Tes' 'ta-Court Reporters i 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

78

- 1 it was based purely on results. It was based on results and

2 insights and an understanding of the physical nature of the 3 problem. I think that's what resulted in --

4 MR. HINZE: That's not quantitative --

4 5 MR. WESCOTT: That's correct, sir. I agree. I 6 think that's_one of the problems right now. When we do have 7 quantitative answers there's a lot of uncertainty associated 8 with them. I think that certainly tends to reduce the value 1 9 of them in a quantitative state to some extent.

10 MS. FEDERLINE: But I don't think we should-11 minimize the understanding of our disciplinary analysts,. of 12 getting an integrated view as they perform these analyses.

13 MR. HINZE: Absolutely.

14 MR. DAVIS: I am still trying to understand the

(

15 overall approach. The next to the last bullet, the need to 16 define realistic failure modes, can I interpret that to mean 17 that you have a reason to believe the failure modes you used 18 before are non-conservative?

19 MR. WESCOTT: Well, I would say ignoring'the 20 failure mode would certainly be unconservative in my 21 opinion.

22 MR. DAVIS: You treated waste release --

23 MR. EISENBERG: In Phase 1 we didn't have any I 24 seismic or structural failure modes. l 25 MR. DAVIS: So it'was not conservative. I

'(O ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

- --, -- l

d 79 1-

. l' 'MR. EISENBERG:- Yes.

2 MR. DAVIS: There was a mechanism that would-  ;

3 result in higher releases and you didn't include it.

4 MR. EISENBERG: Potentially could lead to ---

5 MR. DAVIS: Okay.

6 MR. POMEROY: Mick.

7 MR. APTED: Actually, out of all those modes,-I am 8 most happy with number three. I think one of the key

-9 features in addition-to increasing your communication with 10 other groups is to get some feedback into the program,. of __

~

11 what is important and what isn't.

12 Somewhere along the.line somebody's ox is going to 13 be gored. Not-everything is of equal high importance.- The l.

14 problem always-is that you get metecrologists who feel they.

! - 15 have the most important part of the elephant._to describe, 16 and someone else has -- the geochemists feel they are.

17 important. It's only through integrated in.the sense of

'18 integrated. disciplinary PA,.that you get some weighting of L 19 these apples and oranges.

20 I am looking forward to-further paring down=of. ,

j- 21 this hydro-monster; that, every time weucut off a head it l -: -22 seems like two more issues come up. -We need to really -find-23 something to begin 'to make this a tractable problem ratiher

-24 than a1 full funding for'every sort of_ discipline that:might~-

25: come along.

ANN- RlLEY & ASSOCIATES, Ltd.

RQ' -

. Court Reporters 1612 K. Street, N.W., Suite 300 l -- Washington, D. C. 20006 (202) 293-3950-u y- wv

A 4 k, m.o, 6 .5 &a ~4--

w h }& E ,M _

& ^n'% m,6- &~Av- -,dn--. @ --V-- O,.. 6 M e,464H+' ' ' " ' '

80 l' Out of all of those' bullets I am most comfortable 2 with number three.- In fact, I would.have hoped that 3 something out of each iterative-stage, something could be 4- really laid to rest finally -- colloids or containment for 5 - example. How important is-containment if we -- accepting-

. 6 sort of the substantially complete containment and all that l 7 morass. In terms of actual consequence to release and dose -- j

. 8 down the line looking at that, is 300 years, 3,000 years,

, 9 30,000 years -- I mean, thece are the type of sensitivities 10 that should be looked into and finally put this issue to

11 rest. End of sermon.

l 12 MR. SAGER: First of all, I wanted to correct the 13 impression that the regional study we did was included -- it l 14 was. It was a full-fledged two-dimensional and all kind'of 15 boundary conditions. It wasn't a simple study. -I.think I

.16 see some reluctance.here, that any results'we-get;at this 17 point have a large dose of uncertainty which is true.

18- But not to consider these numerical results, I 19 think, would be foolishness too. This is the best you can 20 get today, and you base-your decisions today on those.- You

-1 21- - temper them with other things you know. But they certainly. 1 I

22 make an important part of the decision making process, as--

23 they should, otherwise, this would all be a waste, from my

- 24 point of view. ..

25 MR. POMEROY: But when there are-. factors of

'Q ANN RlLEY & ASSOCIATES, Ltd.

. Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

81-1 uncertainty, one needs to include those uncertainties in 2 considering changes in priority. That would be my final 3 statement.

4 Margaret, given the fact that nobody has their 5 hand up here, is it possible at this point to take e break 6 before we have Norm proceed.

7 MS. FEDERLINE: Yes, surely.

8 MR. POMEROY: That's a long -- I perceive that's a 9 long presentation.

10 MS. FEDERLINE: How would you like us to proceed 11 to meet our end goal of finishing at 2:00. I don't want to 12 squeeze other presenters. My presentation can certainly be 13 shorted, from 40 minutes. Would you like Dick to attempt to 14 take less than one hour? I would just like to give some

(}

15 guidance to people.

16 MR. POMEROY: I would think at this point -- I 17 don't want to crowd over the 2:00 o' clock time too, because .

18 we have some plane schedules for the DOE people to meet. I 4 19 would like to do it within the current schedule, if that's 20 possible.

21 MS, FEDERLINE: Okay.

4 22 MR. POMEROY: We will take a 15 minute break, and 23 be back here at 11:00 o' clock.

24 [Brief recess.]

4 25 MR. POMEROY: We will reconvene. Do you need any l

Q ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 .

Washington, D. C. 20006 i

(202) 293-3950  !

82 1 introduction, Norm?

2 MR. EISENBERG: I don't think so.

3 MR. POMEROY: Please proceed.

4 (Slides.)

5 MR. EISENBERG: Again, being sensitive to the 6 direction that we try not to duplicate what was done in our 7 presentation on Monday. Unfortunately, a lot of this is 8 repeat. Maybe I can hurry through a lot of it. There are a 9 couple of additional items in the lessons learned here that 10 were not covered for the DOE presentation, some things that 11 are more appropriate for the NRC's advisory group that we 4

12 didn't want to necessarily bring to the wider world.

13 First of all, I should point out that these

() 14 15 lessons learned are based on the work that is in progress, and they are primarily concerned with developing the methods 16 and the models. We don't have the results yet. When wo get 17 the results we expect to get sc.ne ads..tional lessons learned 18 or conclusions that relate to both the methods .ind the 19 models as well as the overall sysuem perfcemance including 20 things like what's important for site cht ;terization.

21 of course, the conclusions that we have about 22 system performance as in Phase 1, are limited by using 23 models and codes which have only been partially supported 24 and tested which is appropriate we believe, for this 25 particular point in time. Certainly, there's a lack of O ^"" ni'ev a Associates' 'id.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

_ _ . _ _ _ _ _ ___ _ _ _ _ . ._ _ __. _ _ _ _ ~ _ _ _ . . _ ._. ._

03 1 mature site characterization data. When we go out, we will 2 only have had limited peer review.

3 MR. IOMEROY: While you are interrupted, can I ask 4 you a question. You have run the entire system code, I 5 assume, even though you have no results. You have, at 6 Icast, run something through all the modules so that you 7 know that the actual system code works as a system.

8 HR. EISENBERG Yes. They are looking at the 9 results. Obviously, there are problems, so we have to fix 10 them. ,

11 MR. POMEROY: Thank you.

12 MR. EISENBERGt One of the large areas of lessons 13 learned for us was in the area of the computer needs. For 14 those of you who were here for the DOE presentation, I think 15 we all agrJe with this concept of abstraction; that one 16 cannot use complex rescarch codes in the context of the 17 extensive Monte Carlo sampling that goes on in the system 18 code treatment and in the estimates of performance for the 19 repository.

20 So, there's a great need to focus on simpilfying 21 and speeding up computational modules. Certainly, we expect 22 that more time and effort has to be put into planning the 23 modules and in their development including some software 24 quality assurance and some other measures that might be 25 taken to improve the approaches to developing the computer O ^"" ni'cv $ ^ssoci^Tes. 'id-Court Reporters 1612 K. Street, N.W., Suite 300 t Washington, D. C. 20006 (202) 293 3950

84 1 codes. l 2 In that regard we will probably need a dedicated 3 staff to develop She codes and maintain them. These 4 computer nords, il you will, will have to work closely with 5 the scientific and technical analysts because one of the 6 things we discovered in doing this work is that you cannot 7 disassociate those two activities. To interpret the 8 performance of the codes you really need the scientific and 9 technical analyst. .

10 As I said Monday, it's not only the computer types 11 that need to be dedicated but considering the long hours, 12 large effort and low pay involved in performance _ assessment, 13 all the staff needs to be dedicated.

14 MR. POMEROY: I feel thoroughly dedicated staff 15 don't need to be high paid. ,

16 [ Laughter.)

17 MR. EISENDERG: Let me assure you, they are not.

18 MR. POMEROY As a serious question from the 19 viewpoint of the NRC Norm, do we need a dedicated staff to 20 do this? We have access, I presume, to extensive code 21 development work being done by the DOE. Is it necessary for 22 us to have code development staff and maintenance _ people 23 when we can draw presumably, on DOE's work?

24 I agree that you need an independent capability, j 25 but is that the best --

O ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

85 ,

1 MR. EISENBERG: I think it's a two part question.

2 For the maintenance of the existing codes and for the 3 maintenanca and usage of the doe codes that we have acquired 4 or that we may acquire in the future, I think that we will l 5 need a dedicated staff to do that. That is a non-trivial [

6 expenditure of manpower. s 7 Now, I think our strategy will be to rely heavily l 8 on the center for that kind of support. But I think it's 9 going to be something that we will have to do. With regard-10 to whethor we need a staff for development, I would say yes.

11 I think that the process of going through and working 12 through an analysis and converting it into a computation -

13 gives you substantially more insights than' standing on the

. 14 outside and just judging whether this assumption is 15 appropriate or that assumption is appropriate.

16 Furthermore, there can be and is, an awful lot of ,

17 unstated inherent assumptions thet are incorporated into the 18 coding that are not necessarily obvious from anybody's  !

19 documentation, and the NRC staff needs to'get experience in 20 both what sorts of things might be incorporated that are not 21 obvious as well as the impact that they can have on the 22 quantitative estimates.-

23 I don't think we can do-it any way, other than to 24 get the experience.

25 MR.-- POMEROY: Do-you think-you have a sufficient Q

ANN RlLEY & ASSOCIATES, Ltd.

Court - Reporters _ .

1612 K.; Street, N.W., Suite 300

. Washington,- DJ C. 20006 (202)_293-3950 s

, ~ . , . . . _ _ , ,- ,_...s._ _ _ _ . . . . _ , . . , _ , _ . _ , , _

86 1 understanding of the codes you have gotten from SNL to know 2 there aren't assumptions, expert judgments and other things 3 built into some of those codes that you haven't identif.'ed 4 yet?

5 MR. EISENBERG: I think we know enough about them 6 to say that some of the mathematical algorithms place 7 constraints on the code that we would not have necessarily 8 realized had we not gone in and done some extensive looks at 9 them as well as trying to write our own. I don't know if 10 Tim wants to speak up on this or not, or perhaps Dick.

11 There are things like the automatic discretization 12 in time, the time stepping, the spatial discretization which 13 you may control from the outside when you specify the 14 running of the code but that the algorithms inside the code 15 may change and make finer based on some other requirements 16 or criteria for convergence or for stability. You need to 17 know what those are, to know whether the answers you are 18 getting are the real answer or just some short cut that was 19 taken in order to produce some result.

20 I think that without getting into the actual 21 insides, the guts of the code, you cannot know that.

22 MR. POMEROY: You will have to come up to the 23 table Tim, because the transcript person is having 24 difficulty here.

25 MR. MCCARTIN: If I could just add one thing.

O ^"" ai'ev $ Associ^Tes' 'id-Couri Reporters i

1612 K. Street, N.W., Suite 300 Washington, D C. 20006 (202) 293-3950 l

87 1 There might be a slight misunderstanding of what code 2 development means. Generally for the PA tools within NRC, a 3 lot of that development was getting existing computer models 4 that a lot of development had been spent either by DOE or 5 other organizations -- USGS -- bringing them in house and 6 doing some modification of those computer programs for our 7 particular use.

8 And that, I think most of that development, is 9 more along the lines of modifying a particular existing 10 computer program to our needs which is not a trivial 11 exercise. As Norm pointed out, you have to do a lot of work 12 understanding exactly what this code is doing, and then how 13 best to modify for your own needs.

() 14 15 Most of the work going back to Sandia in 1976 when they started developing some tools for us, .most of the 16 effort brought in existing codes and modified them to make 17 use of obviously the research and development that had been 18 done outside of NRC. We don't have that much money to 19 spend, development from scratch.

20 MR. STEINDLER: Is it the intent of the staff to 21 dissect all the codes, each of the codes that they are 22 using, so that the underlying assumptions in the mathematics 23 as well as the program shortcuts and their consequences are 24 clearly evident?

25 MR. EISENBERG: I would think no. If I can borrow I] ANN RlLEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

i 88 1 rame terminology, we are using a graded approach in which 2 the codes that are most pertinent to the performance 3 measures that we are most interested in are the ones that we 4 want to do the most work on, and some other codes that are 5 of not such central importance we would do less work on.

6 MR. STEINDLER: Are you planning to require of 7 DOE, such a comprehensive write up so that in the event that 8 a new code arrives on the scene that you have not had a 9 chance to dissect in that way, that someplace you will be 10 able to obtsin that kind of information?

11 MR. EISENBERG: I believe we already have 12 requirements for the DOE to provide the appropriate 13 documentation for their codes as well as the appropriate 14 level of software quality assurance.

15 MR. STEINDLER: Are the specifications for which 16 you consider to be adequate documentation written down 17 somewhere?

18 MR. ETSENBERG: Yes. I think it's NUREG 0856. In 19 that the number, yes.

20 MR. POMEROY: Go ahead, Norm.

21 (Slides.)

22 MR. EISENDERG: Anather thing we found to our 23 dismay but not surprised, is that the computer facilities 24 that we have available for doing Phase 2 are not adequate.

25 But that, we expect the upgrade which has been approved and O ^"" ai'av $ Associ^Tes. d-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

89 1 is in the process of being implemented to take care of most 2 if not all of those problems.

3 MR. POMEROY: What is the cost to make a complete run?

5 MR. EISENDERG: Right noir, a single Vector run of 6 the system code I believe is running about two minutes on 7 the Cray. If we for rule of thumb want to talk about say a 8 300 Vector per scenario and multiplied by 15 scenarios --

9 somebody can do the math in their head. I think we are 10 talking about like $16,000.00 -- around $30,000.00 for doing 11 all 16 scenarios.

12 We have hopes of improving -- as you all heard i 13 Monday -- of improving the efficiency of the modules by 14 perhaps reducing some of the level of input variation.

15 MR. POMEROY Right. Go ahead.

16 (Slides.)

17 MR. EISENBERG We expect that in the future as we 18 do any additional code development, that we will have to 19 have a rigid specifications for these developments because 20 we will have a broader effort with higher complexity and 21 more players, and we will need to specify very carefully 22 what the variables are that need to be produced by preceding 23 modules to put into succeeding modules as well as the format 24 for the data transfers between modules.

25 Another thing we learned from working on Phase 2 O ANN RlLEY & ASSOCIATES, Ltd.

Court Reporters 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 t -. -_ - . -_ . -, . -

l 90 1 is that the computer codos that we have available were not 2 as transportable as we had expected, from one computer 3 onvironment to another. This produced some problems in 4 completing things, and it was not ameliorated to the extent 5 wo expected by having and using standard program languages.

6 A specific regulatory lesson learned if you will 7 is, for carbon 14 it's a difficult and costly issue to 8 model. If we want to make the modeling even more realistic 9 it will be even more costly. It appears to be a problem for 10 demonstrating compliance with EPA and NRC regulations.

11 Hopefully, Dick will speak at longth about that a little 12 later.

13 As far as wasto package design, one of the things 14 we found out by going through IPA that we didn't necessarily 15 hit upon in our previous reviews of the DOE program, is that 16 there does not appear to be any analysis for the post-17 closure performance of the waste package in relation to 18 seismic disturbances that DOE has on the street right now.

19 That was a little bit surprising.

20 (S11 dos.]

21 MR. EISENBERG: Here are some points that were not 22 mentioned on Monday. We fool that technical coordination is 23 an important part of what goes on in iterative performance 24 assessment, at least at the NRC. It definitely is a 25 fulltimo activity. It's because performance assessment is-O ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 j

91 1 inherently interdisciplinary and there are problems that are  ;

2 created because of the distances of separation between the l 3 office of Research and HM.cs which is not very large and then 4 with the Center which is larga.

5 For project management, we introduced some 6 improved strategies for managing the activity including 7 closer tracking of resources expended versus the work 8 accomplished, as well as trying to keep the work on 9 schedule. As I alluded to earlier today, the close 10 interaction and coordination of IPA at NRC with the other 11 technical regulatory activities is a highly desirable thing, 12 but it costs in terms of having the staff available to focus 13 exclusively on the iterative performance assessment which is 14 a resource intensive activity.

(}

15 MR. STEINDLER: What do you mean by the problem 16 brought on by distance?

17 MR. EISENBERG: There is problems in grabbing two 18 or three people and having a discussion about a problem when 19 it comes up, because they are thousands of miles away in a 20 different time zone. There is problems --

21 MR. STEINDLER: Have you discovered a telephone?

22 I don't want to be too flippant, but that's an old adage as 23 you know that has come up a dozen times. I must say that I 24 am a little puzzled by it.

25 MR. EISENBERG: I talk with my mouth but I also O ^"" ai'ev ^ Associ^Tes' 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

92 g 1 talk with my hands, and it's hard to draw something on a

(~

2 board and say this is the point I am trying to make over the 3 telephone. We do not have teleconferencing facilities, 4 which would be a great help to this kind of activity.

5 This is a working level activity. This is not 6 kind of a formal exchange. This is an activity where you 7 want to draw pictures and point, and it is just an 8 impediment to good communications. The other aspect of it 9 is the computer communications. We just have trouble 10 transmitting -- getting good connections, first of all and 11 transmitting data, and transmitting large volumes of data 12 reliably.

13 Dick, do you have a point?

. 14 MR. CODELL: That. was my point. Things that have 15 been standard in universities and laboratories for years, wo 16 are only just getting and we don't have yet.

17 MS. FEDERLINE: We have not had a high speed 18 communications link. They have recently installed one, but 19 it's still not fully operational for the staff to use. So, 20 what we are reporting on is things from Phase 2, and many of 21 the improvements from the advanced computing capability 22 pilot study will rectify some of these, like the high speed 23 communications line for the computers and equipment and 24 software improvements.

25 MR. STEIN 0LER: Has all that been factored into ANN RILEY & ASSOCIATES, Ltd.

[] Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

i 93 1 the new building?  :

2 MS. FEDERLINE: Yes, the design has considered 3 that. The Center did a very good job in supporting us in 4 designing.

5 MR. BOAK: I would like to say this in support of 6 Norm's position. I am Jerry Boak, of the D03. We have 7 recently completed installation of video teleconferencing 8 facilities. I often feel in teleconferences tnat we had to 9 plow the fields many times in both directions to get it 10 taken care of. I was really astonished at how rauch quicker 11 some issues got resolved in a video teleconference than they 12 did in a regular teleconference.

13 [ Slides.)

() 14 HR. EISENBERG Now, I would like to go task by 15 task, referring back to the six tasks in a previous 16 presentation. For the system code we have, as we suspected, 17 discovered that it's a very complex code and there is 18 further effort required for quality assurance, improving the 19 computational efficiency of the code. We need to pay more 20 attention to simplifying complex process models into modules 21 that are suitable for use in the system code. Also, the 22 overall design of the system code which reflects the 23 intellectual context that we are doing the calculation in.

24 It has inherent in it, the entire technical 25 approach towards scenarios, their definition, determination O ^"" ai'ev $ ^Ssoci^Tes. 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 -

Washington, D. .C. 20006 (202) 293 3950

, s _

94 1 of probabilities and even I would venture to say the 2 definition of terms like uncertainty. All of that should be 3 reflected in the design of the system code.

4 We expect, again, that we wi2! need support staff 5 to develop and maintain the system code and that, as we 6 proceed to try to treat additional scenario classes or to 7 treat the existing scenario classes in more robust fashions 8 we may have to revise the overall structure of the system 9 code because of some of the problems of dependencies among 10 the modules and tha way that it treats time progression.

11 We found out that the Sandia scenario methodology 12 needed to be modified in order to work. We found that 13 sampling certain parameters for the consequence modules

() 14 15 actually would affect the probability of occurrence of certain of the scenarios. What this seems to point to is 16 that the approach adopted by the WIPP project for creating 17 the CCDF's, the conditional CCDF's as well as the overall 18 CCDP might be a more appropriate way to tackle the problem.

19 The evaluation of the scenario probabilities does 20 not appear to be adequate. one problem, of course, is the 21 data that is available. Perhaps a more algnificant problem 22 la that there needs to be some advances in the planetary 23 sciences and the earth sciences, in order to be able to do 24 the modeling of the occurrence and manifestation of some of 25 those occurrences in order to determine their probabilities.

O ^"" ni'ev $ ^SSoci^Tes' 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20000 (202) 293-3950 l

l 95 1 MR. STEINDLER: la Item 2 up there a statomont of 2 a shortcoming of the codos, or is that a statomont of a 3 shortcoming of a model?

4 MR. EISENBERG: It's the methodology which is 5 perhaps not as good as it should be.

6 MR. STEINDLIR Methodology, for what?

7 MR. EISENBERG: For treating scenarios, for 8 defining scenarios and the relationship betwoon datormining 9 the probability of a scenario -- which the current approach 10 is to have the probability of the scenario independent of 11 the sampling that is done for the consequence modules.

12 It appears that in some cases to treat the problem 13 realistically you nood to incorporato some of the sample

(~)

%)

14 variables into the dotormination of the probability. A 15 simplo example is that if you have a larger drill you have a 16 higher probability of intersecting the repository. So, 17 that's not a big change but it is a change.

18 That can be treated as a sample variable to 19 datormino the consequences, but wo are currently treating 20 the probability of the drilling scenario as something 21 independent of any of the samplo paramotors.

22 (Slides.) ,

23 MR. EISENBERG: Flow and transport, we are using a.

24 very simplified representation of the flow system. The 25 question is, is it an appropriate one. I suspect we neod to O ^"" ai'ev $ ^SSoci^Tes' 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 ,

(202) 293 3950 ]

1

1 96 1 do many more auxillary analyses to answer the question. We 2 recognize that the effects of heating.in the early phase of 3 the repository could have a profound effect on flow and 4 transport, and we need to address that also, which we have 5 not done in Phase 2.

6 (Slides.)

7 MR. EISENBERG: This gets back to an issue =that 8 was-about the system code which is that if you begin a 9 scenario such as the intrusion of a volcanic dike-it could 10 have an effect on the flow system. This_means that you have 11 to have coden that-can reflect changes midway in a 12 simulation. The Sandia codes were designed to have this-13 restart capability but it was not exercised in Phase 2. We

() 14 15 really need to look at this, to see how big an effect it has on performance.

~

16 The gas flow analysis has many assumptions that-17 need to be evaluated further. Certainly, we found out that 18 even a simple one dimensional steady state flow module --

19 two dimensional, I am sorry, for the unsaturated Zone - is 20 computationally intensive. We would like to see if there's 21 a way to simplify the code for use in this system code.

22 A point related to potential changes in the 23 regulation is that if we need to calculate concentrations 24 rather than integrated discharge as part of your performance 25~ menvurr, it inserts additional technical issues that need to O . ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1812 K. - Steet, N.W., Suite 300

- Washington, D. C, 20006 (202) 293-3950

97

(-) 1 be resolved.

2 MR. POMEROY: Mick.

3 HR. APTED: Norm, I am asking you to speculate.

4 If a change is made to some sort of dose standard, does 5 Carbon 14 begin to fall off the list that has all this vapor 6 transport issues? I mean, the importance of that, is that a 7 topic dependent upon the --

8 MR. EISENBERG: Let me answer it, and then see if 9 Dick wants to add anything to it. I would say that it 10 depends on the approach to the dose calculation. If you are 11 looking at the dose in the nearby neighborhood, I believe 12 the feeling is -- we don't necessarily have the 13 quantification for it right now -- the feeling is that the 14 dose in the nearby regions, say 50 kilometers, would be l()

15 small. It would not be a big issue.

16 The basis for the release limit as I understand 17 it, and perhaps Dan Fehringer can speak to this. The basis 18 for the release limit in the EPA standard is the worldwide 19 dose, assuming that the carbon 14 circulates in the world 20 atmosphere and continues to recycle. I presume that if we 21 used that approach to calculating doses it will still be an 22 issue.

23 MR. APTED: That's about what I --

24 MR. EISENBERG: How is that for speculation. One 25 of the problems with the source term code is that different O ^"" ai'ev a Associates. 'id.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

l 98 1 kinds of releases occur on different spatial scales. The 2 gas releases probably occur on scales of drifts or perhaps 3 the entire repository, whereas the liquid releases would 4 appear to occur on scale of the waste package.

5 The mechanical and seismic failure waste packages 6 is, I believe, an area that has been barely scratched -- no 7 pun intended -- and it probably needs to be investigated 8 considerably more.

9 MR. POMERoY: At the moment these four codes don't 10 consider glass at all, do they?

11 MR. EISENBERGt No. We have stuck to spelit fuel 12 as the waste form.

13 MR. POMEROY Right.

() 14 15 (Slides.)

MR. EISENBERG The source term code is a mini-16 system code in itself. It involves many complicated 17 processes. In some ways it's harder to do than the system 18 code because of the near field effects. One of the things 19 that was done is that the EBSPAC program which is doing the 20 more detailed modeling for the source term, has recently 21 been refocused so that it will provide the simplifications 22 that can be incorporated into a source term code.

23 MR. OKRENT: With regard to possible seismic 24 effects on cans, if the fault ruptures right through the 25 repository then indeed you have a near field acceleration O ^"" ai'ev $ ^ssoci^Tes. 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

l 1

99 1 about which not so much is known. Do you have to institute I

2 some kind of special study in that regard, or what would you I

3 do?

4 MR. EISENBERG: That is something we have not 5 treated so far. I think we would, as a first cut use a very I 6 simplified approach, assuming that if there was a fault 7 movement that it would sheer some of the vaste canisters -

8 that were located in the region where the movement occurred.

9 But the other part of the problem is that even for 10 seismic activity due to fault movement away from the 11 repository itselt /ou will get accelerations that can cause 12 a number of effects, shaking the packages, moving them, 13 hitting them against the sides of the emplacement hole,

() 14 15 dislodging rocks causing spallation of the emplacement holes.

16 All that has to be factored in with the fact that 17 the waste packages over a period of time are degrading and 18 are not as structurally sound as they were on initial 19 emplacement. It's kind of a combination of several 20 different effects that could and will be difficult to model.

21 MR. OKRENT: The earthquake itself is unknown as 22 to when and how large and if.

23 MR. EISENBERG: Yes, but -- not to get too far out 24 of my field of expertise. There are methods that have been 2 r> used in fact even used in the regulatory context, for Q ANN RlLEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

100 1 determining the likelihood of certain magnitudes of

\# 2 earthquakes at a particular site and for assessing the 3 accolorations that will occur.

4 I think we would use those approaches. It's an 5 additional level of uncertainty, that we would have to 6 convolve the two types of models.

7 MR. OKRENT We don't have any nucicar reactors, I 8 believe, knowingly placed within a quarter mile of a fault.

9 That kind of analysis -- there once was one proposed. You 10 may have to hunt to get the analytical techniques. This is 11 what I am suggesting.

12 MR. POMEROY Mick.

13 MR. APTED: Just a point of information, because I

(} 14 15 worked with the EPRI group on a different project.

aware that they just finished .' rather extensive expert I am 16 judgment elicitation on exactly this issue, seismic 17 disturbance and its consequences -- not so much the ,

18 consequences but the recurrence rates and effects and 19 displacements and so on.

20 I wasn't part of that, but the upshot basically is 21 that it's a really looking to be a no-never mind. Again, 22 it's one of these things that everyone's gut feeling turns 23 out to be when you get the experts down it doesn't really 24 amount to -- it doesn't look to be a real show stopper.

25 That's second hand information. I know that report is Q ANN RILEY_ & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950 l

.. . . . .. . . . . . . . . . . .-j

101 1 coming out soon, on exactly this issue, l 2 [ Slides.)

1 3 Mit. EISENBERG We feel that some areas-requiring-l 4 additional work are how the water gets into the waste ,

5 package if at all at Yucca Mountein, how the water would 6 contact the waste within the package, and what near field j i

7 hydrologic processes can produce this kind of wetting.

8 For disruptive consequences, task five, the 9 intractions are very important. They have to be considered 10 at a fairly high level in the design of the system code.  !

11 The disruptive consequence modeling is, again, very labor 12 intensive because you have to treat a whole series of very 13 disparate processes and events and consider a lot of 14 couplings for which there are not sufficient data. Some of ,

15 the processes involved are inherently _ difficult to model.

16 MR. STEINDLER: Does your analysis at the moment 17 includo pr< visions for recovery, retrieval?

18 _MR. EISENBERG:' _ No .

19 MR. STEINDLER: Do you plan to at least 20 contemplate the probleais potentially introduced by design 21 considerations?

22 MR. EISENBERG: I would answer'it this.way. The 23 current recovery requirement is.for the pre-closure period, 24 and we are focusing _our~ efforts on_the' post-closure

~

25 analysis, assuming that the repository is sealed-_with-a-

^

ANN RILEY &; ASSOCIATES, Ltd.

Q-Court _ Reporters

-1612 K. Street, - N.W., Suite 300

- Washington, D. C. 20006 -

(202) 293 3950

..=. =.- = ..--..:. .

. .a

102 1 marker and you walk away from it.

2 Now, again, the new legislation may change all of 3- that. But I think certainly in the past we have taken'the 4 viewpoint that we are not considering recovery because it's 5 not part of the post-closure performance. I am sure in our 6 planning phase we will consider whether we have to change-7 that point of view.

8 MR. POMERoY: Norm, before you go ahead, let me 9 just ask you to repeat a little bit of yesterday's 10 discussion about the Ghost Dance Fault. You are not 11 treating the fault at all in this treatment, as I understand 12 it.

13 Could you'just briefly tell us what you plan to do-

, 14 in terms of the existence _of the Ghost Danco Pault and two, 15 the finite width of that fault if at that finite width >

16 extends at depth. f 17 MR. EISENBERG: I think it's almost true that we j

'18- are not treating it. But we are treating'it to the extent _.

19 that the work we-are using to define the' seismic hazard at [

-20 the repository incorporates movement of the Ghost Dance.-

21- Fault in calculating the probability of various magnitudes -;

22 of' acceleration.

23 In terma of the flow models, I think we would rely- ,

24 on some_of the detailed auxillary analyses:.to look at.

25 ~ There are a lot of questions. First of all, what changes to: ,

t ANN - RILEY- & ASSOCIATES, Ltd.

Q' ' - Court Reporters - i

._1812 - K. Street,- N.W., Suite 300--

1 Washington, D. C. 20006 ,

-(202) 293-3950: .i

l 103 1 the hydrologic system a fault movement produces, whether it 2 creates a flow barrier or creates a flow pathway or both.

3 It might be a horizontal barrier and a vertical pathway.

4 We have to look at some detailed analysis of what 5 it would do to the hydrologic system, incorporating of 6 course the near field effects, since it would be in the 7 region of the repository. But, not in Phase 2.

8 MR. POMEROY: Fine. Proceed.

9 MR. EISENBERG: For sensitivity and uncertainty 10 analysis, this is an area where I think we are in relatively 11 good shape. We probably could evaluate some additional 12 methods and other sampling schemes. There's a lingering 13 doubt about the effectiveness of doing sensitivity analysis

() 24 15 using the step wide progression when the variable ranges are several orders of magnitude and we know that the models are 16 highly non-linear.

17 (Slides.)

1R MR. EISENBERG: One of the things that came up 19 that was a surprise, I think, is that some of the variables 20 that we are using for the sampling to run the system code 21 are not very instructive in terms of providing ideas about 22 performance. We may need to aggregate some of those in 23 order to define variables that do have impact on 24 performance.

25 Perhaps a good example is, we are locating the O ANN RlLEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

104

.

12 MR. OKRENT: I am not quite clear why you call 13 something fuel and something oxide and something grain Yb,:

14 boundary layer.

15 MR. CODELL: Let me dig out another slide here 16 that Mick Apted is responsible for. I will put that side by 17 side up here.

18 [ Slides.)

19 MR. CODELL: This is what we think fuel rods look 20 -like. Actually we know, because it's not me, personally..

21 If;you look at a slice of a fuel rod here you have cracks in:

22 the fuel after it's.used. You also have-grains. The. grains _ l 23 are these small areas here. The fragments'are the areas.

24 between these major cracks.

25 This is about on the order of 20 microns-diameter.

ANN RILEY & ASSOCIATES, Ltd.

1 ---

Court - Reporters - >

1612 K. Street, N.W., Suite 300

- Washington, D. C. 20006 q

-(202) 293-5350 j

.]

112 1- This; bigger-pieces are on the order of two millimeters 2- diameter. ;This is quite evident from slices of. micrographs 3 of slices of spent fuel; that.the oxide-starts'to penetrate 4 from these grain boundaries into the_ grains. You see a 5 -little rind forming around each part of each grain which 6 gets bigger with time.

7 The model here --

8 MR. OKRENT: When you say the oxide, do you mean -

9 - you started with --

10 hR. CODELL: You start with UO2 and then there's a 11 whole series of oxides you would get, depending.on the' 12 temperature. We are starting with UO2, and the end product 13 is U307 in our model.

14 MR. OKRENT: When you say oxide you mean U3077 l 15 MR. CODELL: Yes, U307. -o 16 MR. OKRENT: And fuel, you mean UO2.

17 MR. CODELL: I do not mean UO2. UO2-is fuel.

18 U307 is the oxide.

19 MR. OKRENT: Thank you. I wanted to understand--

=20- that. The grain boundary layer -- it's all made of grains.-

21 MR. CODELL: .The outer layer is just a ,

22 mathematical. convenience. We are modeling the diffusion 23 through_the-grain boundaries as if it were some sort of a-24 continuum, not actual grain boundaries. It's-just a way 25 that we are dealing with the diffusion:that seems to work -

ANN RlLEY & ASSOCIATES, Ltd.

.-h Court Reporters l

1612 K. Street, N.W.,- Suite 300 Washington, D. C. 20006 (202) 293-3950 ~ ,

o

113.

1- okay for our purposes.

2 It's an abstraction of the true picture of the 3 model.

4 MR. OKRENT: Thank you.

5 [ Slides.)

6 MR. CODELL: Based upon the model that DOE is 7 using and Sandia is using for their carbon 14 analysis --

8 it's a 2-D finite difference model that assumes that you are 9 decling only with water saturated air.

10 The model itself is steady state but the 11 temperature is-changing with time. The transients in the 12 gas flow we assume are very fast compared to other 13 transients, which allows us to deal with it as steady state.

() 14 15 We took the equations of the model but made a few significant enhancements, Rex and I did, to the model for 16 our own needs.

17 Once we calculate the flow we assume -- using the 18 2-D model we look at transport of carbon as if it were 19 discreet particles moving along with the velocity field from 20 the repository to the surface. We also consider. retardation 21- of the carbon in the geosphere. It moves slower than the 22 gas itself, the other gas.

23 MR. STEINDLER: Is this the gas velocity field?

24 MR. CODELL: Gas velocity field, yes. Only gas.

25 This is the simplified cross section we used in the Q ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street,- N.W.,' Suite 300-Washington, D. C. 20006 (202) 293-3950

1 l

114 l l

1 analysis, a slice of Yucca Mountain with parallel layern but 2 tilted with respect to the horizon. ,

3 [ Slides.]

4 MR. CODELL: Temperature is the main driving force 5 in the model. We calculate temperature.as a function of 6 time by assuming that the surface of the earth is isothermal 7 but varying with elevation. There is a heat load for decay 8 of the waste radionuclides at the repository. We have a 9 geothermal flux entering. We calculate with a separate two 10 dimensional transient model, with the temperature at each 11 point in time we evaluated it at.

12 (Slides.]

13 MR. CODELL: The next slide illustrates what a 14 typical result from our flow modeling. This shows the

(}

15 chimney effect with air being drawn from the surface not be 16 allowed to penetrate the water table, which is assumed to be 17 no flow boundary, being moved up through the middle of the 18 repository because of the effects of heat.

19 At points in time we use these velocities to 20 release particles along this boundary of the repository, in 21 order to calculate the time it takes to get from here to 22 there and for our subsequent calculations _of performance.

23 The model for transport of the carbon 14 is very 24 simple. It's simply that the particles move along with the 25 flow and that there is a retardation factor caused because ANN RILEY & ASSOCIATES, Ltd.

[]

Coud- Reponers 1612 K. Street, N.W., Suite 300

- Washington, D. C. 20006 (202) 293-3950 i

115-1 of the interaction of the carbon with the other components

-O- 2 in the geosphere.

3 In order to get a better handle we wanted to do an 4 auxillary analysis which looked at a much more complete set 5 of at least chemical equations, to deal with what the true 6 factors in the transport would be. This is a research type 7 of model which was a one-dimensional column. At each point ,

8 we solved bounds equations for carbon chemistry.

9 We pumped into this transient flow, saturation and 10 temperature which we have taken from another analysis.

11 The chemistry shown in the next slide, we 12 considered all of the likely things you would find in the 13 geosphere; water, carbon dioxide, bicarbonate, hydrogen ion, 14 carbonate, hydroxide, water in the gas and liquid phases,

}

15 carbon dioxide in the gas and aqueous phases and calcite 16 solid. We also considered for the sake of maintaining the 17 PH, some representative cat ions.

18 [ Slides.]

19 MR. CODELL: The next picture shows a depiction of 20 our one-dimensional column. The column represents the 21 middle of the repository, so it had a gas flow fed into it 22 from the bottom going out to the top. The repository is 23 located here, which is a source of both heat and carbon 14.

24 At the bottom we assumed that we had ambient but non-25 radioactive carbon dioxide entering at the level it would be ANN RlLEY & ASSOCIATES, Ltd.

i-Q l

Coud Reponers 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

116 1 measured in the ground.

.C~s)' 2 The column consisted of 145 segments, each one of 3 which was a calculation of the geochemical system using an 4 iterative model of speciation of the carbon between solid, 5 liquid and gas phases.

6 HR. MOELLER: A question. The CO2 coming in at 7 the bottom, is depleted at C14.

8 MR. CODELL: That's correct.

9 MR. MOELLER: Then, you are adding to it.

10 MR. CODELL: Right. We are adding to it at the 11 repository. That's the only place radioactive carbon enters 12 the system.

13 MR. MOELLER: Does the fact it's depleted have

() 14 15 much to do with the outcome?

MR. CODELL: It's important, because the PH of the 16 water in the system depends on its carbon content. I think:

17 we found early on in this model that if you set this at zero 18 you would get unrealistic levels of PH permeating through 19 the column. You have to try to maintain the balance of the 20 PH.

21 MR. STEINDLER: You are not doing isotopic 22 exchange kinetics in that column, are you?

23 MR. CODELL: No. In fact, the model -- there is 24 two parts of the model. The first deals only with the 25 carbon -- the total carbon. That is what we used to O ANN RILEY & ASSOCIATES, Ltd.

Coud - RepOders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l l

117 1 determine _where the carbon is moving from, between water, 2 liquid and solid liquid and gas phases.  !

3 Then, we treat the carbon 14 as a trace, which 1

4 does not affect the bulk chemistry. It's only a trace.  !

1 5 I will skip the next few slides which only show  !

l 6 the kinds of representative inputs on temperature and flow 7 and saturation, which are part of the calculation. I would 8 like to move to the slide which shows the results.

9 (Slides.]

10 MR. CODELL: These are two sets of results. The 11 first is the total carbon result vertically in the column.

12 This is the first one, is 100 years after closure. What 13 happens is, the repository is heating up. As you are

() 14 15 heating up you start to precipitate calcite because the solubility increases with increased temperature. Because 16 you see an increase in calcite solid precipitating, it's 17 being depleted from the gas phase.

18 However, part of it is moving up the column 19 because it hasn't been captured. It gets boiled off from 20 the water but does not get captured in the calcite. It just 21 moves up at the speed by the advection of the gas.

22 At 500_ years we bave seen the extent of the 23 calcite has grown quite a bit. At longer times it-has grown 24 in length as the isotherms penetrate into the rock further.

25 However, overall, it's redissolving back into the column.

O ^"" ni'ev a Associates' 'ta-Coud Reponers i 1612 K. Street, N.W., Suite 300  ;

Washington, D. C. 20006 '

(202) 293-3950

118 At 4,000 years there is very little calcite left.

()

fs 1 2 The next slide shows the effect of those phenomena 3 on the transport of the carbon 14.

4 (Slides.]

5 MR. CODELL: What happens depends on when you 6 release the carbon 14. For this particular slide we are 7 releasing it at times zero, slightly below the repository.

8 Part of the carbon 14 gets caught up in the 9 calcite, the solid line and remains fixed in place because 10 the calcite can't move. The large part of it starts moving 11 out of the column, most of it in fact, in the gas and liquid 12 phases. By 2,000 years most of the bulk of the carbon 14 13 has left the column entirely. Only what is left in the 14 calcite remains. The calcite, as it redissolves back into r

)

15 the geosphere, it becomes a source of carbon 14 to the 16 column.

17 These'results are very interesting. LGt me just 18 tell you the bottom line of this, at least for the 19 parameters that we looked at in this auxillary analysis.

20 It gave us a fairly good handle on the kinds of 21 retardation factor we should use in our performance 22 assessment which could not duplicate this complexity. This 23 single run took about ten minutes of Cray time. It would be 24 impractical to include it in our overall analysis. It gave 25 us the confidenco to use a far simpler single retardation

()

ANN RlLEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 j (202) 293-3950

I 119 coefficient,.-because while calcite uptake was interesting it:

- l' 2 did not~ amount to a whole lot. We could get by_with 3 something much simpler.

4 That is the extent of the auxillary analysis.= .

5 MR. MOELLER: Let me pursue my previous question a-6- bit. Would-not the retardation -- and Dr. Steindler'could 7 help me with this -- would not the retardation of the C14;be 8 greater than you would anticipate, due to the fact that the.

9 carbon naturally present-at depth in the repository will 10 have been depleted of the normal carbon 14 component that 11 carbon up at the earth surface has, such as the seuss effect-12 that we talk about.

13 MR. CODELL: It's something that I hadn't thought 14 about until Monday, when Jerry Boak half. jokingly said_that

(

15 the overall effect of the repository.would be to deplete.

16 carb3n 14 from the environment; therefore, is a net benefit.

17- I don't usually take Jerry seriously --

18 [ Laughter.-)

19 101. CODELL: --- but he had a good insight there.

20 MR. MOELLER: - The burning of fossil 1 fuels,_of 21 course, does reduce concentration of C14. .Is this totally 22 insignificant? I 'eun asking.

23 MR. STEINDLER: My guess is, the first order of 24 effect.of isotopic selection is insignificant ~.in~ comparison 25 to the uncertainties of that model. You won't~seeLit. I am ANN RILEY &- ASSOCIATES, Ltd.

=(]- COud Repoders 1612 K. Street, N.W., Suite 300

^

Washington, D. C.. 20006

(202). 293-3950 -

_ . . . - . -._ _ - . - ~ _

~120-

1- ' guessing.-

3 2 -

MR.D BOAK:~ Characteristically, looking in 3 geological processes when you measure fractionations of the 4 carbon isotopes,~you generally-measure':it in per all.

5 Sometimes it gets up into the tens per mil, so it might be a 6 few percent effect.

7 MR. POMEROY: Go ahead, Dick. Can we finish up in-8- a few minutes?

9 MR. CODELL: We are just about finished.

1 10 [ Slides.)

11 MR. CODELL: I just wanted to talk just briefly-12 about how we are using this. First of all', of course, we 13 are summing the releases at the atmosphere in order;to show-14 compliance or non-compliance with the EPA cumulat'ive: release ,

15 standard. -But we are also factoring carbon 14.into the dose 16 assessment, both at concentration in.the air looking at the

17. flux through the surface and assuming that it gets partially.

18 mixed in the air and is inhaled by nearby populations.

19 We are also, along with all the other.- ,

20 radionuclides, considering carbon.14 that enters into.the- -

, 21 water pathways as well.

22 [ Slides.]

23 MR. CODELL: In summary, models for release.from-24 the waste form transport through t.ie-advective -- gas: flow: -

25 and advective transport through the-geosphere to.the- .

.Q ANN RILEY &- ASSOCIATES, Ltd.

. Court Reporters

.1612 K. Street, N.W., Suite -300 Washington, D. C. 20006 '

(202) 293-3950-D ,

u . . - _ _ _ __ _ _ .

- . _ . . = . _ _ _ .

121 1 accessible environment and looked at an auxillary analysis 2 of carbon 14 geochemistry and geosphere.

3 The models are simple but have provided quite a 4 bit of insight on both the physical properties of release in 5 transport at a repository and unsaturated tuff, and also 6 perhaps the importance or lack of importance of carbon 14 to 7 the health and safety of the public.

8 Our tentative conclusion in this regard is that 9 from some very preliminary analyses, is that we have 10 significant concerns about the likely size of the release of 11 carbon 14. Therefore, difficulty in demonstrating will 12 comply with the current EPA cumulative release. At the same 13 time, we are not so sure that these are important health

, 14 effects.

15 We used these conclusions in our dialegue with 16 EPA, to perhaps try to influence them to reconsider their 17 release limits on carbon 14. Thank you.

18 MR. STEINDLER: What is the travel time, if I can-19 couch it that way, from repository to surface by advoction 20 alone?

21 MR. CODELL: It could be very short. It could be 22 just a few tens of years, according to some of our most-23 pessimistic calculations. It varies over a wide range.

24 MR. STEINDLER: Assuming a uniform saturatior. of 25 tuff between the repository and the surface at some percent?

i ANN RILEY & ASSOCIATES, Ltd.

4 O(_/ Coud Repoders 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

1122 L15 MR.ECODELL: The model does not actually take

.O i2  : saturation into. account. It's a single phase-model-which

'3 : considers only' gas flow' saturated with water vapor.

4 -However, ir. determining we consider-:the saturation-in 5 choosing the hydrologic parameters; that-is, the

-6 permeability. For a more saturated it would have a lower 7 gas permeability.

8 MR. POMEROY: Thank you, Dick.- With Margaret's. )

9 consent, we have decided to change the order of the agenda.

10 We are going to take a one hour lunch break, until 1:20. At 11 that time we will reconvene, and go on with.the item 12 originally scheduled for 11:20.

13 (Whereupon, at 12:25 p.m., the meeting recessed, 14- to reconvene at 1:20 p.m., this same day.)

i- 15

-16 t

17 'l e-l' 18  ;

-j 19 i 4

a 20

, 21

~ 22 1 23 i.;

24 j 25 h ANN RlLEY & ASSOCIATES, Ltd. j Court Reporters -

1612 - KJ Street, N.W., ' Suite : 300 Washington, D. C. 20006 ..

(202) 293-3950 l q

123 1 AFTERNOON SESSICN 2 (1:27 p.m.)

3 MR. POMEROY: Margaret, please proceed.

4 MS. FEDERLINE: Dan Fehringer will now talk about 5 the use of expert judgment in NRC's licensing process.

6 (Slides.]

7 MR. FEHRINGER: One of the components of a 8 performance assessment i judgment. Over the years there 9 has been a lot of concern over recent years about how expert 10 judgment will enter into a performance assessment and in 11 pari tcular about how use of judgment will be evaluated 12 during the course of the NRC's licensing reviews for a 13 repository.

14 Over the last several months I have been trying to 15 learn what I can about the nature of our licensing process, 16 in particular what the requirements are for entering 17 information as evidence in the licensing process and how the 18 process evaluates the validity of that information and what 19 weights to place on the different types of judgments.

20 Let me give a disclaimer before I go any farther.

21 One of the sources I consulted was our Office of General 22 counsel but I have also consulted other sources, 23 particularly Staff members with experience in licensing and 24 I have tried to synthesize and interpret all of what I have 25 heard, so I do not represent the General Counsel and the

() ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

124 1 General Counsel may some day tell you something different 2 than what I'll tell you today.

3 First, let me begin by noting that a licensing 4 decision will be based on a combination of fact and opinion.

5 Sometimes in the past we have argued fairly strongly that we 6 are leery of substituting judgments for data that are 7 reasonably available or obtainable. That is not to say that 8 we think everything in the license application can be proven 9 as hard, factual data. We know that judgment will play a 10 big role and we want to plan for that as well as we can, 11 The NTtC's licensing process provides a forum, 12 which is the hearing, where facts and opinions both can be 13 evaluated. The parties to the licensing process have that 14 opportunity at the hearing to present their views on the

[)

15 meritr, of both facts and opinions that are contained .in the 16 license application.

17 In this talk I'll try to describe very briefly 18 what the licensing process consists of, then I'll talk about 19 the role of expert judgment at each phase of that licensing 20 process, and finally I'll try to look at some of the formal 21 methods that have been recommended for developing and-using 22 expert judgment and try to see how those match up with-the 23 requirements of our licensing process, to see if there are 24 any incampatibilities. '

, 25 (slide.]

O ^"" aev $ Associ^Tes' 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

)

125 1 MR. FEHRINGER: First, these are the phases of the 2 licensing process that lead up to a decision whether to 3 authorize construction of a repository. It begins with the 4 pre-licensing consultation phase that we have been involved 5 in for several years now. This is an informal phase of 6 activity where the NRC Staff and the potential applicant 7 negotiate such things as plans for studying a site. That 8 phase ends when DOE prepares and submits a licence 9 application.

10 The second phase then is the NRC Staff's review of 11 a license application. The Staff will prepare its own 12 safety evaluation report. The Staff's report will of course 13 rely heavily on the information contained in the application 14 but it may also include the NRC Staff's own analyses of 15 repository performance and the NRC Staff's own judgmental 16 interpretations of some of the factual information that is 17 in the license application.

18 MR. POMEROY: Dan, by and large that will_ include 19 though either an acceptance or a rejection of the 20 applicant's position, will it not? That second step?

21 MR. FEHRINGER: Yes. The Staff's views on whether 22 the application is acceptable or not will be recorded in 23 that Safety Evaluation Report. ,

24 MR. POMEROY: So at some point if it is 25 acceptable, at some point then the Staff becomes in l I

Q, ANN RILEY & ASSOCIATES, Ltd.

Court Reporters

! 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 L -_ ---

126 1 agreement with in essence the applicant?

2 MR. FEHRINGER: Yes, that's correct.

3 When the Staff has completed its report, then that 4 report plus the original application are forwarded to the 5 licensing board and a hearing will be held for a repository 6 and other parties who think they may be affected by the 7 proceedings may ask to participate in the hearing as formal 8 parties and they will have the opportunity then to identify -

9 the issues they think should be in contention and to present 10 their case.

11 Finally, after the board has reached a decision, 12 there is one opportunity for appeal. In the past there were 13 two appeal opportunities -- there was an appeals board and 14 then if that decision was still unsatisfactory it could be i

}

15 further appealed to the Commission. The appeals board has 16 been abolished and now appeals go directly to the Commission 17 or the commission may review a licensing decision at its own 18 initiative. There does not have to be a formal request for 19 appeal for the Commission to review a decision.

20 MR. POMEROY: Isn't there another step in that 21 also, Dan, in the sense of there is an appeal process to the 22 courts, after that -- or some time during that process?

23 MR. FEHRINGER: Yes, there could-be an appeal to 24 the courts but that would be outside of our own licensing 25 process.

O ^"" ni'ev a Associates. 'ta-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

127 l L g^g 1 MR. POMEROY: I understand.  !

\u-] 2 MR. FEHRINGER: So that's why I didn't list it l

i 3 here but certainly if someone thinks we have violated the I i

4 law in some way, they could go to a court and seek redress l

5 there.

6 [ Slide.]

[ l 7 MR. FEHRINGER: The rules of expert judgment 8 during these phases are, as I have tried to indicate on 9 slide 4, first, during the pre-licensing consultation phase 10 the NRC Staff's involvement with DOE's use of expert 11 judgment is relatively limited. DOR will use judgment for a 12 lot of things that probably should not concern us -- the i

13 internal management of their own program and a lot of

(~'\ 14 decisions about a repository that do not directly affect V

15 safety, and we are taking the position that we have no 36 reason to try to review those types of judgmental decisions.

17 The only cases where we want to have an oversight 18 of DOE's use of judgment is where we_think there may be a 1

1 19 safety implication, in particular where we think that a 20 decision may affect the adequacy of the data that would be 21 collected during site characterization. Arguably that's a 22 big enough loophole to drive a-truck through but still the 23 principle is if it is not safety-related, if it is just

, 24 DOE's management of its own program, we do not want to try l

l 25 to watch over DOE's shoulder, l

ANN RILEY & ASSOCIATES, Ltd.

Coud RepOders I 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

128 i During our review of the license application, it

/-~g

\/

2 is not at all obvious how we will review doe's use of expert 3 judgment.

4 There seem to be two schools of thought shaping 5 though. One holds that the Staff should be concerned ebe 6 the reasoning that underlies the judgment but not concerneC 7 with the process by which a judgment was reached. The 8 alternative school of thought holds that process is 9 important and that the process by which a judgment if 10 formulated has a strong effect on the validity of that 11 judgment. My guess is we'll combine the two a nd have some 12 kind of a hybrid approach to developing our review but we 13 just haven't done our homework yet to find out what our f~}

v 14 underlying philosophy will be when we review the use of 15 judgment in the application.

16 At the time of a hearing judgnents are presented 17 in a written form by experts and then that written evidence 18 is followed up by the expert swearing to the validity of 19 that information that was presented.

20 Following the presentation of evidence a witness 21 is subject to cross examination by the other parties to the 22 licensing hearing and that's perhaps one of the most 23 significant parts of the whole process. Cross examination 24 is the opportunity for each of the parties to probe the 25 reasoning and the underlying basis that supports the

(')

ANN RlLEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 )

(202) 2 0 ' 10

.a !129-

~

-1 judgments of theLother parties to the proceeding. If a cross 2 examination Jbs carried.out well, it will' establish a recordi 3 that the_ licensing: board'can use to determine _which 4 . judgments really have merit and which!ones_do=not or,_ stated.

5 alternatively, how much weight to place:on any of the.

6 judgments that~are presented.

7 MR. STEINDLER: Is the process that you have 8 outlined-here taken from the normal legal approach to-9 looking at single experts, or is this in light of your 10 previous comments what you perceive to be the way the 11 examination -- which is really-a collective process rather 12 than an individual process?

13 In other words, where did you get that procedure

() 14 15 that you just-outlined, having previously said;you're not clear'how you are going to handle an examination of DOE's t

16 use of expert judgment?

17 MR. FEHRINGER: What I am not clear about is how-r 18 the staff will conduct its own review of the application' 19- before_the' hearing begins. The hearing process Obs pretty 20' well established both in~our regulations and in the 21 precedents of previous hearings that"have been carriedJout ,

22 -at the1 agency but it is the Staff's: review prior to-23 ' submittal ~of the application to a-licensing board that is a

-24: 'little bit' questionable.

25 MR. STEINDLER: Have there-been instances in_which.

Q: -ANN RlLEY & ASSOCIATES, Ltd.

-Coud Repoders 1612 - K. Street, - N.W., Suite 300

- Washington, D.~ C. 20006 (202) 293-3950

~

. ,  %. . ~ , -, - , ,

t 130 gr'T 1 the Atomic Safety and Licensing Board or some such board-O 2 have looked at expert judgment where expert judgment is-in 3 the context of how it is likely to be applied-to the high 4 level wastc repository, certainly as we've seen it?

5 MR. FEHRINGER: Yes, there are several examples.

6 I am not personally familiar with them but at a 7 workshop last month that DOE had on this subject several 8 examples were presented.

9 One of the drawbacks is that apparently the nature 10 of the intervention in those cases was not as vigorous as it 11 can be expected to be in this case and so there was not the 12 probing of the expert judgment that we will expect to occur 13 in a repository license, so I am not sure how much weight to

() 14 15 place on those as illustrative examples.

MR. STEINDLER' Well, I certainly would be 16 interested in being pointed to some reference.where expert 17 judgment as it is likely to be used here -- and which is a 18 fairly large effort _ collective process -- was handled by a 19 licensing board.

20 MR. FElfRINGER: Leon Ryder presented a paper at 21 that workshop I just mentioned and he is here today.

22 Perhaps he can fill us in on a little bit of that.

23 [ Slide.]_

24 MR._FEHRINGER: Finally,-let me make one point 25 about the appeal process. .There should be no new

'V ANN RILEY- & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

131 1 introduction of' expert judgment at the appeal. The appeal 2 is a Commission review of the record that is established 3 during the hearing. The hearing is a one-shot opportunity to get all of the cases of all of the parties on tlue record 4

5 and if the Commission finds fault with that record, the 6 Commission will refer the case back to the hearing board for 7 further examination but at the appeal itself there is no new 8 presentation of judgment and there is no evaluation of it 9 other than to the extent that the Commission reviews the 10 record that was established during the hearing.

11 (Slide.)

12 MR. FEHRINGER: One of the basic questions about 13 judgment is is it admissible as evidence, and as near as I

() 14 15 can determine there are really only two criteria for admitting evidence to a licensing hearing. They are the two 16 listed on Viewgraph No. 5.

17 It is necessary to establish the expertise of the 18 witness and it is necessary to show that the evidence the 19 witness is offering is relevant to the specific contention 20 that is at hand. If the contention is the probability of 21 volcanism it's not appropriate to have a witness talk about 22 the consequences of volcanism. It's necessary to limit the 23 evidence to the issue that is being discussed.

24 The second item here says the Federal-Rules of 25 Evidence do not apply and in a strictly legal sense that is O ^"" ai'ev & Associates. 'ta.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

132 1 true. The Federal Rules of Evidence were developed for the 2 Federal courts and the NRC is not a court. However, one 3 should put a mental asterisk beside that. Courts sometimes 4 do rely on the Federal Rules of Evidence when determining 5 whether to admit or exclude a particular witness or a 6 particular line of evidence, so the NRC kind of does and 7 kind of does not follow the Federal Rules.

8 MR. POMEROY: Dan, I'd just like to interject 9 there what is perhaps a personal opinion, but it seems to me 10 that if the Federal Rules of Evidence do not apply in that 11 hearing stage, that a clever lawyer could play havoc with 12 the entire hearing record at a later point in time in a 13 court appeal because in a court appeal a lawyer can have 14 recourse to those Federal Rules of Evidence and if he can

(}

15 prove I believe that they were not followed, I think it 16 would throw the whole system into chaos so I would be very 17 surprised if a hearing board were not very aware of that 18 possibility and force the application at least in this case 19 of the Federal Rules of Evidence and the Frye doctrine.

20 MR. FEHRINGER: That's a point of law I just can't 21 discuss.

22 MR. POMEROY: Right. Neither can I.

23 MR. FEHRINGER: Finally, experts need not have 24 first hand knowledge of the facts that they are 25 interpreting. It's entirely possible to lay out some O ^"" ai'av * ^ssoci^Tes' 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 L Washington, D. C. 20006 (202) 293-3950 e

133

,f-s 1 factual information about a fault for example and ask a i

2 seismologist to interpret that information without the 3 seismologist knowing that.the facts were actually gathered 4 in a particular way or having any other first-hand-5 knowledge.

6 (Pause.)

7 (Slide.]

8 MR. FEHRINGER: I wasn't finished with that one 9 yet.

10 Finally, let's see the next and last item is 1

11 expert judgments are limited to the area of expertise of a 12 particular witness. I guess it's a fairly obvious point but 13 the licensing board or the Commission acting as a decision-14 maker rather than the technical expert-himself will decide f)'N s

15 what evidence is admissible and will decide the relevance of 16 that evidence in terms of deciding the overall acceptability 17 of a facility. That is the role that the licensing.toard or 18 the Commission plays.

i 19 Neither the board nor the Commission is bound to 20 accept any of the judgments that are presented to it. It is 21 not a kind of a pick this or pick that. type of a decision-22 making process. The licensing board may reject all the 23 arguments. The licensing board may-synthesize its own 24 judgment from the information presented to it or reach 25 whatever other finding that the board finds is reasonable.

ANN RILEY & ASSOCIATES, Ltd.

C" Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 4

134 g i The board is not bound to accept any particular ,

J 2 judgment that is offered.

3 [ Slide.)

4 MR. FEHRINGER: The question of combining multiple 5 judgments is one that I thought might have the most 6 potential to cause difficulty in terms of compatibility 7 between the trends in formal use of expert judgment versus 8 our licensing process.

9 It turns out that this precedent has already been 10 set. Panels of experts can testify at a hearing on kind of 11 a round-robin basis. It is-not necessary for one expert to 12 testify at a time and then the next. It is possible to call 13 the whole panel, swear them all in, and take their testimony '

i l

() 14 15 kind of as a group.

I still think there could be potential problem in 16 this area though if the weighting method that is used to 17 combine multiple judgments becomes a source of controversy.

l 18 If the panel of experts themselves determine the weighting 19 method then I think there should be no difficulty but if 20 someone external to that panel combines the judgments, that 21 person is influencing the process and he may end up being 22 the source of scrutiny more than the actual technical i 23 experts themselves.

24 I think the idea situation is where a panel of 25 experts can form some kind of a consensus on how their i

l l

O ^"" ai'ev & Associ^Tes. 'ta-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 L

t

l 135 j 1 collective views are to be combined into an overall 2 judgment.

3 Then there may be a potential problem if members 4 of a panel become unavailable to swear to the evidence that 5 they have produced. There could be a significant time lag 6 between when a judgment is formulated and when it is finally 7 reviewed in the course of a hearing. People can become ill 8 or otherwise become unavailable and a collective panel 9 opinion may not carry the weight one would 3:xe it to if 10 some members of the panel are unavailable tc, as I say, 11 swear to their judgments and submit to the cross examination 12 by the other parties, so there is no bar to introducing 13 multiple judgments but there may be potential difficulties

() 14 15 in getting those judgments to carry the weight one would like to have them carry.

16 A related topic is combining the predictions of 4

17 multiple models of a process. To the extent that one assigns 18 weights to the different models, it's kind of an expression

19 of one's level of confidence that each model is correct.

20 That seems a fairly harmless thing to do, but if one were to 21 take the projections of the models, weight them and combine 22 them into some kind of composite projection, there's a 23 little more opportunity for controversy.

24 For one-thing, the hearing process typically tries.

25 to identify what I'd call the " correct" model, in quotes, O ^"" ni'ev a Associates. 'ta.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950-

136

~ '

1 for evaluating safety. Boards tend to look for a projection

('

2 of how a facility will perform and recognizing there is 3 uncertainty they still would like to have one model that 4 guides them to a decision. If a suite of models with 5 different projections are combined tie board may have a 6 little more difficulty accepting that.

7 In any case, if projections are to be combined, we 8 think that the results should correspond to a physically 9 meaningful situation. Two models of an electric switch, one 10 showing the switch would open when it's supposed to and one 11 showing it would stay closed, cannot very well be combined 12 to show a switch half-open and half-closed. It is not 13 physically meaningful. .

1

() 14 15 Any combination of models should not discard or downplay relevant information, particularly regarding the 16 spread of the predictions that the individual models have j 17 produced.

18 [ Slide.]

19 MR. FEHRINGER: So finally as the bottom line I 20 think a lot of people's concerns about use of expert 21 judgment boil down to this question, how can the applicant 22 be assured that its judgments will prevail during licensing?

23 I can find no way to give such assurance.

24 First of all, judgments must meet some threshold 25 of goodness. The applicant bears.the burden of proof, of O ^"" ai'ev & Associates. 'ta.

Coud Repoders 1612 K. Street,. N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

137 1 showing that his facility will be safe, and it's conceivable 2 that the evidence the applicant offers could fail to meet i O that test. I am told that that has not been a problem in 4 past licensing cases. Applicants are not prone to submit 5 weak applications and the intermediate step of Staff review 6 further helps to assure that applications are of good 7 quality.

8 Still I listed it here because we may have a 9 little different type of decisions to make in repository 10 licensing if we are asked to project human intrusion or 11 other human activities for 10,000 years. Things like that 12 can be so speculative that it is at least conceivable a 13 board could find that the applicant just did not meet the 14 test that he is faced with.

15 Perhaps more important is that there is no 16 protection against new information being developed or new 17 interpretations coming up between the time when a judgment 18 is formulated and the later time when is it reviewed during 19 licensing.

20 Finally, the alternative judgments of other 21 parties cannot be excluded from-the hearing. In fact, the 22 whole purpose of the hearing is to provide a forum where all 23 the judgments can be presented and the parties can question 24 each other about the underlying basis. If one tried to 25 guarantee that any particular judgment would prevail, then h ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006

-(202) 293-3950

saa 1 the whole purpose of the hearing would bo thwarted.

2 That would leave one with something of a 3 possit atic view that the applicant is really faced with a 4 great deal of uncertainty not knowing how his application 5 will survivo. I think it is really not as bad as it might 6 soom. The applicant has the opportunity to determino what 7 the full rango of judgments might be that tho various 8 parties will present. The applicant can ovaluate the 9 strengths and the waaknesses of each of thoso and presumably 10 the applicant can be prepared to defend his judgment and to 11 oint out the weaknesses of alternatives that might be 12 presented so I think if an applicant does his homework ha 13 really should be quito confident that his judgments will 14 prevail or if he is presenting weak judgments will know that 15 he's got a weak caso to present.

16 I don't want to leave on a possimistic note that 17 the applicant is solely at the whim of some arbitrary 18 outcomo of the licensing process. That's the end of the 19 presentation that I gave at the DOE Workshop. I have a 20 couple of words to say about Task 2.5, but why don't I stop 21 at this point for questions on this part of the 22 presentation.

23 MR. STEINDLER: Well, you may not have had 24 intended to cast a pessimistic light on the wholo 25 acceptability of export judgment but you have dono an ANN RILEY & ASSOCIATES, Ltd.

@ Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 203-3950

139 1 outstanding job.

2 (Laughter.)

3 MR. STEINDLERt The implication that I hear is in 4 the absence of rulemaking, which as far as I know the Staff 5 has not yet initiated, there are neither criteria nor 6 limitations on the issue of expert judgment.

7 In principle if the hearing board so elects overy 8 variable or constant that has ever boon used derived from 9 expert judgment, used in models or predictions or whatever 10 have you, are subject to the process you have just outlined, 11 and it's survivable or its qurvivability is wholly unknown.

12 How in the face of that the question is, is this a 13 realistic view of a licensing process that is to be

(} 14 15 completed in some finite time period? Without wanting to put you on the spot, let me answer the question.

16 I think the answer is no. The issue then that I 17 would raise for you is at what -- where in these points that 18 you have raised can procedures be provided that do two 19 things, one, remove from what I sense to be in my personal 20 view endless litigation -- the arguments about whether-your 21 expert is smarter than mine and still maintain some 22 significant evaluation of the technical quality of the 23 decisions that have been made by the applicant.

24 The question I have for you is has the Staff given 25 any thought to this issue and, if so, what is the view that O ^"" ai'ev

• Assoc'ATes 'ta-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

140 1 you have boon able to extract?

2 MR. FEllRINGER: Thoro are two ways that we can 3 address issues separato from what I have described here.

4 One is through the rulomaking procons. Thore the notico and 5 comment process serves the forum for arguing an issue rather 6 than having to do it lator in a hearing, no to the extent 7 that we can tako issues through rulomaking and close them 8 out that way, we can nettle them before thi.; hearing.

I 9 The other way is to have separato hearings on l 10 specific issues. I have described only one hearing for 11 review of the licenso application but it is conceivable to 12 t.glit this up into kind of subsidiary hearings which could 13 addroso very specific inouco and perhaps be conducted simultaneously or some of them conducted ahead of the main

() 14 15 hearing in order to close out lasues. That in another 16 mechanism that could be unod to try to koop within the throo 17 year licensing timo period.

18 We don't have specific plano right now to pursue 19 either one of those although wo recognize both of them are 20 opportunition. When I describe in a ninute the Task 2.5 21 work, that is aimed at doing some of the homework that might 22 load un to development of guidance or perhaps of rulemaking.

23 MR. POMEROY: Okay, why don't we go ahead then, 24 with 2.5.

25 Thora are about a hundred questions that I would O ^"" ni'ev a Associ^Tes, 'id.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

I 141 1 like to ask but I won't.

2 MR. FEHRINGER: Our general approach to dealing 3 with expert judgement is outlined on this wiring diagram, 4 which is really net very legible, but in your hand-outs I 5 think you can make it out.

6 (Slide.)

7 MR. FEHRINGER: We've had, on the left column, 8 something we call the Winkler/ Hora study, named for the two 9 authors -- two decision analysts with experience in using 10 expert judgement -- have been asked to do a study on how 11 well exports perform and what considerations one should keep 12 in mind when carrying out an clicitation to make sure that 13 the judgements are as good as they can be.

() 14 15 That is complete now. I'm not sure if we have the final report in hand, but the study itself is complete. Wo 16 also have a study on-going on the ways expert judgement is 17 used in other gootechnical areas, in particular, exploration 18 for mineral resources -- trying to getting any insights into 19 how people deal with uncertainties 11 those fields.

20 MR. HINZE: Who is doing that, if I may ask?

21 MR. FEHRINGER: I don't know. Do you know?

2 '. MS. FEDERLINE: We are doing it through the 22 Center. Is Bootie here?

• MR. FEHRINGER: The Center in the prime 25 contractor.

O ^"" niu< a Associ^Tes' 'ta.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 )

(202) 293-3950

142 1 MS. PEDERLINE: It came outside the performance O 2 asseusment program elements, so I'm sorry, we don't.

3 MR. FEHRINGER: On the right hand column we have 4 the Phase 2 work that you've already been told about the 5 last few days. There, to come extent, it was a learning 6 process on how to treat scenarios and how to use judgements 7 to estimate some of the parameters.

8 We're now starting on the Phase 2.5 where we will 9 try to pull together a panel of experts and do an example 10 elicitation. The particular subject matter is climate 11 change at the Yucca mountain site. The experts will ba 12 asked to predict the intensity and the frequency of 13 precipitation at the Yucca mountain site over the next

(} 14 15 1 000 years. The progress to date is on the next slide.

(Slide.)

16 MR. FEHRINGER: The progress to date has been to 17 develop some of the training materials and some of the 18 elicitation materials that will be used to carry out this 19 study. An initial effort has been made to identify a panel 20 of experts through requests to professional societies for 21 nominations.

22 When those nominations are received, people will 23 be contacted and asked if they're willing to serve. The 24 elicitations will then be carried out and we're go through 25 this example. We expect to complete this within about six O ^"" ai'ev a Associ^Tes' 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

i 143 months from now, q 1 b' 2 When you got a lot of people involved, it's hard 3 to guarantee you're going to stick to schedules. So this 4 schedulo should be taken with a grain of salt. This is our 5 current planning to have this trial olicitation completed in 6 about six months. i 7 At that time we'll learn whatever lessons ko can 8 learn froE it and ovaluate whethor it's appropriate to try 9 to develor %? "aiv' 3r regulation that more specifically 10 addressos 'w ei .ii;Jxt , Jdgoment in our licensing procons.

11 KR . < - .LnR: What other topics did you consider 12 and discard as being of lower priority than this?

13 MR. FEllRINGER I wasn't involved in that, our 14 contractor picked the topic. Is Aaron here? Aaron 15 DeWispelare from the center has taken a microphone to give 16 you that information.

17 MR. DeWISPEIARE: Maybe I should have either 18 Margaret or Bob comment on what the other topics wore that 19 were competing with climatology.

20 MS. FEDERLINE: Well, we had considered volcanism 21 as well. One of our considerations in doing this 22 olicitation was that we were trying to learn about the 23 olicitation process. We did not want to do an elicitation 24 that would preempt DOE in any_way.

25 We felt the volcanism topic was so controversial

_O ^"" ai'ev * ^8soci^Tes' d-  !

Court Reporters ,

1612 K. Street, N.W., Suite 300 i Washington, D. C. 20006 l (202) 293 3950 l

144 1 that if we got together a panel of exports and they somehow 2 in a definitive way made some conclusion about volcanism, wo 3 would perhaps be out in front of the applicant. We felt 4 that was inappropriato.

5 So, we tried to choose a topic that had )

6 uncertainties related to it, but we would not proompt DOE.

7 We're more looking at the process of elicitation and what we 8 can gain from it. So we tried to focus a specific topic 9 that would just provide us that information.

10 MR. POMEROY: Dan, are you going to carry this 11 through to the bottom line in terms of --

12 MR. FEllRIl1GER: The bottom line being a judgement 13 on what the frequency and intensity of precipitation will 14 be?

15 MR. POMEROY: Yes, sir.

16 MR. FE!!RIllGER: Yes, that's our intent.

17 MR. POMEROY: The only other question that I want 18 to ask right now, Dan, is, in all of your oarlier slidos in 19 the first part of your presentation, you talked about 20 judgement. Phase 2.5 is a formal olicitation of export 21 judgement. It's partially a result of the recommendation of 22 this panel that you're doing this.

23 But in terms of using judgement, there may be a 24 formal elicitation of expert judgement in many areas. I 25 think that will be fairly well documented and least O ^"" ai'ev a Assoc'ATes' 'id-Court Rephrters 1612 K. Street, F.W., Suite 300 Washington, D. C. 20006 (202) D3 3950

l 145 1 debatable in a rational way. There may well be in the 2 license application, a tremendous amount of implicit 3 judgement.

4 I'm not sure I even know how you would try to 5 approach and identify where all those areas are under this 6 process that you've described.

7 MR. PEllRINGER: I don't know how we would approach 8 that either. That's one of the difficulties we have in 9 trying to conceive of how we could develop guidance or 10 regulations that really would be useful. There's a whole il spectrum of ways to produce judgement -- some very formal 12 and highly structured, and some just -- somebody makes a 13 judgement and goes forward with it.

() 14 15 It's really difficult to figure out a set of rules that would be applicable to the full range. That's a real 16 problem we dealing with.

17 MR. POMERoY: I think it isn?t as hard to figure 18 out rules for the formal. I think I agree with you there, 19 and the rules for that, it seems to me after we get through 20 this experiment, you "ight well want to consider going .

21 forward with rule-making on that subject.

22 I think you were next?

23 MR. APTED: Just a question on the expert panel, 24 will you have normative members in addition to these that 25 are reputation members, individuals that are experts in sort O Ann aitev a Associates. tia.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

146 1 of over-all performance?

2 MR. FEHRINGER: Yeah, they've hired a normative 3 expert who will help prepare the preparation materials, the 4 background information, and the training materials. Then he 5 will conduct the clicitations.

6 MR. DeWISpELARE: If I could just add, we're also 7 going to have what we call area specialists involved in the 8 training and the e11 citation process -- folks experienced 9 with the process, if you will, as well as to make sure that 10 their questions are properly being interpreted and so forth.

Il MR. DAVIS: I have sort of a strange question, but 12 how do you establish the expertise of experts if a problem 13 has never been done before? For example, if you equate it

() 14 15 to a court of law, it's basically, "I've studied this.

have experience with this.

I My experience says this is the 16 likely outcome."

17 None of us have done this problem before. I know 18 we all think we're very smart, but none of us have ever done 19 this problem. I would say that even if you're going to give 20 me the answer, that you break it down into the expertise of 21 the people. They haven't done that problem before, either.

22 MR. FEHRINGER: That's one of the areas where the 23 NRC Licensing Doard seems to be somewhat more flexible than 24 the Federal courts. I am told anecdotally that the Board's 25 are fairly accommodating in terms of allowing people to O ^wn nitev a Associates. 'io.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 )

(202) 293 3950

147 1 present information.

2 In fact, one case, a high school science teacher 3 was offered as an expert on a particular subject. He didn't 4 have the normal evidence of expertise like a Ph.D. or 5 research history in a subject, but he said he had read a lot 6 about it. That combined with his training in science, he 7 nald, was enough to allow him to offer useful information to 8 the Board. The Board did allow him to testify.

9 So I don't think there's going to be a real 10 problem with allowing evidence. There may be more of a 11 problem with getting more than the Board can really deal 12 with, in particular some rather speculative information 13 presented.

() 14 15 MR. DAVIS:

MR. POMEROY:

It will all be speculative.

Just to follow that up, of course, 16 there's some experience in backgtvund in the seismic hazard 17 evaluation areas, of which both Mr. Hinze and myself have 18 had a lot of experience. Both of us probably qualify as 19 experts and probably would qualify as experts in that area.

20 Yet if somebody asked us the simple question of 21 what is the basic mechanism that is causing the Eastern 22 United States seismicity, none of us would be able to give 23 that answer, or agree.

24 MR. FEHRINGER: Or agree.

25 MR. POMEROY: Or agree.

I O ^"" ai'ev $ ^ssoc'^Tes' 'id-Coud Reporters 1612 K. Street, N.W., Suite 300 kVashington, D. C. 20006 (202) 293-3950

148 g 1 (laughter. ] l 2 MR. POMEROY: Therefore, if you don't understand 3 the basic positive mechanism offering your opinion, which we 4 both did, may be a very tricky undertaking.

5 MS. FEDERLINE: We often raise these issues.

6 Occasionally it's been pointed out to us that in licensing 7 reactors early in the licensing process there were many 8 problems that had not been definitely answered -- severe 9 accident questions and what not. Yet, the Boards were able 10 to make licensing decisions. So, we are raising the same 11 kind of questions that you're raising. We're continuing to 12 explore it.

13 MR. POMEROY: Okay.

!() 14 15 MS. FEDERLINE: I just wanted to note to Dr.

Steindler that I have a copy here -- I'll make available to 16 Georgio -- this is a presentation from Michael McGary of 17 Winston and Strawn -- industry experience with expert 18 judgement. I think it has in it some of the things that Dan 19 was alluding to earlier, so we'll leave a copy of that to 20 the members.

21 MR. POMEROY: Thank you, Margaret. We may have a 22 copy of that because we had somebody at the meeting also.

23 MS. FEDERLINE: Oh, okay.

24 MR. POMEROY: But check with Georgio to make sure l 25 we do.

l O ^"" ai'ev ^ ^ssoci^Tes' 'ta-COud RepOders 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l

149 1 Thank you, Dan.

2 (Slide.]

3 MS. FEDERLINE: What we need to do is lay out a 4 specific plan of technical issues that need to be addressed 5 by performance assessment. Again, this list is by means 6 inclusive, but I think anyone who sat through the two days 7 of the technical exchange would recognize many of these 8 issues as the key issues that we touched up.

9 What we hope to do in our strategic plan is to lay 10 out how our performance assessment will address these issues 11 prior to the licensing process and in a fashion to provide 12 timely enough guidance to doe so that they can take 13 advantage of this.

14 How, I'm sure you realize what a difficult task 15 that's going to be. We wold plan to iterate the strategic 16 plan, each iteration of the performance assessment. So, we 17 would try to make a projection of what needs to be done, lay 18 out a plan to do it, and then go back and look at it, based 19 on results that we get from the next phase of our 20 performance assessment.

21 MR. IIINZE: Margaret, is that the validation 22 defined by Norm earlier this morning?

23 MS. FEDERLINE: Yes.

24 MR. HINZE: How about verification? Is this also 25 in there?

O ^"" ai'ev ^ Assoc'ATes. d-Court Reporters 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

150 1 MS. FEDERLINE: Yes. As I say, this list is not 2 exhaustive. A lot of it is going to depend upon priorities 3 and how much we can get done.

4 (Slide.) l 5 MS. FEDERLINE: The strategic planning will be a 6 major focus of our fiscal year '93 activities in addition to 7 the other activities that I've listed here. We're all very 8 eager to complete Phase 2 of IPA so that we can gain the  !

9 insights.

10 We want to complete our review of DOE Sandia TSPA.

11 As I mentioned this morning, we're only budgeted for one 12 TSPA review. So we want to try to get some feedback from 13 DOE as to what would be the most useful way for us to 14 comment on the work that they're doing at independent 15 contractors.

16 Dan has just discussed Phase 2.5. I talked a bit 17 about the strategic plan.

18 [ Slide.)

19 MS. FEDERLINE: Another major thing that's very 20 important to us is supporting NRC's interactions with the 21 National Academy of Science and EPA in the development of 22 the high level waste standard. This will receive a high 23 priority during the coming fiscal year as the scope of.the 24 study is defined, and the National Academy gets underway.

25 [ Slide.)

l-O l

^"" ni'ev a Associates. 'ta-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l

151 g- 1 MS. FEDERLINE: We'd also like to work on our plan k 2 for Phase 3. Another very important aspe A. of our 3 regulatory program is the systematic regulatory analysis 4 which will be process by which we'll define our compliance 5 determination methods.

6 (Slide.)

7 MS. FEDERLINE: Let me talk for a few minutes 8 about some of the considerations that we will need to 9 address in developing IPA, Phase 3.

10 I've listed them here, and I'll go into a little 11 more detail on some of them. We need to prioritize the 12 additions and refinements of existing models and codos 13 identified under Phano 2. I think that question came up

() 14 15 this morning as to how we decido how far we're going to take one code or another.

16 (Slide.)

17 MS. FEDERLINE: Some of our considerations for 18 doing that are insights that we gained from DOE's TSPAs, the 19 need for detailed review in specific areas, using NRC models 20 and codes, or whether we can look at DOE codes and examine 21 their assumptions.

22 Another consideration will be the scope of the 23 National Academy study here regarding the EPA standard. As 24 everybody points out, we will got some indication from the.

25 scope of the study, at least the range of issues that are O ^w" ni'ev a Associates. 'ta.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D, C. 20006 (202) 293 3950 l

. . . ~ . . ~

l 152

~

1 going to be considered.

2 Dr. okrent mantioned that it's not limited to the 3 several issues that were identified in the Energy Policy 4 Act. We want to make sure that Phase 3, at least, 5 accommodates the full scope of the study.

6 (Slide.)

7 MS. FEDERLINE: Another consideration for planning 8 of Phase 3 is the technical uncertainties identified through 9 the systematic regulatory analysis. I'll comment in a 10 little more detail on how that process works.

11 (Slide.]

12 MS. FEDERLINE: I'm not going to spend much time 13 on this because Norm has gone through in a lot of detail of

() 14 15 what our considerations from Phase 2 were and things that we need to consider in Phase 3. You can sort of scan down 16 through this in the next slide in your packet. These are 17 just examples of the issues that came out of Phase 2 that we 18 will want to examine for Phase 3.

19 [ Slide.)

20 MS. FEDERLINE: Let me just put the next one up 21 quickly. I think Norn has mentioned all of these in his 22 lessons learned.

l 23 MR. HINZE: This is so far?

24 MS. FEDERLINE: Yes, so far. I'm sure there will 25 be a much more exh 44 ting list.

O ^"" nev a Associ^Tes' 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 i

Washington, D. C, 20006 (202) 293-3950

153 1 (Slide.)

0' 2 MS. FEDERLINE: Some of the considerations for IPA 3 from doe's TSPA are listed on this slide. There's a certain 4 amount of inner-comparison that we feel that we need to do, 5 and particularly in the areas of gas flow, source term, l

6 unsaturated mode 3s. As came out of our discussions over the I 7 last few days, we fool that further work is neoded on CCDF 8 construction and combination, as well as comparing the DOE 9 and NRC scenario methodologies.

10 Paul?

11 MR. DAVIS: Isn't the treatment of matrix 12 diffusion on there? You treat them co very different.

13 MS. FEDERLINE: Well, these are just examples.

(} 14 15 It's not a definitive list. We have not completed the TSPA review, but I just wanted to give you an idea of the thought 16 process that we were going through to try and focus our 17 priorities.

18 Up to this point we may have give you the idea 19 that what comes out of Phase 2 dictates what goes into Phase 20 3, but there are a lot of competing influences which lead us 21 to make priority decisions for what belongs in Phase III, 22 (Slide.]

23 MS. FEDERLINE: We've discussed this as a result 24 of the Energy Policy Act of '92. EPA has been directed to 25 contract with the National Academy to provide findings and O ^"" ai'ev a Associ^Tes' 'id-Coud Repoders l 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

154 g- 1 recommendations. As Dr. Okrent pointed out, the study is

( 2 really not limited to a given set of issues.

3 But we've tried to do some early thinking here of 4 potential impacts that we could see coming from a National 5 Academy study -- ways that performance assessment activities 6 might be changed or enhancements that might not be needed.

7 Although we'll have to wait for the study to see 8 what the scope and eventually the findings and conclusions 9 and how EPA uses those findings and conclusions, we really 10 feel that our PA activities need to be sensitive to thc full 11 scope of the kinds of things that the National Academy is 12 going to be considering.

13 (Slide.)

() 14 15 area.

MS. FEDERLINE: How this is another important This is one of the most important areas in our 16 regulatory program -- considerations for how IPA will 17 contribute to the development of the license application la review plan.

19 As I say on this slide, the LARP provides guidance 20 to the NRC staff on how to review DOE's license application 21 and how to determine if DOE has demonstrated compliance with 22 10 CFR Part 60.

23 We started development of this in 1992 and 24 systematic regulatory analysis will be the process we use to 25 get to our. license application review plan. Our plan is to l

O ^"" ai'ev a Associates' 'ta-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

155 1 issue T draf t LARP in 1993. We will be using the draft LARP 2 for pro-licensing reviews. We hope that it will ensure 3 uniformity in the staff review and it also will be a way of 4 conveying to DOE and other parties the NRC approach to 5 licenso reviews.

6 As I say in my first bullet, wo are very hopeful 7 that ps armance assessment will be used to help focus the 8 review plans on technical areas that are most important to 9 performance. As I note here, 102 review plans will be 10 developed which cortospond to the Part 60 regulatory 11 requirement topics end the associated format and content 12 guido.

13 Initial efforts have been directed toward 14 devoicping strategies for compliance determination for 15 regulatory topics. The plan is to complete the development 16 of strategies and initial identification of uncertainties by 17 the end of 1994.

18 MR. PoMERoY: Before you leave that, is the verb 19 in the first bullet "is being used" or "will be used"? You 20 said "will be."

21 MS. FEDERLINE: Right. Ir. Fiscal Year '92 we 22 completed only a limited number of compliance determination 23 strategies. I think it was on the order of about 10. So, 24 the insights from the analysts' experience in conducting 25 performance assessment woro used as they went through the 10 0 ^"" ai'ev a Associates, 'id-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l 156 1 CFR Part 60 and identified technical uncertainties.

2 We hope that as Phase 2 is complete and we have 3 results, we'll be completing the rest of the 102 review 4 plans, and it will be a much more extensive contribution.

5 (Slide.)

6 MS. FEDERLINE: Here are just somc examples. I'll 7 just touch on those because of the time. For example, the 8 regulatory requirement in the area of potentially adverso 9 conditions. Some technical uncertainties that have been 10 identified are the ability to predict the likelihood of 11 earthquake occurrence, and the adverse effects of future 12 ground water withdrawals on the ground water system.

13 As part of our process, we would hope that our

()

v 14 performance evaluation would shed some light on the impact 15 of these uncertains on performance and how much weight they 16 should ue given in the license review.

17 So, in summary, Phase 2 activities are coming to a 18 close. We hope to issue a report in the spring of '93.

19 We're beginning a focused planning process. We want to 20 thank you for your recommendation that. We nope that will 21 result in a long-term performance assessment strategic plan, 22 and a plan for Phase 3, 23 The planning will use considerations from Phase 2 24 as well as considerations from our review of TSPAs, the 25 impacts of the National Academy study, as well as O ^"" ai'ev a Associates' 'id.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l

1 l

1 I

157

,f-- 1 consideration of uncertainties identified through systematic 2 regulatory analysis.

3 MR. MoELLER: To me, strategic plan is a very 4 difficult document to prepare. As I listened, Margaret, to 5 your presentation, you were very clear that you'll be using 6 the results of the IPA different phases to guido you in the 7 regulatory process and what you need to know and so forth.

8 But in the early portion of it you enumerated --

9 and I know they were just examples -- but you enumerated 10 what you considered to be the pr_.tities issues to address 11 in the next phase of the IPA.

12 It seems to me that an important component -- and 13 I could be wrong -- but an important component of a

() 14 15 strategic plan would be the procedures and methodology you use in establishing whatever it is that's to be the problem.

16 MS. FEDERLINE: Exactly; exactly.

17 MR. MoELLER: okay.

18 MS. FEDERLINE: Those were only meant to be 19 examples. We've not gone through that criteria-setting 20 process.

21 MR. STEINDLER: How many "I's" are there in IPA?

22 How many phases do you expect to have to go through?

23 MS, FEDERLINE: Well, I think our strategic plan 24 will dictate that. We're_ unsure. We've been doing sort of 25 an 18-month to 24 month cycle. Some people have proposed Q ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C, 20006 (202) 293-3950 i

158 r3 1 that maybe we should take smaller bites and maybe look at U 2 smaller parts of a total system analysis and maybe go to a 3 longer 1guency or even a shorter frequency for a total 4 systems number.

5 But what we really have to do is identify theso 6 technical issues and see what the performance dimension is 7 to get us to the answers to those questions. That will sort 8 of dictate how many iterations.

9 I think we're going to all be old, dead, and tired 10 by then -- in that order.

11 MR. HINZE Margaret, does the science ever drive 12 the IPA? In other words, some group within your 13 organization comes up with a problem of whether we should

(} 14 15 know a specific data characteristic to a greater degree of precision or accuracy?

16 Is the performance assessment set up to carry out 17 the sensitivity study on an ad hoc basis to help answer that 18 question? Are you so far ahead of the game that these 19 questions are in mind and solved?

20 MS. FEDERLINE: I think prior to Phase 2 we were 21 not set up, but we did not have a sensitivity and 22 uncertainty analysis. Norm, if you would like to answer, I 23 believe we will be more able to perform those kind of 24 limited analyses or sectioned analyses to answer specific 25 questions once Phase 2'is completed.

O ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 -

159 4

1 MR. HINZE: But oftentimes they interact so that 2 you really have to perform the entire thing to have_any 3 certainty, I would suspect?

I 4 MS. FEDERLINE: Well, that's true.

MR. HINZE I'm wondering whether it is a matter 5 i 6 of funds, access to computers, et cetera? Is it financial 7 resources aspects that might restrain that kind of use of 8 the IPA? Is there any constraint?

9 MS. FEDERLINE: Well, I think that_ issue will 10 really come when we begin to develop our compliance 11 determination methods. Those are the methods that will 2

12 specify what will be an acceptable demonstration of 13 compliance with the regulatory requirement. I think during 14 that time when we're coming up with those methods, there.

(}

15 will be a lot of call on performance assessment.

16 We've been in the building staff' capability. arena 17 to this point. We've been limited by computer capability, 18 limited by staff capability.- So, it's not been a resource  :

19 that people have felt that they could call on,-I think, to 20 this point. Correct me if-I'm wrong.

21- MR. EISENBERG: If I_could_just add another 22~ wrinkle to what I think-you're asking is that remember that 23 for.the total system performance,_we are,;to use DOE's 24- terminology, at a highly abstractive-level where the 25 discretization and the scale may not be'able to reflect the T

ANN RlLEY & ASSOCIATES, Ltd.

'~h ~

Court Reporters _ .e~

1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 ,

. - - .u_.-.-.. _ _ . . _ _ . a .a a _, _ _ , , _ _ , - ,. -_ a. . _. .,

l 160 1 kind of precision that you may be concerned with, s

2 So, we may have to go to what we calling auxiliary 3 analyses to determine how to abstract from them -- or the 4 implications that those have on getting the parameters that 5 we put into the total system codes.

6 I think that's part of what you're asking is to 7 what precision do we need to know a certain parameter? How 8 representative is it? How many bore holes do we have to put 9 in?

10 MR. HINZE: Right. Can I set at my workstation 11 and feed in the performance assessment codes to look at the 12 variables and the sensitivities and the robustness and all 13 those things?

14 MR. EISENBERG:

[Vl You can look at some of the 15 variables, but I think it unrealistic to think that the 16 kinds of variables that we put into the codes are the kinds 17 of things you measure in the field. The kinds of things you 18 measure in the field are often then -- let me-put it the 19 other way.

20 The kinds of things you put into the codes are 21 often inferred from the field measurements through long 22 arduous complicated types of calculations which are unlikely 23 to put automated to an extent that you can change some idea 24 of what's coming in as a field measurement and have it 25 immediately reflected in the performance estimate.

h ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W.', Suite 300 Washington, D. C. 20006

.(202) 293-3950 1

161 1 So, I think for some of the parameters that are 2 important to the abstracted codes and models that, yes, you 3 will be able to sit at your workstation and say, "Well, if 4 we double the numbers of canisters, what effect will it have 5 on performance.

6 You may not be able to do something like change 7 the chromium content of the matorial that you're building 8 the wasto package out of because wo may have an abstracted 9 model which is a fit to a very complicated code. You've 10 have to go through the entire more complicated level of 11 modeling in order to got a different paramotor to put into 12 your system code.

13 MR. HINZE: If a person wants to do that, then

() 14 15 what would they do?

this as a run?

Come to the Systems Branch and proposo 16 MR. EISENBERG: Possibly, but I think the way 17 we're set up, the various disciplines would have available 18 to them a lot of these more detailed codes. They would be 19 familiar with them.

20 MS. FEDERLINE: Performance assessment is a way of 21 thinking rather than a group of people in our location.

22 MR. POMERoY: Before I let you go away, Margaret, 23 I would like to ask my question now. I hasten to assure my 24 colleagues that this is not asked out of either ignorance, 25 nor is it a simplistic question.

O ^"" ni'ev a Associ4Tes 'ta.

Court Reporters 1612 K. Street, N W., Suite 300 Washington, D. C. 20006 (202) 293-3950

162 1 The question is simply, in the context of this i 2 program, from your perspective, what is a CCDF, and how does l 3 your definition -- what is your perception of the difference 4 between your definition and that of DOE, and do you think 5 that difference has a potential for creating any problem at 6 a later point in time?

7 It is a three-part question, and you have five 8 minutes to answer.

9 MR. EISENBERG: I will see what I can do. In the 10 context of the EPA standard, we are calculating cumulative 11 releases over 10,000 years. If this were a nico simple 12 system, we would have one value, and there would be no CCDP, 13 or it would be very simple. We have variability in the

() 14 15 cumulative release, I would say, firstly, because we are unsure of the evolution of the repository, and the 16 environment in which it works. Therefore, we have 17 alternative future conditions which lead to different 18 releases, and because we have the potential for different 19 releases, we can assign a likelihood to those different 20 releases that is based on the likelihood of the conditions 21 which produce them, whether that is earthquakes, climate 22 change, any of these natural eve.nts that could occur singly 23 or in combination.

24 So I think that is one set of uncertainty, if you 25 will, that produces different releases, and when we look at O ^"" ai'av a ^ssoci^Tes. 'id-Coud Repoders 1612 K. Street, N.W., Suite 300

\Yashington, D. C. 20006 (202) 293 3950

163

-~ 1 a CCDP, we are looking at the probablistic accumulation of 2 the different releases.

3 There is another factor, or another set of factors 4 which produces variability in the release from the 5 repository, and that is the variability in the parameters 6 that ooscribe_the system, and these are most notably 7 associated with the natural system, the hydrologic 8 parameters, the porosity, the permeabilities, the 9 geochemical properties. These are actually -- a more proper 10 term might be variability rather than uncertainty because 11 even with rather complete knowledge of the repository, you 12 would know and acknowledge that there is variability.

13 The way it has been traditional in many of the 14 related fields of contaminate transport to treat that

(}

15 variability is to use the variability of certain 16 measurements, the spread in certain measurements, and to 17 pick point values, and have that apply to a rather large 18 region of space, and see what affect that has on the 19 performance.

20 This is another type of variability that we 21 normally incorporate into our performance estimates, and 22 that contributes also to the CCDF, and, of course, with this 23 variability we have a certain amount of probability 24 associated with certain values or ranges of values, and by 25 doing the calculus properly, we can generate a CCDP.

( ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006

'(202) 293-3950

164 1 There are other sources of uncertainty. Another 2 source of uncertainty is the incomplete knowledge of the 3 physical processes, and how they interact with each other 4 and affect the predicted performance of the repository. We 5 have, that is the NRC Staff has, so far, in their approach 6 to construction CCDPs, not included the numerical 7 probabilities associated with different conceptual models 8 into the CCDF. So far, I think the DOE has, and that is a 9 difference, and that is an area that needs to be resolved as 10 we move towards licensing.

11 MR. POMER0Y: Jerry, would you care to comment on 12 that at all?

13 MR. BOAK: I guess I have tried to indicate that

() 14 15 our objective would be to present something that I think is directed by the EPA, a single CCDP, which attempts to 16 capture the variety of uncertainties that we have, and that 17 our expectation is, if we had a standard which required a 18 CCDF, that there would be a great deal of supporting 19 material that showed the effects of many of the kinds of 20 uncertainties we have on various conditional CCDFs that make 21 up part of that single total CCDF.

22 The ways in which we represent our uncertainty 23 include several things. We look at samples of population of 24 a multitude of variables, many, many variables, and we make 25 some estimate about our uncertainty about the population O ^"" nev a ^SSoci^Tes. 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

165 1 using standard statistical techniques. That is addressing 2 the question of variability. But also, we have attempted in 3 a number of ways to capture that component of ignorance that 4 we call uncertainty that is, in some ways, broader than the 5 variability within a system.

6 An alternative way besides expanding or changing 7 your distribution to meet that is simply to assign a single 8 value that we can call a bounding value, or some higher 9 distribution that would be satisfactorily conservative. In 10 a number of places, we think our current CCDP has many of 11 those conservatisms incorporated where we have essentially 12 assigned a much more restrictive distribution, but one which 13 is slanted in order to give us a higher estimate of release.

14 I think that our objective would be to present 15 some single piece of evidence that we said, here is our wrap 16 up of everything we think we know and don't know about this 17 site, and here is what wo expect in the way of performance, 18 and we certainly expect to see that covered by a great deal 19 of backup material to show sensitivities.

20 MR. POMEROY: Mick?

21 MR. APTED: Just a question, and I am not sure to 22 who here, but at the last Performance Assessment Advisory 23 Group Meeting in Paris, the WIPP Site presented some rathtc 24 strange CCDFs which had sort of a median or a base case, and 25 then they had 95 percentile, and so on. Is that the point O Ann ni'ev a Associates. 'ta.

Court Reporters 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 I (202) 293 3950

L 166 1 you ato getting at, that there are those published, rather 2 than cort of one CCDF that combines all the variability 3 uncertainty, they are sort of contouring CCDP space, or 4 something, and I don't know if you have soon that, and 1' 5 whether your comment is on that.

6 HR. EISEEBERG Yes. The so-called " Hair 7 diagrams," yes, and that was one of the things I mentioned 8 this morning that part of their approach was for each sample '_

9 from paramotor space to generate a CCDF representing the 10 variarility with respect to different scenarios. So they 11 conatructed the points in the curve from different 12 scenarios, one for each replication, one for each paramotor 13 vector, and then the whole spread for 100 or 1,000 vectore

() 14 15 would give you this hair-like diagram, and then they bound it with percontiles moving from one of the CCDFs to another.

16 Wo have had many, many discus ~nt 2 which are 17 continuing about the pros and cons of that particular la approach, and the pro,t and cons of tryir.g tc represent the 19 uncertainty about the mean CCDP by using those percontile 20 ranges.

21 What I believe, and it may or may not be correct, 22 but I believe that the WIPP people were not-representing the 23 uncertaip*v with regard to different conceptual models in 24 those diagrams, that that was something that was separate.

25 In fact, I believe they generated different Hair diagrams O ^"" ai'ev 5 ^ssoci^Tes. 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

I 167

~; 1 using-different conceptual models. That is the kind of 2 uncertainty which can be rolled up into the CCDF or cannot.

3 I guess, with regard to what Jerry said, I think, 4 yes, the EPA standard does require a single CCDF, but I 5 think the NRC Staff has, and probably vi continue to 6 encourage DOE to present not only the 2011 up, but also the 7 disaggregated curves, because that gives a lot more insight 8 into what the uncertainties are, what they are due to, and 9 how to evaluate the evidence that is presented.

10 MR. POMEROY: Felton?

11 MR. BING1!AM: Norn's original answer to your 12 question was so economically phrased that I am sure it was a 13 setup, but I think he covered the points that are important.

1 14 The one I would like to add to it is this, let's don't worry

}

15 about it right now. There are lots of ways to make CCDFs.

16 I think we are a little surprised to find that out because, 17 we thought, when we read the standards seven years ago, that 18 there was going to be one. We find that people are clever, 19 they have come up with other ways of presenting their 20 information, t

i 21 It may be that by the time a license application 22 comes around, we will all have agreed on.one curve that 23 tells us everything we want to know, but I think the chances 24 of that are about zilch because, as a part of getting 25 reasonable assurance, people are going to want to knou more O ^"" aev $ ^SSoc'ATES' d-Coud Repoders i 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l 168 1 than just one curve's worth, and I suspect that there will O 2 be a role for all of these ways of making a CCDF because 3 they all contribute some kind of information.

4 They weren't invented because somebody was trying 5 to come up with something that is different than what 6 someone else has done. They were invented because somebody 7 saw a reason for making Hair diagrams, or including, as the 8 DOE is doing now, a lot of uncertainty that is not quite the 9 kind of uncertainty the NRC Staff thinks they would prefer '

10 to see in them. I would say, let's keep the debate open.

11 Let's continue to use all of them for a_ good while, at 12 least, and find out which ones are good and which ones 13 aren't.

14 MR. POMEROY: I concur.

15 MR. BOAK: The EPA standard is a lot like the Tax 16 Code in that way.

17 MR. POMEROY: Right, and there may not be a CCDF 18 by the time we get around to the real problems.

19 MR. BOAK: But there will be an IRS.

20 MR. MOELIER: Speaking of the IRS reminded me of 21 my question which I ask with sincerity. I was reading a 22 book about baseball some time, and they' pointed out the 23 fact, I guess, is it 90 feet between bases, whatever it is, 24 makes it difficult to be a 300 hitter, and probably 25 impossible except for Ted Williams, to be a 400 hitter, in ANN RILEY & ASSOCIATES, Ltd.

Q-COud RepOders 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

169

~ 1 fact, though it makes it possible to have double plays, and 2 so forth.

3 I say that because it is interesting to me that 4 when I look at the CCDFs that you present, either NRC or 5 DOE, the curve is either just a little to the left of the 6 line, or it just barely cuts into the line, and is EPA.this e 7 brilliant?

8 If I didn't have EPA's standards, and I were doing -

9 a CCDF, would it fall exactly where it falls?

10 It really seems fortuitous to me that these curves 11 fall where they do.

12 MR. BINGHAM: Should I answer that?

13 MR. POMEROY: Yes, please.

14 MR. BINGHAM: I don't know whether fortuitous is 15 the right word, but I know from personal experience with the 16 curves I have worked with, and the people I have seen doing 17 them, there has been no jiggerypokery with those curves.

18 They have come out where they have come out.

19 Perheps it is surprising that more of them don't-20 fall over, but when you talk to the EPA people, and try to 21 get their feeling, rather than the things that are in their 22 documents, you'have the feeling that they set those things 23 pretty conservatively. They did try to look ahead to what 24 sites will really do, and although there have been 25 firestorms of criticism of-what they did, and exactly how Q ANN RlLEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

170 1 they did it, I have kind of come to the conclusion that they 2 really did pretty well. The probably did set a standard 3 that is quite a bit higher, or somewhere near what a 4 reasonable site will be expected to do, and they say that 5 they don't think there should be wide margins of safety 6 between their limits and what a CCDP turns out to be because '

7 they set the standard that way.

8 So perhaps it is not surprising that they come out -

9 that way, and there are a few of the partial CCDFs, of 10 course, that are far away from them.

11 MR. DAVIS: Without a definition of what you 12 expect the CCDP to represent and a way to come up with it, I 13 don't see how any of us expect to use performance assessment 14 to drive site characterization or research at all. I have

{)

15 no idea how you do that.

16 MS. FEDERLINE: I think that is why we proposed it 17 as a topic that we need immediate or, if not immediate, 18 near-term discussion on. I don't think we are --

19 MR. EISENBERG: As a first cut, I think we are 20

~

talking about what kind of uncertainties to include in it, 21 and how to roll it up. The fundamental is still there. We 22 are still looking at cumulative releases from the 23 repository. The parameters and the variables that will 24 affect that most are still the things that we have to focus 25 on in site characterization.

O ^"" ni'av a Associates. 'ta.

Coud Reponers 1612 K. Street, N.W., Suita 300 Washington, D. C. 20006 (202) 293-3950

171 1 Of course, also the things that will cause

'~

2 disruptive events that could influence the performance are 3 the things that need to be focused on.

4 MR. DAVIS: But how do we define the stopping 5 rule, I don't know how to do that. How do we do what Mick 6 would like to do, and that is throw things out from 7 consideration because they are of no concern to meeting the 8 standard?

9 I can't do that unless I have guidance that says, 10 this is the source of uncertainty that will be included in 11 the CCDF, and this is how you construct it. If I have 12 confidence in that, then I can say climate change doesn't 13 matter because I show it doesn't matter to the CCDP, it 14 doesn't matter to safety, 15 If I have multiple approaches to doing this, one 16 which lies on one side of the rule, one on the other, I 17 can't throw climate change out.

18 MR. EISENDERG: I would again say that one of the 19 things that we don't like to include are the different 20 conceptual models, just because it is not included in the 21 CCDF does not mean it doesn't have an influence on what we 22 think is important.

23 If there are alternative conceptual models that 24 will shift that curve, then, as far as we are concerned, 25 that is an issue that needs to be explored. If DOE puts it l l

O- ^"" ai'ev a ^ssoci^Tes' 'ta-Court Reporters j 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 l (202) 293-3950 l

172 1 in, they would put it in as a uncertainty in the conceptual O 2 model, and -it will move the curve over.

3 So I am not sure it matters a whole lot whether 4 you do it outside the CCDF, or do it within the CCDF. I 5 think the effect is going to be the same. Maybe Felton can 6 convince you.

7 MR. POMEROY: Let me cut it off there because I ,

8 think we have to get on to the DOE presentation.

9 Margaret, thank you very much for all of the 10 presentations, and for the presentations in the face of the 11 somewhat overwhelming obstacles here of no PA system, and so 12 forth. We deeply appreciate it. Many of our questions, of 13 course, may seem critical, but they have an underlying basis 14 of admiration for the great deal that you have accomplished.

[}

15 MS. FEDERLINE: We always go away and think about 16 your comments. We appreciate the interaction.

17 MR. POMEROY: Thank you very much.

18 We are going to take a twelve-minute break, and we 19 are going to start at 3:00, 20 (Brief recess.]

21 MR. POMEROY: Back on the record.

22 I would like to turn now to a discussion of the 23 DOE Total Systems Performance Assessment. WE are fortunate 24 to have with us this afternoon several representatives from-

, 25 DOE and its contractors, Jerry Boak from the Yucca Mountain

~h ANN RlLEY & ASSOCIATES, Ltd.

Coud RepOders l 1612 K. Street, N.W., Suite 300 l Washington, D. C. 20006 i (202) 293-3950 l

173 1 Project Office; Holly Dockery-and Felton Bingham from Sandia

-D -

2- National 1 Labs; and! Paul Eslinger from1 Pacific = Northwest 3 Labs. _I believe Holly is going to_ start this afternoon.

'4 Welcome, and thank you for being patient with us

-5 today.

6 MS. DOCKERY: Does-anybody else have_a sense of 7 deja vu, didn't we just do this yesterday. You guys can all 8 go back and go to sleep now. -

9 What the title of the talk says was something r 10 about overview from PACE to TSP-91, but it is really going 11 to be more of an overview of the TSPA-91 with a few 12 reference to PACE, because PACE was really the_last 13 benchmark,'if you will, from which we built this particular 14 TSPA,-perhaps it was the last published document that you

}

15 saw that-involved an performance assessment calculations.

16 ,

(Slide.)

~

17 MS. DOCKERY: -Most of you have seen this 18 yesterday, but the purpose of this?particular_TSPA,qTSP-91, 19 for DOE - -and_I will try to be clear when I am talking. -

-20 about Sandia's TSPA document-versus'the-DOE TSPA effort, 21 which included both Pacific Northwest Iabs and Sandia -- ,

-22 was really to.try to-start:a series of iterative-performance 23 assessment.- Something_that we1could continuously _buildJon; 24 and-learn from our previous inadequacies and mistakes,-andi 25 do better with-as time went on.

ANN RILEY &L ASSOCIATES, Ltd.

Coud RepOders 1612 K. Street,- N.W., Suite- 300-Washington, D. C. 20006 i (202) 293-3950 -

174-1 We wanted to try to review the system under 2 conditions that we thought were reasonable. I know that we 3 had some argument yesterday about whether the right 4 distributions were picked, and we don't want to go to great 5 lengths to defend any one distribution, because this is not 6 our licensing argument, this is simply ranges of parameters 7 that we think we will probably be working in when it comes 8 time to actually start doing something real.

9 We also wanted to have some tie-in to guiding site 10 characterization and, indeed, this particular document was 11 one of the bases upon which the Integrated Test Evaluation 12 Task Force based some of the information that they used in 13 their assessments to determine which tests were most 14 important, or most likely to give us valuable information

(

15 for our site characterization.

16 We wanted to have the teams that would be involved 17 in total system performance assessment working with the {

18 steps that were outlined in the SEP, the six-step method [

19 that Raleigh has talked about in the past in great detail, ,

20 and I think Felton may be talking about it just a little bit 21 in his talk.

22 From Sandia's point of view, what we wanted to'try i

23 to do was generate abstracted models. We hadn't really had 24 a concerted effort to try to do this, and we wanted to find 25 out, how difficult is this process, is this appropriate, _can )

l Q ANN RlLEY & ASSOCIATES, Ltd.

Coud - Reponers 1

1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

175 1 we capture the sorts of processes or parameters that we need 2 in models that could be run for Monte Carlo simulations as 3 opposed to having to run full up 2D and 3D models through 4 thousands and thousands of simulations. We really didn't 5 thin} that anybody wanted to wait for the license until the 6 craze 9-; through in about 10,000 years or so. So we feel 4

7 like this is something that is fairly important.

8 on my slide, there is still a word left from back -

9 when that we don't believe any more and that is 10 " simplified," we really do believe that these models are 11 abstract, that they have been condensed and distilled to 12 capture the essence of the processes we are trying to model.

13 In some cases, they were not simple at all. They

( f 14 were fairly complex, and certainly the method to get to 15 these models was very complex.

1 16 The participants that took part in this exercise 17 included Sandia National Labs. We did coordinate the 18 initial stages of the TSPA and did the calculations using 19 the abstract models. Pacific Northwest Lab also performed

• 20 calculations, and they were using 2D calculations, so they 21 were trying-to be-our touchstone back to, if we used 22 abstracted models, were we still getting the same range of 23 answers that we would get with the more detailed 24 calculations.

25 They also took the releases generated from both-tO ^"" ai'ev & Associates 'ta.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006

, (202) 293-3950

l 176 j=s 1 the ID simulations at Sandia, and the 2D simulations from 2 PNL, and put them into the dose calculations, and that 3 particular information is reported in the PNL report that 4 they have available. I guess Paul has them out in the mail.

5 They will be there for your after Christmas mailings from 6 the IRS'and from PNL will come about the same time.

7 MR. POMEROY: We have one from PNL already.

8. MS. DOCKERY: Are you lucky or what, u 9 MR. POMEROY: We are just waiting for IRS'.

10 MS. DOCKERY: We don't have ours yet.

11 Los Alamos Lab did provide quite a bit of 12 information to us in the form of data for the volcanic 13 analysis that we did, not only the parameters but also some

() 14 15 of the conceptualizations. Greg Valentine and Bruce Crowe, as you probably all know, have been involved in that for a 16 long time, so we will defer to the expert on their 17 information. "

18 Also Aaron Meyer who was at that time at Los 19 Alamos, provided us information on the parameter 20 distributions for retardation.

21 Lawrence Livermoore Lab specified the. source term 22 for the Sandia exercise, and gave us quite a bit of new 23 information that we didn't have for our previous exercise.

24 [ Slide.] i 25 MS. DOCKERY: You can see_the schedule that goes O ^"" ai'ev & ^ssoci^Tes. 'ta.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

d 177 1 from hydrocoin through PACE, ESSE, and the early site 2 suitability evaluation, and into the TSPA. Although we did 3 have a very concerted effort during this TSPA, much of the 4 information that was developed for the site suitability 5 evaluation study at the time we were not able to actually 6 put it into the.t document, that massive document that Jean 7 Yonker and her team actually published. They could only 8 deal with referenceable material, so it was only alluded to 9 indirectly, but we had done a number of analyses, and those l

10 analyses were the departure point, or the building point of 11 where we went for the TSPA, but PACE was also a big building 12 point, 13 Before I go into a lot of detail on the previous

( 14 exercises, I just wanted to briefly remind you of what PACE 15 did, what it was about, and if you have more detailed 16 questions, maybe I can answer them, or Raleigh, or Mick 17 Apted, or some of the other folks, Paul Eslinger, that were 18 involved in the PACE exercises.

19 There were 18 layers involved in PACE. This was 20 one of the big features of the PACE exercise, it was an 21 attempt to more precisely specify the lithologic layering 22 that would be present at the Yucca Mountain Site. It was an 23 attempt to try to determine whether detail of that degree 24 was going to be necessary in our simulations in the future.

25 The geohydrologic parameters that were assigned to l

O ^"" aev $ ^SSoci^Tes. 'id-Court Reporters 1612 K. Street, N W., Suite 300 Washington, D. C. 20006 (202) 293-3950-

1 178 e~ 1 each layer were single value. They did come from USGS down

\' 2 hole data, but they were extrapolated from different points 3 into individual layers that had been determined for PACE.

4 The calculations were run from the repository to 5 the watertable, P.nd the calculations were run in 1, 2 and 6 3D. TOSPAC did some 1D runs, SUMO from PNL, TRACER from Los 7 Alamos, NoRIA from Sandia all did 2D calculations, and then 8 the Los Alamos folks with TRACER did a couple of 3D 9 simulations just to look at the impacts.

10 There were very low infiltration rates that were 11 used, and these were single points, but there were three 12 points -- .01, .1 and .5 millimeters per year were used.

13 It turns out that that exercise showed that .5 was

() 14 15 about the point at which it got transitioned into fracture flow. Given the models that we were using, given the 16 stratigraphy, and the way that we set up the problem, below 17 that we were getting essentially all matrix flow. As a 18 result, we were getting on the order of ten to the fourth to l

19 ten to the sixth years of travel time from the repository 20 down to the watertable.

21 The source term that was used, only used for l

l l 22 radionuclides -- the neptunium, cesium and technetium -- and 23 that was trying to capture a range of responses by the 24 different radionuclides, solubility limit, alteration limit, 25 different ways of representing the entire source term just l

O ^"" aev * ^Ssoci^Tes. 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006

(202) 293-3950 l

179 p) 1 V

1 using thoca four radionuclides.

2 What we did when we moved on to the Total System 3 Performance Assessment '91 was model a number-of different 4 phenomena as opposed to simply the groundwater flow from the 5 repository to the watertable, we expanded the suite of 6 phenomena.

7 We did have releases to the accessible environment 8 along two paths, both through the groundwater path and to -

9 the surface. We performed stochastic simulations on a 10 number of different parameters, geohydrologic, some of them 11 were the volcanic parameters. There were a few sensitivity 12 studies done. This was an area that we really wanted to 13 expand. We would have liked to have don e more, but we

[J L

) 14 15 didn't have the time to complete all the sensitivity studies we would have liked, but there were a few sensitivity 16 studies run for volcanism and for the human intrusion. ,

17 This was the first time that we used the results 18 in some sort of a dose calculation, so we were beginning to 19 hear the whispers in the wind that we were going to have to 20 think about dose in the future.

21 Going back to the source term for a minute, the 22 source term was quite a bit more sophisticated. Livermoore 23 did give us information that we were able to use in the 24 simulations. We had a better understanding of the water 25 contact modes. It was computationally somewhat similar, but

(' ANN RlLEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 i

_.--________.,___m___ _ _ - _ _ _ _ _ _ _ _ _ .._ _ . _ _ _ . _ . . _ . _ _ _ _ . _ _ _ _ _ _ _ _ _ . _ _ . _ _ _ _

180' 1 we also had a-larger suite of radionuclides. In terms'of 01- 2 mobilization, there were somewhat better mobilization 1

3 mechanisms, and there were.some near-field interactions that 4 were incorporated.

5 The suite of radionuclides added plutonium,:

i 6 americium and uranium isotopes-to-try to deal with the 7 larger percentage of nuclides that were available in the 8 inventory. We put in Selenium 10-for the affects on dose, 9 and, of course, Carbon 14 because we wanted to try to-deal--

10 with gaseous release.

11 This is the suite of processes that were modelled 12 in TSPA. It included, as you can-see, groundwater flow.

13 So PACE was that, and the-total system performance

() 14 15 assessment -- so we expanded the groundwater flow to have the composite porosity, and we also got a annar simplistic 16 model: to represent fractures, and in the saturated zone 17 transport, gaseous flow via surface release, human intrusion 18 in both the surface release and a direct saturated zone 19 release, basaltic volcanism as a. surface release, and PNL 20 tectonism-as water table rise and additional gas flow 21 through fractures.

22 Sandia did not incorporate tectonism as'an aspect.

23 And then each one of these was rolled up'into a 24 conditional CCDF, and I think Felton may talk a'little bit

-25 about how those were combined and weighted. And we h ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 ,

(202) 293-3950-

181 r' 1 certainly did talk about it yesterday.

(/

2 And then the ultimate goal was to come up with c 2 3 total system CCDF that would represent these features.

4 (Slide.)

5 MS. DOCHERY: The transect that was modeled by 6 both PNL and Sandia is this one right here that went from H-7 5 on the west across the Ghost Dance Fault which was modeled 8 with 14 meters of fset by PNL who was doing the 2-D -

9 simulation through G-4, UE-25 A #1, and then 500 meters to 10 the east.

11 The reason that this was extended a little bit 12 farther was to try to account, or try to take any modeling 13 artifacts like ponding that you would get up against the

[J h 14 15 boundary far enough away that it would be -- it wouldn't interfere with the understanding underneath the repository.

16 We picked this particular domain or this 17 particular transect as being representative of the entire 18 repository, although we know that there may be changes as 19 you move south in the repository.

20 However, it was at this area that we had the most 21 information on geohydrologic parameters.

22 [ Slide.]

23 MS. DOCHERY: Again, the release pathways for each 24 of the individual scenarios was for volcanism, gaseous flow 25 and human intrusion, it was direct to the surface; O)

ANN RlLEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300  ;

Washington, D. C. 20006 )

(202) 293-3950

l 1

182- ,

l 1 1 tectonism,. aqueous flow and human intrusion =down~'into~the j 1

2 saturated.tuffacious zone, and-then another human intrusion

-3 down into the carbonate aquifer and the paleozoics.

4 I'll change this slide and never use'it-again.

-l

- - 5 Common factors. '!

6 [ Slide.)

7 MS. DOCHERY: The stratigraphy was common between ,

l 8 the two exercises. The distribution of geohydrologic -l I

9 pt- ameters, at first we thought the source term, but that' 10 didn't work out, and the-boundary conditions.

11 MR. MOELLER: Excuse me. On the two slides back, -l 12 when you showed human intrusion, you had it go through the 13 groundwater layer, and then down into the carbon'-- whatever-14 it is.  !

15 MS. DOCHERY: Carbonate aquifer.

16 MR. MOELLER: Is there some significance to that?-

17 Could you have had human intrusion just into the' aqueous l

18 flow where they are drilling for water, and - '

19 MS. DOCHERY: Well, the first one here, human-20 intrusion here, was-to taks into account drilling into_the 21 -- just to the water table. But-we also thought in this

^

22 kind of environment, the likelihood that you-would just.

23 drill.for water-was fairly unlikely.

]

24 However, if you'were an exploratory' driller for 25 minerals,- you would want:to-tag-the-basement.

ANN RILEY & ASSOCIATES, Ltd.

Court Reporters-1612 K. Street,- N.W., Suite - 300

. Washington, D. C. 20006

-(202)-293-3950 d

183

-1 MR. MOELLER: Okay. So you do have both.

2 MS. DOCHERY: Right.

3 MR. MOELLER: I missed that.

4 MS. DOCHERY: And there was a reason for that, 5 partially because.the information we have in here is not 6 very fast flow, but it is much more rapid in this region, 7 and so there were technical reasons both in terms of 8 drilling strategies as well as transport, to try to 9 understand these differences.

10 (Slide.]

11 MS. DOCHERY: The cross section that was developed 12 for the TSPA included up to five layers, and we will see, 13 depending on where you went down to the water table, you

(} 14 15 might go through either all five layers or through only a few of the layers before you hit the water table. And so 16 you will see in Felton's presentation, I think, there were 17 some differences in the calculations that were run in this 18 region versus in this region.

19 I mentioned that PACE had 18 layers, and here we 20 have five, and you might wonder why that was. I think this 21 is -- part of what we hoped to gain out of abstraction, we 22 did the 18 layers, we found out that the hydraulic 23 conductivitics, the contrasts were relatively small, and we 24 just didn't see that there was a big difference that was 25 caused by having all the layers. People that do remember

] ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., _ Suite 300 Washington, D. C. 20006 '

(202) 293-3950

1 l

184-1 1 back to PACE may_ remember that there.was a layer in the --

2 there's a-TPT-T&B that we artificially used a very different 3 conductivity.

4 However, that information was stuck in.more to try-5 to stress-the codes than it was bas 6d on site data.. So we 6 went back to try_to fit the site data more realistically, 7 and simplified the stratigraphic section, making it 8 computationally quite a bit.more simple.

9 MR. POMEROY: Holly, in terms of this-cross 10 section, is-that what you used, or did you extrapolate from.

11 that for a 3-D? Did you analyze this?

12 MS. DOCHERY: We didn't do a 3-D on TSPA. Sandia

)

13 did 1-D,-and PNL did'2-D.

14 MR. POMEROY: Doesn't-the regional flow of the-15 water -- isn't that almost. perpendicular to this section?

16 MS. DOCHERY: Raleigh, do you want-to --

17 MR. POMEROY: But you showed it going off to the-18 east.

19 MS. DOCHERY: That was a cartoon. That was just 20 schematic. The regional flow fields that were used for 21 Czarnecki's --

22 MR.-POMEROY: Where are-the flow----

23 MS. DOCHERY: The Czarnecki fields were the 24 -northeast to southwest flow fields.

25 Raleigh?

- ANN RILEY & ASSOCIATES, Ltd.

_ Court Reporters 1612. K. Street, N.W., Suite- 300 Washington, D. C. 20006 (202) 293-3950 R

1 18F jr-) 1 MR. BARNARD: I am Raleigh Barnard from Sandia.

V 2 Labs.

3 In the unsaturated zone, we have as much 4 confidence as we have in anything to say that the water 5 essentially flows down.

6 [ Laughter. ]

7 MS. DOCHERY: And we don't have a lot of 8 confidence in that.

9 MR. BARNARD: Rather than laterally. Certainly 10 there's a lateral component.

11 When you get to the water table, there would be 12 lateral flows both in the plane of that diagram and 13 perpendicular to the plane of the diagram, and that is the

() 14 part that we ignored to a certain extent, although by using 15 flow tubes which don't have to go in the direction-shown in 16 that diagram, we did take into account the actual direction 17 based on the Czarnecki regional flow model.

18 MS. DOCHERY: Does that answer your question, 19 Bill, more or less?

20 MR. POMEROY: Yes, thank you.

21 MS. DOCHERY: The data that we obtained came from 22 this H-5, G-4 and UE-25A, the three holes that were logged' 23 by the USGS. We obviously had to lump multiple layers 24 together, and so we just picked the dominant lithology in 25 the individual section, and represented that stratigraphy or l

.O Ann ni'av a Associates. 'ta.

COud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

186 1 that layer by the dominant unit.

2 The saturated zone, as you may recall from the 1

3 diagram that I had before, was represented as two layers.

4 The elevations really didn't matter because in the human 5 intrusion we assumed a direct release where something was 6 just dropped directly down and there was no travel time, but 7 we did base that on the information that was available in 8 the lith logs, and in the upper part was simply the 9 tuffacious saturated zone. We have used the term tuff 10 aquifer. The USGS doesn't like us to call it an aquifer, 11 because it is not very permeable, it's non-welded tuff, and 12 water really doesn't flow through it very readily, as 13 opposed to the carbonate aquifer which truly would be

() 14 considered an aquifer in a geologic sense, and the point 15 that Bill is asking about is that the flow fields were 16 represented using Czarnecki's model that flows from the 17 northwest to the southeast across the repository.

18 (Slide.]

19 MS. DOCHERY: As far as the geohydrologic data 20 that were used for the two exercises, we felt the need to 21 have a range of hydrogeologic properties to do the 22 stochastic simulations, and so those ranges were derived 23 primarily from site data, but there was some information 24 that was taken from analog data, and in the case of the Kds, 25 in some of the volcanic information from expert opinion O ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 -l (202) 293-3950 1

l

187

_j- 1 elicitation, the hydrologic properties were represented as

\~' 2 probability distribution functions, and I would really hate 3 to go into great detail on Paul Kaplan's use of Ken's 4 informational interfeed theorem and trying to explain in 5 detail how we came up with the elicitations, but if that is 6 something you would like to hear, we could-talk about that 7 or you could read about it in any one of a number of papers 8 that Paul has written, or in the TSPA.

9 However, the ultimate goal was to try to reflect 10 not only the natural variability in these parameters, but 11 also our uncertainty. So the shapes of the distribution 12 were chosen to reflect what the experts felt or what the 13 data showed about the uncertainty in those particular 14 parameters.

(}

15 Again, we ate not going to stand beside those 16 parameters and say we absolutely believe these and we are i 17 never going to change them. We have already changed them.

18 The document, the data document that we were generating in "

19 association with this TSPA is there's a data group that's 20 already working on new and better distributions, and so as 21 data become available from the site, we are continually-22 revising and changing that.

23 There was no cross correlation structure 24 incorporated in Sandia's TSPA, and that is another aspect' 25 that we are working on and hope to be able to incorporate in

(] ANN RlLEY & ASSOCIATES, Ltd.

Court Repoders 1612 K Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

188 1 the next go-round of the performance assessment.

2 The parameters -- the specifics of-the parameters 3 that were -- for which there were PDPs generated were 4 saturated hydraulic conductivity, velocity, water content, 5 the Vandenuken parameters, residual saturation, and fracture 6 density and apertures. And there was a distribution that 7 was generated for each of the five layers. So there were 8 different distributions for those properties within each one -

9 of the layers.

10 The additional parameters that we had PDFs for 11 that weren't necessarily used in the aqueous flow 12 calculations were for volcanism, dike length, width, 13 orientation, extrusive volume and percentage of wall rock 14 entrained, ar.d I have a little bit more data on the Kds if

(}

15 you are interested in seeing that. I've got the viewgraphs 16 that have what is used for those particular -- for the 17 radionuclides for Kds.

18 [ Slide.]

19 MS. DOCHERY: The data set was -- this 20 geohydrologic data set with PDFs were applied for the flow 21 and transport calculations for Sandia, both the unsaturated 22 and saturated groundwater scenarios.

23 PNL, because they were running disturbances as 24 excursions on the groundwater scenario, used them for all of 25 their scenarios.

Q ANN RlLEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 I (202) 293-3950

189 gx 1 The tuff retardation coefficients that were used

%-]' 2 -- this came from Aaron Meyer at Los Alamos -- there were -

3 -the rock types were subdivided and there were PDFs 4 generated for certain radionuclides for vitric, nonvitric, 5 and zeolotic rocks.

6 Briefly, the nuclides that had no retardation were 7 technesium, iodine and carbon. Those that were assumed to 8 be essentially completely retarded greater than 100 9 milliliters per gram were amersium, plutonium and tin, and 10 then we had specific probability distribution functions 11 assigned to uranium, selenium, cesium and neptunium.

12 The range of retardations was established for pH 13 based on J13 water,'and oxidizing conditions were assumed as 14 a conservatism.

15 [ Slide.)

16 MS. DOCHERY: The carbonates, we don't have any 17 information from the site on retardation in the carbonate 18 aquifer. So we turned to the WIPP data. They have a report 19 out that includes information on retardation in the Culebra 20 Dolomite, and so we took their matrix values -- there were a 21 lot of clays in the fractures, and so Aaron didn't feel like 22 that was appropriate for representing the Yucca Mountain 23 site.

24 The water chemistry again assumed oxidizing 25 conditions, and those of you who are familiar with WIPP know ANN RILEY & ASSOCIATES, Ltd.

t]'- Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

190 1 that it's a salt site. However, at that point, it was l 0 2 assumed that chlorides didn't have any effect on the Kds.

3 There was information available in the literature that 4 seemed to indicate that one of the NDWTRB interchanges -- I 5 think it was Pat Domenico, someone suggested that maybe 6 there was additional information that might suggest 7 different. So that is something we need to look into for 8 the next go-round.

9 And in this case we only had our Kds for 10 plutonium, americium and uranium.

11 [ Slide.]

12 MS. DOCHERY: The boundary conditions established 13 for Sandia and PNL were for the 2-D calculations, no flow 14 lateral boundaries. The calculations were run-from_ initial 15 saturation and flux to steady state that was appropriate for 16 the percolations that were imposed.

17 The ranges in flux -- and this is where we had --

18 this is what took up an hour yesterday, was the range picked 19 was zero_to 39 millimeters per year. We believed that this 20 range of values would a tually take into account what one 21 might expect with climate change.

22. I know that there are people who may not agree 23 with some of the. winding calculations that were done in 24 association with PACE, but the inverse calculations done 25 using that specified stratigraphy and the flux rates showed

.O i

^"" aev $ ^SSoci^TES' 'ta-Court Reporters l 1612 K. ' Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

191 7 1 that the -- given the saturations available at the site e

2 right now that -- the measured saturations -- it was -- the 3 implied flux values would be most appropriate between zero 4 or minus zero and .01 millimeters per year. And we do 5 realize that those are speculative, but because of that 6 information and the information that Allen Flint gives us to 7 indicate that Rainer Mesa may be our best analog for climate 8 change, being a vetter area that has something on the order 9 of .1 millimeters per year percolation, that we could weight i

10 our values more toward the lower end of a distribution l 11 function.

I 12 (Slide.] I 13 MS. DOCHERY: I cringe, but this is what it looks l

()

14 like, guys. Don't throw anything from over there, okay?

15 [ Laughter.)

16 MS. DOCHERY: We were very spirited yesterday in 17 our discussion on this particular viewgraph.

18 One of the reasons that we pushed it to 39 was 19 because we wanted there to be some values that would be 20 sampled in this distribution that would force the system 21 into fracture flow. We wanted to, as opposed to PACE, where l

l 22 there was basically a composite porosity matrix flow sort of 23 calculation, we wanted to force the issue and swjtch over 24 into the fracture flow domain. And so that was part of the 25 reason of having the extremely large values.

, O O ANN RlLEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 o

\

192 1 We want to get more information from the survey, 73 2 and I think we decided that this was the point where we 3 might want to have an intorchange with the NRC and the USGS 4 present to try to get a better handle on what sort of 5 infiltrations might be appropriate, but at this point we 6 didn't feel like this was an unreasonable representation of 7 what might happen, what is happening currently, and what 8 might happen in the case of climate changes. With the 9 exception of not having any negative values, and there is a 10 contingent out there that believes we ought to have negative 11 values, that we are actually having drying in the Yucca 12 Mountain region. So that that might be more appropriate 13 than to not have zero as a possibility.

,O 14 (Slide.]

U 15 MS. DOCHERY: The caveats on these exercises is 16 that we did our best to try to deal with reasonable values 17 of parameters and processes that were relatively reasonable.

18 However, we didn't model all the components. We have event 19 trees that have multitudes of scenarios that we could model.

20 We thought at the time that we might be picking 21 end members, most dramatic effects, highest consequences, or 22 most likely, but one of the things we have learned is that 23 in the case of volcanism, it may actually be that some of 24 the indirect effects are more important. So we obviously_

25 have to model more and more components as we go along.

O ^"" aev & ^SSoci^TES' d-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

193 1 We have no validation of the data and models, but 2 we had to start somewhere, and this is where we chose to 3 start. And the ranges of values that we chose were very 4 broad to indicate as much ignorance as possible. We hope, 5 as time goes on, and site characterization proceeds, that 6 some of those values, some of those parameter ranges, may be 7 narrowed somewhat.

8 So the last thing I will say again is we don't 9 expect to use these as the answers. I doubt seriously that 10 we will come to the Licensing Board with these particular 11 calculations, but they were useful for determining the 12 sequence of, or the prioritization of testing for site 13 characterization, and in that sense we think it's a useful 14 exercise, and will become even.more useful as we get time to-15 do more sensitivity studies, and find out where the really 16 important parameters may lie.

17 I think, although I haven't seen Felton's 18 viewgraphs or Paul's viewgraphs, I think what they are 19 planning on presenting is looking at the calculations, both 20 the detailed models and the abstraction models. You will 21 see some preliminary CCDFs that have been constructed.

22 Paul, are you going-to talk about the dose-based 23 limit, or -- okay. And then Jerry will follow up with a 24 little bit of information on where we hope to go with the l 25 next iteration of the total system performance assessment.

l Q ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 l Washington, D. C. 20006 (202) 293-3950

I 194 1 MR. POMEROY: Thank you, Holly. j O' 2 MR. APTED: Holly, the TSP was based on the SEP l

3 near-field design vertical package -- l t 4 MS. DOCHERY: Yes.

5 MR. APTED: You know, obviously there is some talk l

6 about changing and coming up with those more robust wasto 7 package designs. Are your models themselves robust enough 8 to take into account --

9 MS. DOCHERY: I think Sandia's were kind of like 10 brown paper bags. So we can certainly make them robust.

11 MR. APTED: Much bigger bags.

12 MS. DOCHERY: Much bigger bags. We want sono 13 Hofty bags, trash bags. And that's -- when you see where we 14 would like to go, we woro just beginning to got a little

{

'5 more sophisticated in the source term. I think Mike would be mor >1 eased than anything if he could actually have a 17 container that would be a little bit more reasonable. As it 18 is rig'ai now, we're taking very, very low credit for our 19 engineered barrior system. Cortainly I think much loss than 20 in reasonable.

21 MR. STEINDLER: Did your transport include only 72 matarlais that were soluble in the groundwater, or did you 13 include colloids, or --

24 MS. DOCHERY: There were no colloids included in 25 this particular iteration.

O ANN RFLEY & ASSOCIATES, Ltd.

Coud Repoders 1812 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

195 l

1 MR. STEINDLER: Are you going to look at that to 2 see whether --

3 MS. DOCllERY: That's one of our wish lists, and 4 there is a workshop that is going on -- I believe it is in 5 May -- between some of the PA people and some of the 6 geochemistry working group people to try to get -- find out l 7 what is in the literature, what could we use to make a more 8 reasonable assessment of colloidal transport.  !

9 MR. PoMERoY: Looks like that's it, llolly. Thank 10 you very much.

11 The next presenter is Felton Bingham.

12 MR. BINGilAM We tried to imagine what would be a 13 good thing to present here, and after lintoning to the talks

[ 14 this morning, I have decided that I guessed wrong.

15 (Laughter.)

16 Mh. BINGilAM t Sc what I would like to do is to 17 kind of depart from what I had really intended to do. I 18 think it might be more valuable to deal more with some of 19 the kinds of questions that I have heard here.

20 I noticed that the NRC Staff, who are more 21 experienced in dealing with this particular committee, 22 didn't do the thing that I had planned to do, which is kind 23 of to run through the whole total tystem performance 24 assessment in more or less detail. ~

25 I have still got the viewgraphs to io that, but I O ^"" ai'ev a Associates' 'ta-Court Reporters 1612 K. Street, N.W., Suito 300 Washington, D. C. 20006 (202) 293-3950

196 1 think I will cut back on doing that, because I would like to 2 try to deal a little more with some of the issues that the 3 committee has been raising. And how well I do that, I'm not 4 sure of, because I got to thinking while llolly was talking 5 what I might do, and decided I'd better explain to you that 6 I'm departing, because you will have two things to do:

7 one is you may have to kind of flip through some 8 of these viewgraphs as I have decided to discard them, and 9 you also may have to listen to statements that I will 10 immediately retract, because I'm not sure exactly in advance 11 what they are all going to be.

12 You have seen several times what the purposes of 13 this TSPA are, So this time instead of going through those, 14 I want to point out one or two things.

15 I was delighted to hear Norm speak of this 16 abstraction process this morning, to hear him say it in 17 terms of there are going to be simpler models that produce 18 the CCDFs, you are going to have to use the more complicated 19 models to tell you how good those are.

20 That's what we mean by this abstraction process.

21 We keep picking at it because most of the time our audiences 22 seem to greet us-with silence when we say something about 23 it. But I remembered that four years ago in this very room, 24 we put up a picture that looked like this.

25 (Slide.]

l {} ANN RlLEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l 197 1 MR. BINGilAM And I assumed that everyone would Oi 2 remember, of course, what it meant; that Norm was saying 3 that the modeling you do dtwn here with these complicated 4 ^

• tais. tic models is the real support for the total system 5 rOdeling that is up near the top of this pyramid.

6 Well, that was, to me, anyway, the primary thing 7 we were trying to get out of this exercise. We were trying 9 to get lots of things out of it. Those of you who are good 9 at counting beans will notice that each one of these 10 purposes slides --

11 (Slide.)

12 MR. DINGilAh: -- has a different number of 13 purposes on it. But this one is the one I would like to try

() 14 15 to emphasize in some way.

And there is one other point on here I want to 16 bring out, too. I have heard questions coming that seem to 17 be things like how close are you to the DOE, how well do you 18 understand what they are doing? Are they doing what you 19 vant? I think that's the kind of question you'd be 20 interested in hearing. And somehow or other I was lucky 21 enough to guess in making up the viewgraphs that that might 22 be something you'd want to hear a word about.

23 I think it is worth pointing out that a lot of 24 what we are doing, we are doing because of the influence of 25 our interactions with the NRC Staff.

O ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 j l

198 1 I think the site characterization plan itself was O 2 heavily influened by those. But in performance assessment, 3 I think we have tried to como up with things that will 4 satisfy the needs that the Staff has expressed to us as they 5 grow and learn themselves what to do. And I hope to be able 6 to point out somewhere during the going through the TSPA 7 what some of those things are that we have done.

8 But the point here, and the reason the dot is 9 underneath the abstraction process, we think that is what 10 the NRC Staff has always asked us to do. It's connected 11 with the difficult word that -- to me, the dirtiest of the 12 two dirty words in this whole business is the ten-letter i

13 dirty word, validation.

14 This is all connected with that thorny, difficult

(

15 issue. To me, the validation really means I am a regulator, 16 you have brought me an analysis, I want to know why I should 17 believe it.

18 Well, believint; it, a total system analysis, is it l 19 seems to me in luirge part depending on one's ability to take l

20 the complicated tnalyses and make from them the simple ones 21 that really contri.buto directly to the performance measure 22 that you are going to have to look at.

I 23 This is the first example, and I may forget to do 24 the others, because I hadn't planned to do it, but this is 25 one of the examples where we think we were heavily Q ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006

-(202) 293-3950 1

L

199 1 influenced by our interactions with the NRC Staff.

2 (Slide.)

i 3 MR. BINGHAM: As you look through your things, you l

4 will find -- I believe I will skip the next one. I had 5 something about the schedule because I believe there are 6 some quirks to this TSPA that are explainable only by 7 looking at the schedule on which they were produced. But 8 let's skip that for the time being, unless we have to.

9 Let's skip the next slide about the EPA standard.

10 Holly has already talked to you about the 11 conditions we modeled, and let's go on to the -- something 12 thrit I think is really very important, and that is the names 13 of the people who worked on this.

14 There are a lot of names here. And they are, of 15 course, as they are traditionally, roughly in order of the 16 amount of time and lifetime they gave to it. It's probably 17 worth pointing out that, of these people, the ones who were 18 devoted almost full-time are the first four names, and the 19 first three names are the ones whose lives will probably 20 never again be the same because of the permutations they l 21 went through to produce this.

22 (Slide.]

23 MR. BINGHAM: Holly has already showed you a more 24 colorful version of this, and I'm now going to try to go l

25 back to what I had intended to say for a while. This is O ^"" ai'ev & Associates' 'ta-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

200 1 sort of an outline for the talk, except that I want to do it 2 in a different order. I want to talk a little about each of 3 those components that we did, except, of course, for 4 tectonism, which PNL did, instead of Sandia.

5 I am going to try to go through them, emphasizing 6 things that I think haven't been discussed well before, and 7 then enough just to give a little flavor for the folks who 8 weren't here during the last -- during yesterday, to hear 9 the details of how they work. We do stand able to answer 10 detailed questions though, not by me, but by doctors Wilson 11 and Barnard, who stayed over today to be able to answer 12 those difficult ones that I would look ridiculous trying to 13 answer.

G[ T 14 (Slide.)

15 MR. BINGHAM: The next slide is in here not 16 because I want you to follow through with it, but because .I 17 want to make a more abstract point with it. This TSPA is 18 one of the way stations on the route to fulfilling the 19 process that the site characterization plan lays out for 20 producing the total system part of the license application.

21 One of the steps in there is first to decide what 22 are all the things that can happen at the site -- what 23 processes can go on there? Which ones are going on? Which '

24 ones may go on, and to give everybody confidence that you've 25 thought about everything. Well, that's not an easy thing to O ^"" ai'ev a Associ^Tes' 'id.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

201 1 do. But, we are devoting quite a bit of effort to trying to 2 decide what are all the things that could happen at the 3 site?

4 The bookkeeping devi::e we adopted for doing this 5 is the use of trees, and this is one of them. This is a 6 piece of the tree for volcanism. We developed these thirigs 7 separately from the total system effort, in many ways, by 8 trying to consult with the experts we have mostly within the 9 project, on what things could happen at the site. We found 10 that the use of this tree is useful, not only because it 11 helps us remember what we thought of and what we didn't 12 think of, but because those experts tend to respond to it.

13 It's hard to go to an expert in say volcanism, and say tell f 14 me bhat I need to know to make a CCDF?  !!a'll say get out of 15 my off'.co. But, if you ask him what would happen if a 16 v'alcano came up at the site, and let me right it down, and 17 then you tell me what would happen next, you could l

18 comuunicate with him. And, in fact, we found that they're l 19 kind of interested in doing that. They sort _of think it's 20 fun. It allows them to express their erudition in ways that 21 performance assessors can only goggle at.

22 (Laughter.)

23 MR. BINGilAM: As I say, this is the piece of the 24 tree for volcanism that we've developed so far. And we 25 thought, in doing our modeling, we ought to try to mimic O ^"" ai'ev ^ Assoc'ATes' 'id-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950 .

202 1 what we hope we'll be able to do for the final license 2 application, and that is, go back to these trees and decide 3 what in them needs to be modeled and what doesn't. Well, 4 we're not at all ready to say what doesn't need to be 5 modeled yet, but we can look at this tree, and all the 6 branches that have of course disappeared, because they're 7 shown here as dashed lines, to think what is it that we can B model? And we decided that probably the most important 9 thing to model right now is the one that is a little bit 10 darker here -- that we get some intrusion, that directly 11 acts on the repository through the formation of a dyke -- a 12 cone, and waste that comes right up to the top.

13 It was pointed out yesterday that now maybe we're

(~h 14 not so sure. Maybe there's something else that might even l G

15 have been more important. But, for the time being, this 16 looks like a good way to get a handle on the problems with 17 dealing with volcanism. So, we began by going to a tree 18 like this.

19 If you take the pieces that were in that tree that 20 were dark and draw a picture of them, the picture looks like 21 this. Here's the repository, it's back-filled, and here are 22 some canisters sitting on the floor, and the dyke comes 23 whooshing up, pushing the rock out of the way as it comes, 24 it intercepts the canister in this picture, and fragments 25 some waste and carries it up to the top. And the question Q ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N W., Suite 300 Washington, D. C. 20006 (202) 293-3950

203 1 then becomes how do you model that? If that's the picture 2 of what you think is going to happen, how do you model it?

3 We found these pictures, by the way, to be very 4 useful too for communicating with people. If you go to the 5 trouble of making them, they do becomo an effective way of 6 finding out what it is the expert really has in his head.

7 Well, we developed a model for doing that sort of thing, 8 and, as part of this economy effort, we're not going to go 9 into the details of the model unless there are questions 10 about it.

11 The point is -- maybe I should say something --

12 that to make a model like this, you have to know a lot of 13 parameters. You have to know things like the width of this 14 dyke. You have to know something about where the canisters 15 are and how they are arranged. You have to know something 16 about the ability of this dyke to rip rock away from the 17 walls as it comes up, because that's something like it's ,

18 ability to rip spent fuel fragments out of the containers.

19 The model that we developed has something in it for all of 20 these things, and, for most of them, they are distributions 21 that reflect what we think the distributions in rature are 22 for those parameters and, as llolly has pointed out, 23 frequently contain our estimates of what our uncertainty 24 about them is.

25 For eliciting these distributions, we came up Q ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington,- D. C. 20006 (202) 293-3950 l 1

9 . . . . .

204 1 with, as an experiment, a little technique that not much has 2 been said about in -- even yesterday. Holly mentioned it a 3 little today. It's sort of like -- it involves having the 4 expert come and sit down and be asked some questions like 5 what do you think the width of the dyke is? What data do 6 you have that makes you think what that is? What variation 7 have you seen in nature? How uncertain are you? Do you 8 know a range, a mean? What do you know? When the expert 9 gives these answers, we have somo software set up so that he 10 sees immediately on a computer screen the distribution 11 that's being elicited from him, and he can say, oh, ay 12 heavens, I didn't think it's anything like that. Or, after 13 a few iterations through a process like that, he can say

(} 14 15 that's pretty close. I believe that expresses what nature has in it, and maybe even how much my uncertainty is as 16 well.

17 We did this as an experiment. When we reported 18 it, up till now, it seems te provoke some kinds of 19 controversy. I think, as it settles down a little better 20 just what it was we did and how simple it is, it may turn 21 out to be a useful tool. Someone may be able to find 22 something wrong with it. But, it's the sort of thing that 23 we would like to propose as a way of dealing with some of 24 the expert judgment problems we have, not all of them. One 25 of tae objections I have heard raised to it is it doesn't O ^"" ni'ev a Associates' d.

Coud Repoders l 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

205 1 solve all the problems. Yes.

2 Well, with this model, most of whose features I 3 have left mysterious, we can come up with releases at the 4 surface. this shows what they're like, and basically 5 they're kind of like the inventory, because we are just 6 ripping up pieces of the spent fuel and bring them to the 7 top, without much selection with processes like dissolution 8 or precipitation or all those other things that drive most 9 people crazy.

10 The differences in here are that, since the time 11 at which the event occurs is a random variable, it is not 12 exactly the inventory at any particular time. With that, 13 and some estimates of the probabilities, we can come up with

(} 14 15 a conditional CCDF just for volcanism. It looks like this.

And, at this point, I would like to try to make 16 some philosophical remarks that are along the lines of those 17 that I'm hearing. There are questions being asked that seem 18 to be -- is there ever going to be a way to get enough 19 empirical dat6 to feel confident about any of these things?

20 Well, I certainly don't want to claim that this answers that 4

21 question. But, I would like to claim that it is pointing a 22 way towards possibly answering it.

23 This, of course, has got to withstand the scrutiny 24 of a great. deal of expert opinion and inexpert opinion. So, 25 it's not at all complete. We know we haven't modeled all j 1

l C ANN RILEY & ASSOCIATES,- Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950 i

i l

206 I 1 the sequences that we need to look at. But, if we continue 2 along this vein, and we continue to get distributions that 3 look like this, it seems reasonable to me that people who 4 have to assign priorities to what needs to be done would 5 look at this and say, if you can convince me that this is 6 anywhere near right, let's don't worry a lot more about 7 volcanism.

8 Now, it's striving towards that point that I think 9 -- I feel the license application would like to go. We'd 10 like to get to the point where we can say no further work is 11 required on overy one of the issues. This kind of analysis 12 that we've got here, reported in full, with all the 13 distributions and all the ways it was obtained, leave open 14 the possibility for something that is a practical way of 15 deciding whether enough has been gained. We would like to 16 see people take this analysis and go through it step by 17 step, viewgraph by viewgraph, answering, at each point, 18 what's wrong with this? Can it be improved? Is it good 19 enough yet, or not? That will lead us, I think, to the 20 point where we can change any one of the viewgraphs, or any 21 one of the parameters, and finally decide for sure that 22 we've done enough.

23 This is maybe the most trivial example of how that 24 process may work, but I think it's enlightening. I think it 25 shcws us what the basic steps are. The foremost among them, O ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

207 1 I think, is getting the answers out in a reasonable, easy to IO. 2 understand form, so that critics of it can criticize it, not 3 on the basis of hey, that must be wrong, but on the basis 4 of, your picture of a dyke is wrong, or your distribution 5 for a dyke is wrong.

6 Even though that looks a little -- we did try to 7 do some sensitivity studies, and since I'm really not trying 8 to convince you that it is either right or wrong, I won't 9 dwell on this much, except to say that we did do some things 10 like suppose the wall rock fraction, that's the fraction of 11 rock that can be ripped out and carry spent fuel with it is 12 increased above what the empirical data that we had at our 13 disposal say it is? Well, you can make the CCDF get farther

(} 14 15 over to the right. Suppose we say the eruption volume, another one of the parameters has changed? We can still get 16 farther-to the right. But, we still aren't getting very 17 close to the limits. This is another part of that process 18 that I think eventually leads to a real answer to.that 19 question of will we ever have enough empirical data.

20 The next thing I want to talk about is the 21 groundwater flow, and the gaseous flow all at once.

22 MR. POMEROY: Felton, before you do that, can we 23 just talk ono -- 30 seconds perhaps about that question? I 24 must have missed something. But, someplace you must have 25 told me -- your said you brought an expert in -- one or more ANN RlLEY & ASSOCIATES, Ltd.

Q'- Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

208' 1 1 experts in, I don't know how many.

2 MR. BINGHAM: Uh-huh?

3 MR. POMEROY: I think you said largely in-house 4 people. Then you've got their composite views. Could you 5 just expand on that a little bit? Is that a hundred percent 6 Bruce Crow, or --

7 MR. BINGHAM: Bruce Crow and Greg Valentine.

8 MR. POMEROY: And Greg Valentine.

9 MR. BINGHAM: Yes. Those are the two people that 10 we use for-doing this -- for getting things like that' width 11 -- that wall rock fraction. And, of course, we try to pin 12 them down to what -- we really don't want your speculations, 13 guys, we want to know what you know. What do the field data 14 show?

15 And, I guess, this is another place that-I want to ,

16 mention how we worked under the-influence of.the-NRC.- Among 17 the early comments on the site characterization plan, were

-18 those that revealed the fear-that somehow we're-going-to use ,

19 expert judgment in place of going out and getting data. And ,

20 I-noticed that the replies to our replies to those comments

'21 still seem to be revealing'a little of that fear.

22 Well,.we are trying to deal with that by showing 1

23 you a little of how we expect to go. We don't intend to go 24 to a bunch of experts and say, okay,--give me the-wall-rock-25 fraction, thank-you, sir. We'll put it into our license-Q- ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612' K. Street,7 N.W., Suite -300

.. Washington, D. C 20006 (202) 293 3950

I 209 1 application. We would like to pin them down and make them 2 give us some references on field data where we can. And I 3 like using this as an example -- I am afraid it may be a 4 trivial example -- it may be the easiest one -- of a place 5 where data turned out to be available. Now, how well those 6 data will stand up to the scrutiny of further experts, is a 7 matter of practicality, rather than theory.

8 MR. POMEROY: Okay. Thank you.

9 MR. STEINDLIR: Can I take this philosophical 10 discussion one step further? Is it your aim to ultimately 11 set aside questions of the kind we've just looked at because 12 they are so far below the EPA limit, assuming the limit 13 holds, and have it set aside by a process of rulemaking, so

(} 14 15 in fact you not only can put it behind you, but you indeed never have to look at it again in the licensing process? or 16 is it that you're trying to -- let me give you two choices, 17 and then you can exercise third -- or is it that what you're 18 trying to do is tell the Department that hey, guys, there's 19 no point in spending any more resources on this, and we've 20 got some other things that are more important?

21 MR. BINGHAM: Yes.

22 (Laughter.]

23 MR. STEINDLER: Have you conveyed the first to the 24 NRC and the second to RW?

25 MR. BINGHAM: The first you asked did have the Q ANN RILEY & ASSOC!ATES, Ltd.

Court Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

._ m __

210 1 word rulemaking in it.

2 MR. STEINDLER: Yes.

3 MR. BINGliAM: And, of course, that gives me the 4 willies. That is the question the DOE has got to answer, I 5 think, rather than some contractor, who probably can barely 6 spell it. But, it seems to me it would probably be a good 7 idea, if you could deal with issues by rulemaking and get 8 rid of them before the hearings and all of that starts. And 9 I just don't know whether that's practical, or whether 10 that's a good idea. It seems to me, to a country boy, that 11 it would be a pretty good idea.

12 MR. STEINDLIR: So, your first answer of yes was 13 qualified by not knowing now to spell rulemaking?

() 14 15 MR. BINGilAM: Yes.

One of the things I want to get to a little later 16 is the question that has been raised. I think the question 17 was asked from the NRC staff, and should be asked us as 18 well, or Jerry, how are you communicating this stuff to the 19 rest of your program? You were ir.terested in how do you 20 communicate with Mel Silverberg, and we're interested in 21 communicating with the whole raft of people who have 22 business in site characterization. And I would like to talk 23 about that a little later, after I have got a few more 24 examples.

25 MR. OKRENT: I probably ought to note that when C ANN RlLEY & ASSOCIATES, Ltd.

Coud Reponers 1812 K. Street, N,W., Suite 300 Washington, D. C. 20006 (202) 293-3950

211 1 this was discussed yesterday, Jchn Trapp stood up and noted 2 that the NRC has reservations t.;wnfterning volcanism, and 3 there is to be a meeting on that subject. So, there was not 4 necessarily concurrence with what seems to be a very big gap 5 between the EPA step criterion, and this result.

6 MR. STEINDLER: My comment was based -- ignoring 7 facts, as I sometimes do -- just trying to get at the rolo 8 that you see of setting things aside, and how you intend to 9 implement them. What is it you're going to do with this 10 interesting bit of information?

11 MR. BI!1GIIAM: And, by the way, I may have been a 12 little coy in talking about the kind of criticism that wo 13 would like to get. I was delighted yesterday to hear John 14 say that he intends, at least, to wrestle with some of these 15 issues in another workshop. I think that we, in performance 16 assessment, would like to see those people in that workshop 17 put a little of their attention into telling us what we did 18 wrong or right, if anything.

19 MR. TRAPP: This is John Trapp. I would like to 20 just make one comment from a staff viewpoint. DOE lately 21 issued a topical report on the area of origin. And this 22 topical report was hoped to solve some of these technical 23 problems, bring them on the table, et cetera, discuss them, 24 and see if we can get at least to a staff agreement. This 25 report I am talking about, the interaction I'm talking O. ^"" ai'ev $ Associates. 'id-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

a_ -

212 1 about, is supposedly a similar topical report in the area of 2 volcanism which will go back and forth. Whether we'll come 3 to closure or not is really -- these are the first two 4 topical reports, and I have no ides if we are going to come 5 to closure, or raise more questions. Come back to us about 6 the end of June, and we will be able to tell you how well 7 the process works.

8 MR. STEINDLER: You know, John, the point I guess 9 I would make is I am heartened to see that people are at 10 least spelling the words closure. There was a time not too 11 long ago when that thought some how never rose to the 12 surface. And I think great progress is being made.

13 MR. TRAPP: That's really the whole point. We're 14 trying, in staff closure, I don't know if we will make it.

15 And we will see what happens.

16 (Slide.)

17 MR. BINGHAM: Yesterday there was a lot of time 18 spent talking about the things that are pictured on this 19 viewgraph, because this is the source material for the 20 source term. And that is of great interest to everybody.

21 So, I am going to say very little about it, only partly 22 because I don't know that much about it, also because it 23 looks like it takes so long to do, if you open the door a 24 crack, to discussing it.

25 But, I do want to say that, for predicting aqueous Q ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

213 1 and gaseous releases, we had to answer the question that 2 appeared on some of the NRC staff's viewgraphs prominently 3 this morning, which is how does water get to the container?

4 I have boon in the program for nine years, and nobody has 5 told me that yet. When one first joins the project, and 6 roads about how dry things are, and how they will stay dried 7 out for a while, and how water and unsaturated rock doesn't 8 really jump across air gaps, the first question you think 9 of, I believe, is okay, so, if water is what's going to 10 degrade these canisters, how does it get there? I think wo 11 are utill trying to figure that out. There are ideas in 12 here. But, to me, they are mostly kind of ad hoc. They 13 just assume that water gets there some way.

14 This model that's used here for the source term, 15 has it in both terms for advoction and diffusion for the way 16 wastes can get out of the canister, and move out through the 17 engineered barrier system. It has to deal with degradation 18 mechanisms, which are not modeled mechanistically in this 19 model. All of those things are put in pretty much in terms 20 of distribution functions. This is one of those quirks of 21 this TSPA that I would say is attributable to the three 22 months' time that we had to do it. The source term is 23 defective, in that it includes very little of the bottom-24 level modeling. It's an almost automatic roll-up of 25 processes that are going to require a great deal of study.

O ANN RlLEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

214 1 One of the things that you will see that we are 2 hoping to do in the next iteration is to improve that source 3 term quite a bit. But, as it stands now, it does appear to 4 be useful, especially for sensitivity study kinds of things.

5 Because when you represent those variables, and describe the 6 processes that are pictured on here, as well as some that 7 aren't pictured, you can play with those and try to decide 8 which of them are important.

9 If we had an hour, we would explain what all those 10 parameters were, but we won't. We'll move on to one of the 11 more interesting and basic facts about the current state of 12 performance assessment of Yucca Mountain.

13 [ Slide.)

14 MR. BINGHAM: The fact is that no one is really 15 sure how water moves through the mountain. We adopted two 16 separate models, two alternative conceptual models, for 17 purposes of trying to see at 1 cast something about_what the 18 range of possibilities is-for releases at Yucca Mountain.

l 19 The first of these is the composite porosity model. It has 20 been described as a classical kind of mocol. It allows 21 water to flow either in the matrix, or in the fractures, or 22 in both, depending on-the flux level. It tries to adopt a 23 mathematical formulation for flow that's appropriate for 24 unsaturated media. And, the distribution of water in the 25 mountain might look something like this cartoon that's put O ^"" aev a Associ^Tes' d-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

l 215 1 up here. I tend to say the water is sort of ubiquitous in 2 this. It's going through all the mountain more or less 3 evenly, except when the flux values are high enough that 4 there is a good deal of flow in major fractures.

5 Now, at the other extreme of what might happen, is 6 that maybe the water flows almost entirely in these 7 fractures that are around the Yucca Mountain. Maybe there 8 la very little flow in the matrix. Maybe all the 9 significant water is in those fractures. So, we've made 10 statements, or we've made them towards the end of our 11 presentations, and they seem to startle people, so I want to 12 make it now -- that this looks to us as though it is kind of 13 in members of a npectrum -- that, at least you can say this

() 14 15 Weeps model, just go ahead and assume the water is in the fractures. Let's see what the consequences of making that 16 assumption are.

17 Now, yesterday, Michael went through quite a bit 18 of explanation for just how this works, how you decide how 19 many fractures must be flowing, how much water they're able 20 to carry, what the relationship is between fracture, 21 aperture, and length and depth, and width and height. We 22 won't do that today. I'll just point out that there are 23 these two models, that we are treating them in ways that we 24 think the NRC staff would like to see for alternative 25 conceptual models. That is, we are carrying them separately

'h ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 i Washington, D. C. 20006 '

(202) 293-3950

l 216 1 through the entire set of calculations for which they are

>O 2 appropriate, to see what the consequences of each is.

3 Just to see if we could get Norm excited, we did 4 show a graph yesterday where we weighted those two models 5 and combined them, but he didn't bat an eyelid, so I guess 6 we're going okay.

7 With no further explanation of what's in these 8 models, I'll just show a picture. The principal 9 radionuclides that get out, and these percentages are in 10 terms of EPA limit, not curies, are these. With both of 11 them, it is only the rapidly moving radionuclides that get 12 out. The ones that are essentially unretarded, are, at 13 best, only weakly retarded. I suppose that's what you would 14 think.

15 I think it is probably worth repeating -- and I 16 hope that nobody will ask a question -- that the 17 distribution use for the flux that drives all of these 18 releases is really quite wide. Holly showed it to you and 19 ducked down behind the podium a little while ago. So, we 20 have included, in releases like these, some allowance for 21 future weather conditions, or even present weather 22 conditions, if somebody eventually shows us that they're 23- there. We have a long way to go with doing this.

24 And one of the questions that I've heard asked or 25 brought up is are there some issues for which there may O ^"" ai'ev a Associ^Tes' 'ta.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

217 1 never be complete empirical validation? I guess this would 2 be my candidate for them, if there is one. I don't know how 3 clever the hydrologists are going to be in the future, so I 4 certainly wouldn't want to say they will never solve this 5 problem. But, in view of the difficulties that there seem 6 to be with determining flow at a dry place like Yucca 7 Hountain, it may be that conceptual models like this will B have to be carried through for a long time yet, maybe in the 9 license application.

10 The good news, I think, is that, so long as the 11 results still tend to follow the ones we've had in these 12 preliminary ones, it may not make much difference. One of 13 the other things I was delighted to hear norm saying this 14 morning was in his definition of validation, that it has

}

15 something to do with adequacy for the purpose to which the 16 model is being applied. If the purpose for which these 17 models are applied doesn't require that they be extremely 18 accurate, maybe there is a solution to the difficulty that 19 seems possible to me that we may never be able to decide 20 completely between them.

21 MR. APTED: Just a quick question. You've given 22 us the relative percentages, can you give us some idea of 23 their relative size of those two circles, one to another, 24 for the same sort of infiltration rate after 10,000 years?

25 I mean --

O ^"" nev 5 ^ssoci^Tes. 'ta-Coud Repoders 1612 K. Street, N.W.,- Suite 300

. Washington, D. C. 20006 (202) 293-3950

218 1 MR. BINGilAM: That's the CCDP for the two models.

2 Yes.

3 MR. APTED: Okay.  !

4 (Slide.) l l

5 MR. BINGilAM: In fact, here we have a CCDP with 6 the two models on it, plotted separately. You'll see that 7 the blue line here is composite porosity, the green line is 8 Weeps. One of them comes a little closer than the other.

9 Depending on what you think is significant, there may or may 10 not be a significant difference between those two. I think, 11 for the modelers, there is a significant difference. These

17. aren't just numerical instabilities, or something that are 13 produced -- the models really are saying something that's 14 different.

15 Now, there's lots to be done to improve these 16 models to make them better, and when we get to our plans for 17 the future, we will point out what some of those things are.

18 If, in the future, the better CCDF's, the ones that survive 19 peer review, and all kinds of difficult scrutiny tend-to 20 look like this, I would say we are probably reaching towards 21 an answer for that question are there some issue that can 22 never be resolved, say to the complete satisfaction of the 23 scientific community, but may be resolved adequately for 24 purposes of licensing.

25 MR. MOELLER:. Excuse me. Were these two models O ^"" aev a ^SSoc'^Tes' 'id-Court Reporters i 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l

219 1 developed exclusively for your analyses, or did you build 2 upon something that someone else had done in the past?

3 MR. BINGHAM: The composite porosity model, of 4 course, has been used, in one form or another, for a long 5 time.

6 MR. MOELLER: Okay.

7 MR. BINGHAM: At Sandia, in particular, we have 8 had a code that's embodied in it since the early '80s.

9 MR. MOELLER: No , I-meant -- well, even.

10 MR. BINGHAM: The Weeps model is new essentially 11 for this exercise. __

12 MR. MOELLER: Well, are the out-growths of models 13 that hydrologists or someone used 30 years ago, or for 14 something else?

15

~

MR. BINGHAM: I don't know about the 30 years ago.

16 MR. MOELLER: Well, whatever.

17 MR. BINGHAM: The mathematical formulation or.the 18 composite porosity model,. Richard's equation-is, I gather 19 well-known to hydrologists.

20 MR. MOELLER: Okay. That helps.

21 MR. BINGHAM: And the ability to use things-like 22 the Van Gunuchten formulation is'-- those are -- I can't'say 23 whether they're standard or not, but they show up a lot.

24 And there may be hydrologists'who would like to attack'them.

25 So long as'they attack.them instead of me, that's all right.

O ANN RILEY & ASSOCIATES, Ltd.

-Court Reporters 1612 K.- Street, N.W.,- Suite 300 Washington, - D.' O. 20006

-(202) - 293-3950 ~

y

j 220 1 The Weeps model, though, is something now. And, 2 in fact, there is very little to it. It is almost a 3 geometry model, as you saw yesterday. It is pretty easy to 4 do, once somebody has shown you how.

5 (Slide.)

6 MR. BINGHAM Those are the aqueous releases.

7 Let's look at what happens to gas. For gaseous releases we 8 had to use the source terms, of course, that came out of 9 that aqueous kind of analysis. Because the thing that makes 10 gasses come out of the package are the same as the things 11 that make the solids come out. It's the interaction of 12 water with the container. So, those source terms still 13 apply here, except that all that we're talking about for the

(} 14 15 gas flow is carbon.

happens with Carbon-14.

And, in particular, we want to see what 16 We use Ben Ross' model, which is accepted -- the 17 same one that the NRC staff talked about today, except that 18 they said they tinkered with it a little. Wouldn't they?

19 The idea is that gas and air, in fact, gasses of 20 all kinds can float fairly freely through Yucca Moontain, 21 maybe more freely than people thought a little while ago, 22 and that when they're driven by the heat from a repository, 23 the flow lines may be substantially different from the way 24 they are when there is no heat there.

25 This is a picture that shows flow lines under an O ^"" ai'ev a Associates, 'ta.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 1 (202) 293-3950

l 221 1 elevated repository temperature. There are couple of 2 interesting things about it that I probably can point out 3 without violating my rules of not going into too much 4 detail. Maybe the biggest -- the two things are, one, 5 you've got some interesting phenomena, like confection 6 cells. And the other is that, over here, some of the path 7 lines can be very short. If you look at one similar to 8 this, the cold -- or no repository, these lines aren't over 9 there. Those are interesting features that need to be taken 10 care of in the modeling.

11 In trying to get -- of course, we can get a source 12 term for release of Carbon-14, it won't be accepted by 13 everybody, and will be attacked for some time to come, but

(} 14 15 we can get one. Then we need to decide how long does it take for the Carbon-14 to move out of the mountain along 16 path lines like those. Well, Ben Ross has been clever 17 enough to come up with some travel time distributions, that 18 really depend on about two things: The permeability of 19 those layers in the mountain, and the possibility that 20 chemical exchanges will occur as the CO2 from the canisters 21 moves through the CO2-rich environment.

22 Then Ross has been able to do both of these things 23 really, of course, in a pretty preliminary fashion. He 24 didn't have a lot more time to do this than we did to do the 25 whole total system analysis. But, he was able to come up O ^"" ni'av a Associ^Tes' 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

. - - ~. - . ..

222 I with a graph like this for probabilities of travel times.

' 2 You will notice that over here in the cold repository, 3 everything is longer, and that's for two reasons. One is 4 that the gas moves out faster when it is driven by the heat, 5 and the other is that there are path lines that don't exist 6 for the cold repository. Over here at the 360 one you can 7 get some pretty short ones.

8 To combine all these, as a function of time, is a 9 task that is cor,r'.icated, but not impossible. Michael 10 talked yesterday about how to do that.

11 MR. MOELLER: Excuse me. 60 degrees makes all of 12 that difference? I realize that's Kelvin. But, it's just 13 60 degrees between? Okay. I see it.

I 14 MR. BINGHAM: This is what the CCDP looks like.

15 There must still be two of them. We still have to carry the-16 two conceptual models for water flow, because they affect 17 the source term differently. This time we get the bigger 18 releases from the composite porosity model instead of the 19 '4eeps.

. And the answer is very obvious. In the composite 20 porosity model, every waste canister, or most every waste 21 canister gets wet, and in the Weeps Model only a few do.

22 The few that do get more water than the ones that don't --

l 23 or, the few that do get more water than they would if they 24 were only ubiquitous flow everywhere. But, this is -- when 25 those facts are balanced out, this is what our CCDF looks Q ANN RILEY & ASSOCIATES, Ltd.

Court F.eporters 1612 K. Street N.W., Suite - 300 Washir$ tor , D. C. 20006 (2Uj 293-3950 l

223-1 like.

O 2 And there ar several things, I suppose to point 3 out about-it. The first is that this is the only one of our 4 CCDF's that seems to be threatening to violate the standard.

5 a, it's appropriate to say how real is that? Well, we 6 think it is not very real at all because of the lack of 7 development of things like good source terms, good container 8 degradation models. We made a lot of assumptions in coming 9 up with this, and we invariably tried to make them 10 conservative. We tried, as has already been explained, to 11 include our uncertainty. So, I suppose we would have a hard 12 time thinking that the CCDP could lie any farther to the 13 r.'gnt and up than this one does.

14 But, we regard this as a place to point out to the 15 site characterization program, to the DOE, to anybody else 16 who wants to listen, that here there are some parameters 17 that we may be able to nail down. Perhaps a site 18 characterization program can really help with this. Perhaps 19 it can show why our conservative numbers are too 20 conservative.

21 I cite this as an example of guidance to the site Tu characterization program. And I will say a little more 23 about that when we get to the end and wrap up.

24 [ Slide.)

25 MR ., BINGHAM: We did human intrusion. We otarted O ^"" ai'ev a ^ssoci^Tes. 'ta-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

m - _. . ._ _ _ . . . . _ . _ _ . _ _ _ . - . _ _ _ . . - _ _ _ . _ _ . _ _ _ . . - . ._-

224 1 off,-.as always, with.a tree for human intrusion. I didn't:

l-2-

show you'the tree for aqueous releases. It's humongous.-

3 It's so enormous that'I-didn't even~want to show a little' -

4 piece of it. But, we did-start with the tree.to try to ,

5 figure out what to model, because that's the systematic way 6 to do it. And the SNL stands for systematic.

7 (Laughter.)

8 MR. BINGHAM: For this we picked the path-that's 9 shown here in dark line. Two things are happening. A' drill-10 intercepts waste, pulls it up to the-surface in the drilling 11 mud, or it knocks the waste down into one of the underlying-12 aquifers so it can be carried away through aqueous means.

13 Here's a picture of a drill stem coming down, 14 coming close to a waste package, banging-up some rods.

15 T1.ere are some -- our drilling fluid.here is rather an odd 16 color, but it does carry spent fuel back up to the top. And 17 that, of course, is a direct release to the accessible 18 environment.

-19 How this.was done has always been. easier for:me to 20: understand by looking at this particular graph.-- I think 21 when Raleigh_Barnard made this thing._he couldn't understand 22~ why I kept insisting on-looking at this instead of the CCDF.

23 _To me this is a little; easier to-understand than the.CCDF.

-24 The way the calculation-is done is to throw darts-at a 25 repository.

ANN .RILEY _& ASSOCIATES, Ltd.

_h

'=

Court : Reporters-1612 K. Street, N.W.,- Suite 300 Washington,- D. C. 20006 (202) 293-3950

225 1 We assume that the probability that the human

2 intrusion is going to occur is one. But, we follow the 3 EPA's guidance for how many holes in 10,000 years are likely 4 to happen there. Three per square kilometer, that works out 5 to be 17 for our site. So, 17 times during 10,000 years, 6 the computer throws a dart, and decides did that thing hit a 7 waste canister? Did it come close to one, or did it just 8 miss everything? If it hits a waste canister, then a 9 variable fraction of the waste is pulled up. If it misses -

10 - if it doesn't hit a waste canister, but comes close to 11 one, there is an algorithm, a routine in the code to decide, 12 through the use of a diffusion equation, how much waste it 13 may pull up, assuming that the package is busted up enough

() 14 15 that some waste is diffused out into the rock.

everything.

It may miss 16 What this shows out of -- this is 20,000 17 realizations I believe -- that, of those 20,000 18 realizations, we got a few direct hits, and that when they 19 happen, it is possible to pull up an entire EPA limit on a 20 very small number of the cases.

21 Most of the time what happens is a near miss. And 22 there we're winding up with a thousandth or a millionth or 23 so of an EPA sum. And, in surprisingly few cases, I guess 24 it's not surprising, be-suse the mathematics of this are 25 really very simple. You don't get anything at all. This is (O

ANN RILEY & ASSOCIATES, t (d.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

226 1 really just geometry. It's what is the difference betwaen

O 2 the -- if you examine the size of the drill and the size of 3 canisters, and how they are soon to be uniformly put 4 throughout the area, it is easy -- well, I don't want to say 5 it's easy, it wasn't easy was it, Ron? In principle, this 6 can be done without advancing the state of the art of 7 mathematics let's say, or computer science.

8 MR. MOELLER: To help me understand. If we have 9 the estimate of 17 holes, and you've looked at 20,000, or-10 whatever you say here, to get this frequency distribution -

11 -

12 MR. BINGHAM: It's 20,000 sets of 10,000 years.

13 MR. MOELLER: Oh.

14 MR. BINGHAM: So, it's 20,000 times at which --

(

15 MR. MOELLER: Oh, okay.

16 MR. BINGHAM: -- in each of which 17 holes are --

17 or 17 darts are thrown.

18 MR. MOELLIR: Even so, then, if I actually drf' 19 17 in 10,000 years, then what is the probability of those 20 17, one is going to be in the direct hit?

21 MR. BINGHAM: There is a young man there who is 22 just itching to answer that question. Raleigh, would you?

23 MR. BARNARD: For each realization, what you do is 24 you throw the dice and find out whether you have a direct 25 hit, a near miss, or a complete miss. And what you are O ^"" ai'ev a Associatesta-COud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l 227 1 trying to do over 10,000 years is accumulate over 17 0 2 instances of this to find out what the grand total release 3 of the surface is. It turns out, for 17 holes, the binomial 4 theorem will tell you that the most likely event is to have 5 not a hit -- zero direct hits. Okay? Sometimes you get 6 one, sometimes you get two, sometimes you get three. Three 7 is pretty much the most that you would get.

8 Okay. Now, if you also include in there this 9 factor which Felton said that you bring up between a little 10 bit and all of the waste nackage, and you also include a 11 factor which says sometimes you drill at 300 years, and 12 sometimes you drill at 9,999 years, the sum total of that is 13 that the accumulative release for any one trial, the worst 14 trial that you_see there, comes out to be about one release 15 -- an EPA limit of one. And this may have occurred because 16 of two hits, but each of which only brought up half a waste 17 package, or one hit way late, and so forth like that.

18 Anyway, it's enough rolls of the dice that it's just 19 statistics at this point, iO MR. MOELIER: Thank you.

21 MR. BARNARD: Does that take care of it?

22 MR. MOELLER: Yes. Thank you.

23 [ Slide.]

24 MR. BINGHAM: I especially like having this in the 25 package. Because, if you really would-like to know what was Q ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

228 73 -

1 done, you can look at this for a little while, and all of a

's 2 sudden, things start becoming clearer. 'At least it seemed 3 that way to me. When I looked at the CCDF, I scratched my 4 head for a long time.

5 But, this is the CCDF. You can make it just from 6 that graph, by adding up columns-and working from right to 7 left. And you can see the kind of two pieces in here, one 8 that has to do with direct hits, and then the one that comes 9 -- is almost all of the probability here of the near misses.

10 This one doesn't come close either, in spite of 11 having what I believe we think, again, are a number of 12 pretty conservative assumptions. Frit dealing with this one, 13 we tried to deal with experts in drilling technology, people 14 who have actually been drill fields, and know what sizes

(

15 bits are and what happens. They tended to be a little 16 skeptical of this, because most of them seemed to feel that 17 a drill bit isn't going to go through a waste container 18 anyway, unless it's in awfully bad shape. It's probably 19 going to hit the steel thing, and veer off to one side. We-20 didn't allow for that to happen. We just assumed that 21 people drill so nicely, that they can go down into even 22 uncorroded containers.

23 This is another example of a place where I think 24 it's possible to start coming to some conclusions. Now, 25 once again, we haven't modeled all of the sequences that you Q ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

229 e- 1 can think of that would be like human intrusion. But, this-

\- 2 seems to me to come pretty close to following the kind of 3 guidance that the EPA has dealing for dealing with human 4 intrusion. And I believe it's explicit enough that those 5 who would like to understand it, explore it, and decide 6 whether it's adequate, can do so, because the numbers are 7 there in front of them, and the numbers were speak for 8 themselves, with a little bit of explanation from Raleigh 9 Barnard, of cource.

10 One question that seems to be showing up a lot is-11 what good are these TSPA? What use is being made of them?

12 We use some of these data in talking to the National 13 Academy's panel, who were charged with looking at the DOE's

( 14 reply to the EPA's questions about the new standard, or the 15 old standard, or whatever its status is now. These results 16 were handy to have, because they could be pointed to as 17 examples of calculations -- of some technical analysis, that 18 has some bearing on some of the questions that the EPA was 19 interested in.

20 We did a sensitivity study on this. It turns out 21 the best way-to make those curves move to the right here is 22 just to increase the number of holes. So, if you do 17 in 23 10,000 years, you get a curve that looks like this; if you 24 do 170, you get one here. You can get closer, and closer 25 and closer. And, in fact, yesterday, Raleigh showed one-ANN RILEY & ASSOCIATES, Ltd.

Q'~ Coud Repoders t

1612 K. Street, N.W.,. Suite 300 Washington, D. C. 20006 (202) 293-3950

230

, -) 1 where he was finally able -- 600 holes or so -- able to get

'~

2 over there into that forbidden region. And this, or course, 3 tells us something that we all knew; that human intrusion is 4 a squirrely thing to deal with, because, unless you decide, 5 at the beginning, that we are just going to make some 6 assumptions about the future, it will be possible to get any 7 CCDP you want by making different assumptions about the 8 future.

9 It think this sort of thing made us feel pleased 10 that the EPA has included in their guidance some statements 11 to the effect that, since the future is essentially 12 unknowable, we just have to make some assumptions, and get 13 on with it.

() 14 15

[ Slide.]

MR. BINGHAM: To combine all these components --

16 Michael talked at some length yesterday about how you go 17 about combining them. And there are some interesting 18 technical questions there about what's the right way to put 19 them all together.

20 It might even be worth pointing out that the 21 process that the SCP suggests as the way to put everything 22 together is different from this. It doesn't involve 23 conditional CCDF. But, since we are a very long way yet 24 from following the procedure that this SCP has, it is a 25 little irrelevant at the time.

f O) ANN RILEY & ASSOCIATES, Ltd.

Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 l (202) 293-3950

231 1 Here are the separate ones, and then there's a 1 '

2 total shown for them. The total is blue. It hardly appears 3 on here, except down here, because it is about the same as 4 the gaseous one. The gaseous releases are so much biggr4r 5 than the others that they just dominate the rest of them 6 here.

7 We have to do this separately for the two 8 conceptual models, of course. This is the composite -

9 porosity model, the one with the big Weeps.

10 MR. HINZE: Felton, in your tree for the 11 volcanism, you had one branch going down to the indirect 12 effect upon the subsurface water. Is that included in that 13 CCDP that we're looking at?

[ } 14 MR. BINGHAM: Not, it's not. The only thing that 15 includes is the sequence that was dark in that. And, I 16 think we realize that one of the things we need to do is to 17 look at that other sequence. Could it possibly have greater 18 effects?

19 MR. HINZE: So, one plus one could equal three?

20 MR. BINGHAM: On CCDF that may be possible.

21 [ Slide.)

22 MR. BINGHAM: Let's see. We still have the Weeps 23 Model to show. And this one has, of course, the lower 24 overall releases associated with it. But that's what the.

25 thing looks like. There's food for discussion in all of O ^"" ai'ev * ^SSoc'^Tes' 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l

232 1 these curves. And getting things out on the street is O' 2 intended to provoke that kind of discussion.

3 We thought we ought to talk a little about what 4 conclusions we were able to reach. I talked about them 5 yesterday. I had intended to spend three minutes talking on 6 them. After about 25, I decided people may have had enough.

7 Since a lot of you heard them yesterday, maybe we'll be able 8 to run through them quickly this time.

9 In this study, and in fact that phrase, "in this 10 study," probably should come in front of every one of these.

11 We're quite aware that it's someone else's study, making 12 different assumptions, and using different data. This is 13 likely to produce something different. But, in this study, 14 at any rate, the gaseous releases are the most important.

(}

15 This is something we would like to tell the site 16 characterization program, 17 The composite porosity model predicts higher 18 releases than the Weeps Model.

19 One of the questions that has been asked is how 20 much have you learned? Have you learned anything-you didn't 21 know before? I think this is probably something we didn't 22 know before. Until somebody really tried to take a-hard 23 look at what it means to have water flowing exclusively in ,

24 fractures, and how many fractures that' involves and so on, 25 it was hard to answer this question. We now have an answer.

] ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612. K. Street, N.W,, Suite 300 Washington, D. C. 20006 (202) 293-3950

233 1 I think this is something we learned from this study -- that O 2 human intrusion, we can't ignore it; but it does appear to 3 give releases that are lower that threaten to. violate the 4 standard less than just what we call nominal conditions. We 5 call them nominal conditions, by the way, because all the 6 words like undisturbed and unanticipated, and expected, and, 7 I'm probably going on now, seem to have been preempted to 8 mean something else. So, we have tried to think of 9 something that nobody had used yet, l

10 I predict that what we will discover is somebody 11 will start using it in a different way eventually, and tell

! 12 us we were wrong.

13 This is possibly an inflammatory statement. But, 14 I would like to make it to be provocative. If they are not 15 inconsequential, let's find out why, and let's find out 16 which of the parameters that makes them look inconsequential 17 is wrong.

18 The first one on this slide is probably something 19 most everybody knew. I think people have realized for a 20 long time that the way water flows through Yucca Mountain is 21 probably the most important thing about it. How much is 22 there, and how does it get through? So, I can't say that 23 that's a new conclusion, but we have a new basis for making 24 it -- some more data.

25 The next one is, I suppose sounds like motherhood.

Q ANN RILEY & ASSOCIATES, Ltd.

Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

234.

q 1 It's probably true of almost everything. But,fwe do=think 2 that_there are some particular places -- and in the 3 presentations yesterday,.we tried to point out where-we were 4 -- where we really have lots more conservatisms than we 5 really ought to.. Even at*the present state of knowledge, we-6 can be less conservative than we have been so far.

7 We need to model more sequences certainly. 'The 8 sequence for indirect effects of volcanism is one that needs L

9 to be thought of. There are probably some more in human 10 intrusion. There are lots of them. Those trees are 11 enormous. There are lots more things to be modeled before 12 we can say we have done a complete job.

13 And then, finally, the last one here is one that 14 the NRC staff has picked up on too. Including thermal. i: 15 effects is going to be important. 16 MR. HINZE: If I may ask? Your first bullet is 17 about gaseous releases. And_the NRC has also evaluated 18 this. Is there anything that.we can learn by comparing the 19 results from what you have done with-what NRC has done? 20 MR. BINGHAM: I'm sure-that there are thing that. 21 could be learned from it. Because -- of course, the first 22 time we saw their stuff was the day before yesterday, and I 23 can't say that I understand everything-they have-done. But, 24 -a technical exchange on that-topic, something where the

          .25=  modelers can-actually talk to each other, I am confident; Q                          ANN RILEY & ASSOCIATES, Ltd.

Coud Reporters - 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

                                                                                       .s

235 1 will reveal some insights. I am sure they thought of some 2 stuff we didn't. Maybe less than you expect, since we both 3 used Ben Ross' model. But, to have Ben Ross present, as 4 well, might be a good way of finding out something new. 5 MR. HINZE: The barometric variations on the 6 surface were not taken into account in either of the gaseous 7 models, right? 8 MR. BINGHAM: That's right. There are some 9 offorts underway to look at that a little more. Because 10 it's clear that they can have an effect. The mountain does 11 breath. 12 MR. EISENBERG: Let me ask you a question I didn't 13 ask yesterday, since we have you up there. O V 14 MR. MOELLER: Microphone. 15 MR. EISENBERG: Sorry. 16 MR. BINGHAM: Sorry, Norm, we can't take your 17 question, you didn't use the microphone. 18 MR. EISENBERG: Regarding your first and fifth 19 conclusions. 20 MR. BINGHAM: I can find that one. 21 MR. EISENBERG: That's the first. But, then you 22 say that your results are most sensitive to the choice of 23 the conceptual model for flow, which I would -- as I 24 understand it, was decoupled from your gaseous analysis. 25 MR. BINGHAM: No. Because the conceptual model O ^"" ni'ev a Associateste. Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

236 1 for flow determines the source term for the rest of the -- tO 2 MR. EISENBERG: Oh. Okay. 3 MR. BINGHAM: Yes. Okay. 4 MR. EISENBERG: And the fifth one wa, subsumed in 5 that question. 6 MR. BINGHAM: It's at this point that I would like 7 to depart from the viewgraphs I have got. I would like you 8 to take up your hymn books that were presented by the WRC 9 staff this morning, and turn to hymn number 37. I thought - 10 - we are hearing so many questions about are you learning - 11 - well, not a lot of questions -- but, I think the Committee 12 is interested in are you learning the same things, or are 13 you learning different things. () 14 15 As Norm went through these, I got to thinking, you know, I could have used his viewgraphs probably. There are 16 some lesscns that he has learned, and he puts on here as a 17 result of this IPA that we probably learned some time ago. 18 Although, we haven't been doing very many cycles of total 19 system performance assessment, this being the first one. We 20 have been doing other things called performance assessment, 21 and we have learned some of those. And I realized that I 22 didn't put them on our lessons learned, because they didn't 23 come from this TSPA. I thought it might be worth while just 24 to run through these ones-that norm had, and I'll give you, 25 off the top of my head, what I thought of them, as Norm was O ^"" ni'ev a ^ssociATes' 'ta-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l 237 g 1 saying.

 '~    2                                          The first one, on page 37, has to do with computer 3  needs.          And he talks about the need to simplify models, to 4  make them run faster, and to make them do better in a lot of 5  ways. I wish there was some way to be 150 percent 6 supportive of Norm's statement.                                                                                                                                                                                                                                                                                That's something we found 7  out quite some time ago.                                                                                                                                                                                                   We would like to make the codes 8 run faster, and there probably are ways to make them run 9 faster.                        Most of them rely on 1970's or so technology.                                                                                                                                                                                                                                                                                                                                     They 10  don't -- they are not up to date.                                                                                                                                                                                                                                                                                                 Paul Eslinger was 11  pointing out to me at lunch today that even things like the 12   FORTRAN code display that people use routinely for graphics, 13  has been updated about four times in, I think he said, the
 . 14   last 18 years.                                                                                        And they start from scratch.                                                                                                                                                                                                                   They don't 15   just tinker with it.                                                                                                                                                         We have got a lot of codes that 16   probably would work a lot better.

17 The prospects for massively paralleled computing 3 seem, to my ignorant state, to be tailor-made for things 19 like working with many different flow paths. These are 20 things we thought of that we would like to do. I am sad to 21 report that every year I take this out of the budget. There 22 isn't money to do it. 23 One of the reasons, maybe -- I hope I don't sound 24 like I'm trying to undermine the DOE here, but, from a 25 contractor's point of view, there may be a lot of outcry (] ANN RILEY & ASSOCIATES, Ltd. Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 , (202) 293-3950 l.

l 1 238  ! 1 from folks like the NRC that we need to upgrade these codes 2 would help make it better. I wouldn't expect you to do_that 3 without thinking carefully, of course, about what the 4 payoffs would be for doing it. But, to me, the payoff is 5 not so much increased speed, but, in the ability to make the 6 system models contain more mechanistic models, more of the 7 detailed modeling than we can do now. If we can make the 8 codes run faster and better, we will have less problem with-9 abstraction. Because abstraction, and I might as well admit 10 it now, is really the thing that keeps me awake at night. 11 The site characterizers will do their thing, and they will 12 do the best they can to get us some data. We, in 13 performance assessment, have got to do the best we-can for 14 that abstraction process. And I don't know of a way -- I (} 15 cannot draw you a flow diagram that will convince you that 16 we are able to do that well. There have been attempts to do 17 such a thing, and they wind up with a lot of boxes and 18 arrows, and an arrow going from every box to every other 19 box. And what have you learned when it is all over? 20 I think this is an intellectual enterprise, 21 intimately connected with details of modeling, details of-22 data, and with our increasing understanding of what the term 23 " reasonable assurance" means. 24 MR. BINGHAM: Well, that's page 37. And, page 38 25 is about the same thing. l O ANN RILEY & ASSOCIATES, Ltd. Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

239 g- -- 1 Page 39, transportability of computer codes. This

  '-                                                      One of the ways of 2  also is something that we have found.

3 dealing with this a little bit is through-the process called 4 benchmarking. We used to, in the old days when money was 5 plentiful, conduct benchmarking exercises. We would have 6 people using different codes, from different institutions, 7 some of them using the same codes, but on different 8 machines, or at different institutions, construct a problem 9 that would be difficult for all the codes, and then try to 10 solve it together, and then get together to decide why they 11 haven't been able to solve it. Because they invariably 12 found out that they got different answers, 13 They found a number of very interesting things (} 14 15 when that happened. Beginning with -- the first troubles you find are that people are using different value for the 16 gravitational constant and some things like that. You iron 17 those out of the way, then you find -- or maybe you find, 18 ~ first, that just using standard _ program language, as Norm 19 says here, isn't enough to ensure the codes are going to 20 work the same on different machines. A lot of insights to 21 be gathered that way.

       .22             When I see the codes that the NRC is using I 23   think, goodness, wouldn't it be great if we could have a 24   benchmarking exercise with them and with us?        Wouldn't that-25   help us to understand why our results are different?        And I O#'                        ANN RILEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

                                                                                      'l 240
          'l am sure they will be.       There have been difficulties in doing g-)g
 \_        2 that in the past, some of which are the sort of things 3 country boys don't know about, like problems of agencies 4 working together.       But, the benchmarking seems to be an 5 awfully good thing to do, if you can forc s the code writers 6 to get together and be frank about what their limitations 7 are. We think that can be done.

8 Once again, I am sad to say, that every year I 9 take that out of the budget, because when the axes start 10 falling, that's nv: something anybody is screaming to do. I 11 would like to propose it as a way of helping to decide the 12 answer to a question like one that at least subliminally 13 seems to be around here, which is are these codes giving --

       \ 14   are you getting the same answers, or aren't you?

![/

 \_

15 MR. EISENBERG: Felton, can I interrupt? May I? 16 MR. POMEROY: Yes. Please. Go ahead. I am going 17 to interrupt after you do. 18 MR. EISENBERG: I think, in-all fairness, we ought 19 to point out that the DOE and NRC have participated in a 20 number of internatior.al benchmarking-studies, starting with 21 Intercoin, going through hydrocoin, and now into Intervalve, 22 which do just that. And there maybe a couple that I have 23 left out that also look at benchmarking exercises. 24 MR. BINGHAM: That's certainly right. It was 25 completely wrong of me to give the impression, if I did, O ^"" ai'av a ^SSociATes' 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l l

241' g- 1 that there's been nothing. But, I think, the kinds of

~

2 benchmarking that I saw working in the early '80s are a  ; i 3 little easier to get something out of than the international 4 efforts, which require plane trips, and flights across the 5 Atlantic, and 45-page reports to the. agency on what you did j 6 while you were there. 7 The ability to do something close, where people 8 can talk to each other and grab each other by the neck, if 9 they dare -- grab each other by the neckties -- is a very 10 valuable thing to have, when what you are dealing with are 11 details of codes, rather than just the big answers that come 12 out. 13 Paul, you were going to jerk me around about () 14 15 something? MR. POMEROY: Let me just ask a question. I am a 16 little unclear. Because of the earlier inbreeding here, 17 with regard to Sandia-developed modules, could you comment 18 just on how different the modular construction, and the -- I 19 realize you're using different codes -- but how different, 20 in reality, are the NRC and your codes? 21 MR. BINGHAM: I think they're-very different. 22 MR. POMEROY: I guess I would ask you to speculate 23 that if I use an EPRI code, for example, would I, for 24 instance, would I be likely to get a wildly different-25 result? I realize we don't have any data on that. h ANN RlLEY & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

242 1 MR. BINGHAM: I guess that you would probably get 2 pretty different results. And the reason I say that is 3 remembering those old early '80s benchmarking exercises, 4 where codes that were thought to be very much alike.at first 5 gave different answers. After people understood why they 6 were getting different answers -- and some, to their 7 embarrassment, found mistakes in their codes -- they began 8 to get much the same answers. It was very valuable in 9 developing the code. So -- but, I would guess that you 10 would find differences. And I do not think that the codes 11 that Sandia developed for the NRC are likely to be very much 12 like the codes that Sandia developed for the DOE for this 13 project. We were forbidden to speak to each other or have 14 much to do with each other then, and probably to the

  .( )

15 advantage of independence of the agencies, but probably to 16 the disadvantages of those who could have learned something 17 from having them. 18 MR. POMEROY: Thank you. Mick? 19 MR. APTED: Mick Apted. I echo your support for 20 benchmarking. Although, having gone through it myself on a 21 limited basis, and now doing it on a larger basis,.I think 22 the most that could be gotten out of it is to try to 23 anticipate the differences by studying the code before 24 jumping into the calculations. I think, if you jump in too 25 soon, you spend all of your time saying oh, you know, that l t ANN RILEY & ASSOCIATES, Ltd. l{ l Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 L L l

243 1 would have been simple to eliminate, and then you are back 9 2 to doing the calculation yet again, I think. 3 Plus, there is some advantage to saying oh, 4 because this is a one-dimensional model, versus a two-5 dimensional model, or spherical versus cylindrical geometry, 6 or something, you feel better if you make some predictions 7 on how you think it is going to be different. And so, then, 8 once you do the benchmarking, you get a little bit of ah- - 9 hah, at least out of it and to anticipate. Otherwise, it 10 just is sort of always stumbling around correcting one 11 thing, coming back and correcting another thing. It is 12 better, I think, if the people think conceptually first. 13 That is the hard part. And doing the calculations is 14 usually a matter of minutes to hours, depending upon the 15 computer. 16 MR. BINGHAM: I would agree. In fact, I would 17 expand on it by saying the way I think the benchmarking 18 should be done is to begin with a very long, hard problem 19 definition phase, where the users of the codes, who are 20 aware of their code's difficulties, work on the problem, 21 being frank enough to admit that their codes may not work on 22 some of them. That's always kind of a difficulty. And the 23 second point that I believe you are raising, is this should 24 be done with fairly mature codes. 25 Just as soon as you have written out 7,500 lines g ANN RlLEY & ASSOCIATES, Ltd. Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

244 fm- 1 of FORTRAN, we heard about, probably you are not ready _to

 'Q) 2    jump into a benchmarking exercise right then.      But, after 3    you've learned some things about it, then you probably can 4    make the benchmarking exercise.      In fact, in the days when 5    we used to put this into the budget, we would tend to have a 6    benchmarking exercise last about three years, and the first        i 7    year of it was getting those people to agree on a good 8    problem definition. We thought that was every bit as 9    important as running the code.      And we learned that the hard 10    way from the early 1980's exercise.

11 Carbon-14. I guess I have got to agree with this. 12 The Part C, on page 39, that it may be a problem. Yes. It 13 looks like it might be here. We need some more site () 14 15 characterization data, and probably some better modeling to deal with this red flag, that we think this system 16 assessment has raised. 17 Point four -- that there doesn't appear to be much 18 for seismic disturbances. That's quite true. There is a 19 reason that it hasn't been done. And the reason is not 20 well-documented, and perhaps not even very well thought out. 21 In view of the worries about what happens if a volcano comes i 22 through? What happens if the packages get real wet? What 23 happens if people drill into it? It seems a little of l 24 secondary importance to me. What happens if the packages 25 get shaken a little? And there are simplistic arguments you l l Q ANN RILEY & ASSOCIATES, Ltd. Coud Repoders l 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

245 1 _1- can make. If the expected ground motion is this many 2 centimeters, and the air gap is this many centimeters, it 3 doesn't look like much is going to happen. Yesterday we saw 4 some suggestions that we ought to say that a canister swings 5 like a pendulum. And maybe that's a way to force more 6 action in the seismic disturbance business. But, that is 7 something that we have on our list of things to do next. 8 The remarks about technical coordination and 9 project management I suppose aren't technical, but they sure 10 are true, and they sure do interfere with the technical 11 work. Coordinating things may be more difficult than 12 predicting. 13 On page 41, the system code that is complex is 14 certainly true. There may be a feeling -- I think I used to 15 have it -- that the system codes should probably be kind of 16 system. Well, they're not simple. They're hard. And the 17 one that the SCP points to, which hasn't been even started 18 yet -- I used to take it out of the budget every year, I 19 don't even put it in anymore. It's going to be a formidable 20 effort. It will require a lot of the kinds of support staff 21 that Norm says will be required. 22 We don't, at Sandia. We have a real system code. 23 This_TSPA was done by Mike Wilson -- calls it the total 24 system analyzer, the TSA. By giving something initials, of 25 course, you make it seem real I suppose. But, it largely O ^"" ni'ev & Associates. 'ta. Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 2000S (202) 293-3950

246 r~ 1 exists in his head, and in the collection of subroutines

  ~#  2 that he can run off of his work station.

3 Scenario methodology. That the old SNL stuff 4 needed some modifications, we would -- we're not surprised 5 at that at all. In fact, that is one of the reasons that 6 what the SCP suggests is a good thing to do is different 7 from that. 'That some modifications are necessary is not 8 surprising. We thought it would be. 9 Point number two that he makes about the 10 effectiveness of the WIPP approach is probably a more 11 technical thing than to be discussed here now. I feel that 12 it is already-subsumed in my plea earlier today. So, let's 13 not discard any of the CCDP methods yet. Let's try to keep (} 14 15 them all on the table, and find out which ones are most useful. l 16 On page 43, the -- is the flow part acceptable? 17 That, of course, is the question we have. And we really 18 feel strongly about number two. We certainly agree with 19 that -- that thermal effects, not just in the early phase, 20 but throughout, could be important, and they can be l 21 important on the aqueous pathways that we've treated 22 isothermally so far. 23 MR. APTED: Could you expand on that? You say L 24 important. Important adversely? Positively? I mean, how 25 do you expect? O ^"" aev * ^SSoci^Tes. 'id-Coud Repoders 1612 K. . Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l 247 g-) 1 MR. BINGHAM: Well, I've learned never to say when V 2 something is adverse and when it isn't in a complicated 3 system like this. I wouldn't be surprised to find that they 4 help. 5 There is, of course, the difficulty that has been 6 mentioned several times, of putting a lot of effort into 7 this. Now you worry a little, since we don't know whether 8 we're going to have vertical emplacement, or drift 9 emplacement, or a big, huge, cask or a bunch of little what. 10 But, there is, we think, a good deal that could be done, 11 particularly about the source term, which ever one it is. 12 The thermal effects on it are important. .And we would like 13 to know a lot more about them possibly through testing, f~h 14 possibly through more use of the models at the bottom of the J 15 pyramid. 16 MR. APTED: But, won't it, in a sense -- hopefully 17 a decision on hot or cold would be based on performance 18 assessment, wouldn't it? or would it be based on some other 19 -- 20 MR. BINGHAM: Yes. I was wondering if someone 21 would ask that question. You are asking, are you not, 22 shouldn't the decision about which kind of cask and which 23 kind of emplacement be based on performance assessment? 24 MR. APTED: As a leading factor, at least? 25 MR. BINGHAM: And the answer to that is, of (] ANN RILEY & ASSOCIATES, Ltd. Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006  ! (202) 293-3950

248

                 -1     course, yes.                        But, I don't think the answer is quite as 2  obvious as that.                        And the reason I don't think so has to do 3  with the state of the art of total system performance 4  assessment, as opposed to subsystem performance assessment.

5 If you were to come to us now and say, okay, model 6 it with a big cask, and model it with a little cask, we 7 would scratch our heads a while and then say, I guess we had 8 better widen some of our distributions and see what happens, - 9 which might yield some insights, but probably doesn't tell 10 you what you would really like to know. The more profitable 11 kinds of modeling for that question, at least right now, 12 seem to me anyway -- somebody here may agree with me -- huh, 13 Bill's nodding his head, maybe he does -- to be done closer (} 14 15 to the bottom of the pyramid, with detailed models that really look at what's happening within the first few meters, 16 or the first 10 or 20 meters around the repository. 17 There are some big decisions to be made there. 18 And, of course, the decision analysis is going to rope in 19 all the things like, _if we have a big cask, do we have to 20 put a railroad all the way downstairs? From the viewpoint ' 21 of performance assessment, quasi-technical.

                                                                                                                                     ?

22 MR. STEINDLER: Do you believe that the inadequate 23 state of the art of the total systems analysis is so poor as 24 not to be able to make some commentary on potential benefits 25 for prolonged cooling prior to emplacement? O ^"" ai' ev * ^ S S o c'^T E S. 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l

249 1 MR. BINGHAM: So long as you include that word 2 potential, I think it can be done. Sensitivity studies, 3 using distributions that display our ignorance, where we 4 don't know what values to use, may or may not yield some 5 insights, depending on how they come out. 6 Is that an answer you don't -- you look as though 7 you like it very much. 8 MR. STEINDLER: It's got so many qualifiers that I H 9 am not -- I've got to mull it over for a while. 10 [ Laughter.) 11 MR. BINGHAM: Well, I must have succeeded. 12 Yes, I think there are some insights that can be 13 gained. I think the more valuable insights probably come 14 from some more detailed modeling. Because those simplified 15 ones that are total system analysis we'll have to do really 16 ought to be based on some good, fundamental details modeling 17 that underlies it. I would believe them a lot better if 18 they were. 19 MR. BOAK: Much of analysis that_has been 20 addressed to questions of varying thermal loads has been 21 done by Lawrence Livermore. It actually has been funded 22 under the near-field environment studies and it tends, at 23 least at this point, to strongly suggest that, rather_than 24 their being benefits from prolonged cooling, there are 25 substantial detriments to the long cooling. That may well Q ANN RILEY & ASSOCIATES, Ltd. Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

250 be a >:onsequence of a technical preference. But, the main, (q/ 1 2 the critical point, I think, is valid geochemically that, in 3 fact, even at low temperatures, that there is nothing 4 mystical or magical about the boiling point of water at 5 Yucca Mountain -- that substantial alterations to the system , 6 will occur, down to temperatures I think as low as 50 7 degrees centlgrade, and tat the kind of coolino you would 8 have to do to fuel -- or tha areas you would have to occupy 9 would be subacantially different from anything we envision 10 now, without obvious benefit. 11 That, at least, is the conclusion drawn by the 12 group that has been looking at that issue at this point. We 13 would like to do a great deal more analysis to determine 14 that that choice is, in fact, the appropriate one. And our 15 next iteration is intended to address some of those 16 questions. 17 MR. STEINDLER: With some reservations, I might 18 agree with you, were it not for the fact that that same 19 group has ignored the up to recently compelling nature of 20 the EPA requirement on Carbon-14. Now, all of that may 21 change, and they may in fact turn out to have lucked out. 22 But, up to relatively recently, that was a risk of which 23 they elected to ignore. 24 MR. BOAK: Not entirely. Because so much of their

    .5  case was based on the presumption that it would be far O                        ^"" ai'av a ^Ssoci^Tes. 'ta-Court Reporters 1612 K. Seeet, N.W., Suite 300 Washir.gton, D. C. 20006 (202) 293-3950                                                                          l
                          . . _              - - - - - - _ - - - - _ _ - - _ _ - - _ _ _ - - - - - - - - - - - - - - - - - - - -                         ---------------------A

251 1 easier to validate our expectations of the engineered

2 barrier system performance, and make the assertion -- their l

3 contentien was that we would be butter able to defend the } 4 assertion that the vaste package lifetime is, in fact, 5 substantially larger than it is. Right now, we have no data 6 to support anything but imaginary distributions for 7 container lifetime. 8 MR. STEINDLER: Based on a rather abysmal 9 ignorance of mechanism corrosion problems. And I think 10 there is some merit to be gained if you are forced into a 11 mechanistic approach. And I noticed that you also made the 12 comment that it would be nice to be able to drive toward 13 more and more comprehensive mechanistic models. Were it not (} 14 15 for the late hour, we could talk some time about whether or not that is really a sensible drive, considering that you 16 would like to have a repository within the lifetime of the 17 youngest person in this room. 18 ( Laughter. ) 19 MR. BINGHAM: We probably shouldn't go through 20 every one of these lessons learned. I hate to praise him so 21 hiahly by agreeing with everything he has got here. Let me 22 point out, on page 44, number five, where he is Laying that 23 oven going down to two dimensions taic- a lot of work- One 24 of the slidos ve had yesterday, which I don't intend to show 25 again today, was suggesting that, at one dimenulon, things O ^"" ai'av a Associates. 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

_= 252

     ,  1 are still too computationally intensive, at least with the

( 2 slow-running codes that we have today. 3 on page 45, I believe we would agree with all of 4 these things here. This really is having to do with the 5 problems we vero -- the issue we were just ta? king about, a 6 mechanistic modeling -- of understanding, in more detail, 7 what is happening around the waste package. I believe it is 8 our bias, and it looks as though it may be the NRC staff's 9 as well, that we need some complex modeling here. There is 10 a little bit of differerce, I think, here, between the way 11 they're thinking of things, and the things we are thinking. 12 I believe they are thinkina of a system code tnac is 13 probably including more meet.anistic modeling than we have () 14 15 been thinking of it. semantics, mostly. But, I be.lieve it is a question of If the mechar', tic modeling is done, 16 whether it is separately done, and then incorporated into 17 the system code, or built into it, may be a question of what 18 the words mean. 19 I would really like to p:3nt out, on page 46, the 20 number 4-A, how water gets into the waste package, if at 21 all. Our problems, even with Carbon-14, which you point out 22 the group has kind of steadfastly ignored for the time-23 being, sort of go away if the water really can't get at the 24 waste package. 25 I believe I will stop at this point, because we I O ANN RILEY & ASSOCIATES, Ltd. Coud Repoders ( 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

253 1 pretty much agree with the rest of those things that Norm

 ,O          2      has to say as well.                  But, I thought it might ce worth giving 3      you a little feeling for what we thought of, or at least 4      what I thought of when Norm was up saying these things.                                    I 5      believe they are finding pretty much the same problems we 6      are, and that our differences with them are going to turn 7      out to be differences in details and, as always, in 8      semantics.

9 [ Slide.) 10 MR. BINGHAM: I wanted to show kind of an ending 11 slide. I am going to skip the ones that propose what we 12 want to do this coming year, and just leave them there for 13 you to look at. But, as far as meeting our purposes, the 14 first chapter of the TSPA has really four major purposes ( 15 and, I thought to wind up, I should say something about 16 them. The first one was to develop the abstraction process. 17 I think we made some progress there. I think the human 18 intrusion and the volcanism both have some -- have showed us 19 that we're learning some things about it. 20 Do we know how to do it yet? Can I stand up in 21 front of you and say yes, we know how to do-that now? No. 22 I don't think that's so. And I don't think we're going to 23 learn how to do it better until we do some more of the 24 detailed modeling. I guess you are hearing this pretty PS strong pitch that we ought to do more of this detailed, t O ANN RlLEY & ASSOCIATES, Ltd. Coud Repoders i 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

254 1 mechanistic modeling. The total system analysis are all 2 very 1. ice, and they do yield insights, and they yield 3 important insights, but there are some very important 4 insights yet to be gained, I think, from the detailed 5 modeling. 6 Assembling data combinations. And our purpose -- 7 we pointed out that it is not clear that wo are going to be 8 able to do that. When you have complex combinations of all 9 sorts of data, some probabilistic, some something also, 10 people have expressed some skepticism that you would be ablo 11 to do it. I think we are finding that it does seem to be 12 possible to do it. It seems to be possible to do it in ways 13 that don't provoke a lot of controversy. There will always (} 14 15 be some disagrooments over details. by what this analysis said on the insights I could gain from But, I felt encouraged 16 it by seeing what we had to go through to make these 17 combinations. 18 Guidance for sito characterization. Th.$a is a 19 particularly important point that you have been going after, 20 And this is where I have been promising I would say a little 21 more. I have tried to point out a few places where we think 22 we have some specific recommendations to make. 23 You are probably interestod also in hearing from 24 ti.e DOE a process that ensures that those recommendations go 25 into somebody, and are acted upon. That is a process that I O ANN RlLEY & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C, 20CX)6 (202) 293-3950

255 1 expect the DOE could give you a presentation on. There are -~) 2 a number of documents that deal with it, and some of those 3 have been sent to the NRC staff, in response to a comment 4 into their SCA. 5 Instead of citing those, I would like to point out 6 something that a couple of other people alluded to, and 7 that's this. The most effective way I think of 8 communicating this is not through those processes. It's not 9 through things you draw up on a diagram with boxes and 10 arrows. The way we are finding -- the most effective way of 11 communicating this is by talking to the people who are doing 12 the characterization. We have learned things from dealing 13 with people in site characterization, let's say Alan Flint. ( 14 I believe Alan Flint has learned some things from dealing

    )

15 with us -- from working with us. From -- even the 16 possibility of some joint papers, if we can work out the 17 inter-agency between the USGS and the DOE National 18 Laboratory. The best way to make that happen, although it 19 is not a way that can be drawn up and will satisfy a process 20 diagram, is through the hearts and minds of the people who 21 are doing it. And, to me, the second most encouraging thing 22 about the TSPA is this thing has worked as a vehicle for 23 that. It has helped that interaction occur. And it's no 24 secret that that's an interaction that a group of warring 20 tribes, like the Yucca Mountain Project, in the old days, O ^"" ni'ev a Associates' 'ta-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l 256 1 was not easy to bring about. O 2 The last one is guidance for future analyses. My 3 answer to that was going to me my next four viewgraphs that 4 are in your package that I am going to skip. We have some 5 very detailed ideas of what we would like to do in the 6 coming months. And those are put there for your inspection. 7 MR. POMEROY: I see no questions, so I'm going to 8 move on. 9 Our next presenter is Paul Eslinger, from pacific 10 Northwest Laboratories. 11 MR. ESLINGER: It's hard to follow, this late in 12 the day, such an eloquent gentleman as Felton Bingham, so I 13 will try not to put you to sloop. 14 A couple of things to notice here. (} Other than 15 having my name spelled wrong, and my laboratory spelled 16 wrong, we did okay on these. 17 The only thing I want to acknowledge is, like 18 Felton's group, there was a group of nine people that worked 19 on our analyses, some more than others. I was involved in 20 that work. But, I certainly didn't do near all the work. 21 And there is lots of credit to be given to those -- those 22 people listed on the title of the document that is in 23 publication, and I have got a few advance copies to pass 24 around to people who are interested. 25 Like Felton, based on some teleconferences, we Q ANN RlLEY & ASSOCIATES, Ltd. Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l

257 1 guessed at what we wanted to present today, and guessed 2 quite long, so I have a whole page full of notes here that 3 have come up during the day. I may talk from those as much 4 or more than the slides. 5 Let me talk some about differences between what we 6 did and what Sandia did from a philosophic basis. And we 7 will get down to showing some -- a few results as well. 8 As I start this out though, I do want to make one 9 comment relative to a discussion earlier in the day about 10 climate change and infiltration rates. I suppose it could 11 be said about several things that were kicked around. That , 12 is, doe has been funding a study, at least since 1985, to 13 look at global climate change in the presence of ice sheet, 14 carbon dioxide forcing and other things. Those are global 15 stuff that is going to feed into a regional model which will 16 hopefully give us some sort of range, from a modeling 17 perspective, on infiltration rates, or at least 18 precipitation rates, to be translated into infiltration-19 rates, that go along with expert solicitation of the same l 20 thing. So, I know about that one, because my organization 21 is involved in some of that work. 22 PNL used, as I say here, much more detailed 23 process models in several areas: 2-D, instead of 1-D 24- models. When we went to gas-based transport we used a 2-D 25 transient model to look at the gas phase, where Ben Ross's , Q ANN RILEY & ASSOCIATES, Ltd. Court Reporters l 1612 K. Street, N.W., Suite 300

                                    - Washington, D. C. 20006 (202) 293-3950

258 I g-s 1 is a steady-state at a series of repository temperatures. 2 One of the consequences, back to the abstraction 3 sort of principle, was when we got down to running 4 iterations, we didn't complete near as many iterations as 5 they did, in terms of getting CCDF's. And the other basic 6 difference in philosophy here is we went on and did some 7 individual dose calculations, based on releases from some of 8 the total system scenarios. They weren't based on an 9 attempt to go out and do the dose scenarios correctly, it 10 was take these releases as a first cut, and see what happens 11 from that. 12 (Slide.) 13 MR. ESLINGER: And these slides may be out of 14 order, in terms of what I am going to talk about, versus [} 15 what I thought I was going to talk about. 16 The processes -- Holly has talked about these in 17 detail. And those of you who have been here in the last 18 couple of days have seen them. We went to the same sequence 19 of calculations with the exception. We added some stuff 20 from tectonic activities. And I will finish that discussion 21 on that one right quick, in that it was a water table rise, 22 and that the range if infiltration rates we got we didn't 23 get any transport to the water table. So, that one sort of 24 short-circuited and stopped at that stage. There are a lot 25 of other ways you could look at tectonic activities, and the O ANN RILEY & ASSOCIATES, Ltd.

 \J                                Coud Reponers 1612 K. Street, N.W., Suite 300-Washington,- D. O, 20006 (202) 293-3950 t

259 1 one way we looked at it, but it had not consequence 2 different from the base case. 3 (Slide.] 4 MR. ESLINGER: Let me slip a slide in here from 5 yesterday to point out one of the differences. This one has 6 to do with basalted intrusion. I have a conditional CCDF. 7 This one starts at one, if you would go back to -- on the 8 probability, if you would go back to zero releases. 9 Sandia took an approach where they used dyke 10 formation that was based -- the dyke properties were based 11 on an elicitation of experts, in terms of width, breadth, 12 those sorts of ideas. We had done a different approach, in 13 that we took a simplified model out of the literature, and () 14 15 took ranges of parameters from a dyke formation model that were world-wide ranges, and used a very similar sort of 16 process than they did, in terms of generating dyke 17 properties. Well, the real interesting thing about it is 18 our CCDF -- they had two of them -- lies right between the 19 two of theirs_with very little difference. 20 Another interesting thing on this was car human 21 intrusion on this analysis. When we ran transport 22 scenarios, we started with the 2-D model, and you get a 23 concentration field changing in time for nuclides within the 24 mountain. When we modeled at volcanic intrusion, to get the 25 amount of contaminant, again -- we looked at a fraction of Q ANN RILEY & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300

                               -Washington, D. C. 20006 (202) 293-3950

260 1 rock entrain, and those sorts of things -- we would put a 2 location somewhere in our 2-D modeling domain at c time -- 3 random time -- and then take a look at what the 4 concentration was at that point. So, we nested these a 5 little bit different than the way they nested them. Quite a 6 bit different, in terms of the modeling scenario; but you 7 come up vith about the same result. So, that was an 8 interesting difference in approach, but very similar in 9 result. In fact, if you look at the results, other than one 10 or two things I'll point out, they are very similar. It's 11 just that one or two things change the shape of the whole 12 curve when you get down to the end. 13 MR. STEINDLER: What conclusion do you draw from 14 the fact that they are very similar? Are they both wrong, 15 or are they both right? 16 MR. ESLINGER: The conclusion I would draw is kind 17 of this -- you can look at it two ways, and you get very 18 similar results, in terms of performance measure. You are 19 starting to get some sort of robustness in the idea of 20 conceptual model uncertainty, which is a real concern in the 21 whole modeling approach, can somebody pick a different 22 conceptual model and get something that is very different? 23 Whether they are both right or wrong, I think we 24 will have to wait till the exchange, at the very minimum, 25 that was mentioned earlier in the day, about the - t O ^"" ai'ev $ ^ssoci^Tes. 'id-Court Reporters l 1612 K. Street, N.W., Suite 300 Washington, D C. 20006 1 (202) 293-3950 l i I

1 263 1 volcanologists getting together and hashing over details and 2 talking about this. But, it does kind of approach the 3 conceptual model uncertainty, which is something that we 4 wrestle with a lot. 5 It is also something that the PSAC group, from 6 OECD is wrestling with, as an organization. And, currently, 7 I happen to be a member of that group. And it will be 8 interesting to see what comes out of their next exercise, 9 which is aimed at looking at conceptual model uncertainty 10 for a similar sort of approach for several national 11 programs. 12 I was going to put up a human intrusion CCDP, but, 13 by the time you put them on, my look so much like Raleigh's, 14 that they came out essentially the same. 15 Let me circle back to the scurce term model a 16 little bit, and make a point relative to a question somebody 17 asked earlier. The source term rodel that we ran -- Mick 18 should know a lot about it, it's the arrest code -- is a 19 mass transport away from waste package, in the very near-20 field. And the source term is based on some solubility 21 considerations and water flow considerations. So, it's a 22 much more detailed analytic model than the response function 23 approach that Sandia used. l 24 One thing to point out here is we looked at both a 25 glass dissolution model and a spent fuel model. And there O ^"" aev $ ^SSoc'^TES' 'id-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l

I 262 1 are plans in the -- work plans in place that we will 2 continue to look at the glass modeling as well as the spent 3 fuel modeling. So, some of that work is going on. 4 When we looked at combining scenarios, and this 5 was t&lked about earlier -- I'll put another slide up from 6 yesterday. 7 [ Slide.] 8 MR. ESLINGER: There have been as many ways of 9 combing scenarios as we have done exercises so far. And it 10 may continue to be that way for a while. I am not sure -- 11 my major background and formal training is in mathematics 12 and statistics, and I have no difficulty with any of the 13 approaches, from a statistical perspective. And I'm not () 14 15 sure we have to reach convergence on any one approach, as long as you have the right statistical properties embedded 16 under it. 17 The kind of approach we.took in adding things 18 togather were that human intrusion and volcanism were nested 19 within base case runs that we were doing. So, if you look 20 at the future, and you go down different paths, you have 21 base case, some sort of human intrusion, drilling, which 22 exhumes the soil column, which is the 8-0 here, or drilling 23 which pulls-the waste container up, or puts it in the 24 saturated zone -- waste in the saturated zone. And those 25 are all nested within each other. But, you end up finding a O ^"" ai'ev * ^SSoci^TES' d-Coud Reponers 1612 K. Street, N.W., Suite 300 l Washington, D. C. 20006 (202) 293-3950

263 1 way to get groups of results which have the properties that 0 i l 2 you can add them statistically to get the overall CCDP. ) 1 3 And, at least from all the stuff I've seen, we may i 4 have to be a little creative, but you can find a way through 5 it to do that sort of approach. 6 (Slide.] [ 7 MR. ESLINGER: Let me show now a combined CCDF for 8 the total system. We will talk about some of the bumps on 9 it, and how our bumps then differ from the kind of bumpt 10 that Sandia showed. If you start from the right most tail, 11 way down at the bottom, this is the contribution from the 12 volcanism condition on CCDF I showed, scaled with an 13 occurrence probability, which is down the order of 10 to the 14 minus four, or two times 10 to the minus four. We ended up

15. using this same sort of occurrence-probability, based on 16 Bruce Crow's work that I think Sandia did.

17 The consequences can leave you above the. EPA 18 limit. But, if we agree on occurrence probability, then-19 you're not above the' limit in such a region of the space to 20 be af regulatory concern. But, again, I am not going-to re- , 21 do all of the caveat. slides we have seen all" day. But, l- 22. thone caveats all apply to this. 23 _If you look at the-next hump,-going _up on this -

                                                                                       ~

24 shde from the right, have human. intrusion where'you 25 intercept one or more-waste containers-in.a '.Satticular Q ANN RILEY & ASSOCIATES, Ltd. Court - Reporters-1612 K. Street, N.W., Suite 300 Washington, D :C. 20006 , (202) 293-3950- - u

             - - . . - , - .    ~ . , . . - -        ,            , - . - . . - - , , - _ . - _ _ , - - , , - . - - . . -

264 1 model. Instead of drilling just 17 holes, we took a 2 poissant process, with an expected value of 17 as the 3 determinant number of holes, which then corresponds, in a 4 statistical sense, to exponential waiting times between 5 drilling events and did random drilling times, with a mean 6 value of the 17 holes in the repository. You como up with 7 the same sort of approach, and the same sort of result that 8 Sandia did. 9 One point to note here, if you are comparing 10 viewgraphs, is that particular portion -- I believe, in 11 Sandia's approach, they looked at 41 nuclides, and-we looked 12 at 10 nuclides. The 10 nuclides we were looking at, at 13 least in reasonable timeframes, has an EPA ratio of about 14 .1, if you would exhume the whole waste container and look 15 at its contents. Where, if you throw in all nuclides, you 16 are up closer to one. So, this is like a loth of the 17 inventory in most of this analysis. So, you need to keep 18 that in mind. 19 If you go back now to the next thing, where I have 20 a sharp break point on this, now we start talking about gas 21 phase analyc.is. The difference is significant, in terms of 22 releases, between our results and Sandia's results, come 23 mostly from gas phase, that's one point. There's another 24 point as well. We used a 2 D transient model to calculate - 25 thermal effects, and then derive gas flow in the mountain. O ^"" ni'ev

• Associ^Tes' 'id-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

                             -   2----_--_------_-_-__.__-_-_.-----.-__.__.---__.____--____.-__.____.--__--_-----_-a

265 1 1 We get very little releases out. Only at a -- the bounding 2 cases of no recharge do we get anything out. 3 The result of this seems to be based on a couple 4 of ideas. One is, when you heat the repository region, at 5 least in the model we ran, you get a redistribution of water 6 within the mountain that reaches very near saturation some 1 7 distance above the dry-out zone. That very high saturation 8 acts as an effective gas flow barrier. So, in very 9 simplified terms, that's one of the reasons that we don't 10 get very much gas flow out. 11 And the other reason is our permoabilities, even 12 though we have to look at processes and other things. Our 13 permeability value we chose happened to have been a couple () 14 15 or three orders of magnitude different than the one Ben Ross used in the results he did, and much lower. So, you would 16 expect less transport. There's lots of room here for 17 sensitivity studies, and lots of things to play with there. 18 But, those were the results that we got. 19 When you go on to the left most tall, the little 20 blip, and that is human intrusion, where you brought up a 21 contaminated soil column, but you didn't get a waste 22 container. And it is down on the 10 to the minus seven, 10 23 to the minuc eight EPA, which is the same place I think it 24 was showing up on Felton's slide. 25 MR. STEINDLER: Did you, by chance, make comments O ^"" ai'ev a Assoc'ATes' 'id-Coud Repoders ! 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

266 1 about how one could test your assumption on the block of gas 2 flow going to what appears apparently to be a drastic 3 increase in saturation? 4 MR. ESLINGER: I think, if I remember right, some 5 of the G-tunnel heater experiments showed some of the same 6 sort of things around the very small-scale heater 7 experiments. So, that's one way that you could go about it. 8 I think the other thing that's kind of interesting has been 9 in Alan Flint's studies of saturation profiles in the charde 10 base, above the repository above the Topopah Springs where, 11 at least in every hole they have looked at so far, they find 12 a layer which has some clay, which has a saturation of very 13 near one. If there are holes in it, if that is not aerial () 14 15 continuous across the whole region, then you've got maybe more potential for gas flow. If it does exist across the 16 whole repository region, the significance of gas flow is -- 17 you could find a barrier there to flow that would be very 18 interesting to continue to look at. 19 One thing also. We found out -- we already knew 20 it. We were using much more detailed models than Sandia, 21 and they were busy doing their abstractions. And we thought 22 we had our detailed models ready to run on this data set. 23 In fact, we had got some results out a couple of months 24 before the deadline. And then we found out that we had one 25 of the data values wrong relative to what we were going to. O ^"" ai'ev * ^Ssoci^Tes' 'te-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

267 1 And we changed it and submitted the code again, and we ran 2 into convergence problems, and only completed some of the 3 cases we wanted to run. 4 The cases we didn't complete were those of very 5 high infiltration rates. So, our results, as presented -- 6 the documents presented here are limited to the low 7 infiltration, below a half a millimeter a year, and did not 8 include the 1020, up to 39 distribution that Sandia used. 9 As such, mass transport was dominated by diffusion, so you 10 don't get large invective flow, even though there is an 11 invective terminal model, it is small enough to not be of 12 too much significance over 10,000 years. 13 Somebody asked about the independence of codes of {} 14 15 Sandia. I guess the one thing I will point out here are these codes, even though we have been in performance 16 assessment for -- I can remember for quite a number of 17 years, they are all independent of the Sandia codes. So, 18 they have a different pedigree. In some cases, you get very 19 similar results. And the prices -- we don't -- we believe 20 there are as much data-driven reasons, as code reasons, why 21 they are different. 22 Let me switch directions here a little bit, and 23 focus more towards some dose calculations we did. 24 (Slide.) 25 MR. ESLINGER: And I couldn't resist putting up l Q ANN RILEY & ASSOCIATES, Ltd. Coud Repoders i 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950 1

268 1 this slide that talks about regulations covering performance O 2 assessment, from a total system perspective. Starting with 3 1985, where dose had a very limited part of the regulatory 4 process, and was limited to the first thousand years, in 5 terms of ground water protection. It went on to working 6 draft four of 40 CFR 191 in early 1992, after we basically 7 finished our calculations, but before we finished the write-8 up on it, Where it looked like dose could play a much bigger 9 part. And, of course, here in october, Congress changes its 10 mind again, and have opened up the regulatory process. But, 11 it looks like individual dose will have some regulatory 12 significance as we go on. 13 The results I am going to present, in terms of 14 dose, at least to allude to them, come from ICRP 26, as 15 modified on through ICRP 30 and 40 kind of models. We 16 looked at several scenarios. I am just going to talk about 17 a couple of them today. But, we have both base case sort of 18 analysis, and human intrusion analysis for doses. The human 19 intrusion, or the disruptive conditions weren't required on 20 the very earlier versions of 40 CFR 191, what has showed up 21 in the population dose portion of the working draft -- at 22 least it was out early this year. Who knows where they are 23 going to end up now. 24 In terms of scenarios, I have a 70-year exposure, 25 70-year lifetime, except_for some driller scenarios, where ANN RILEY & ASSOCIATES, Ltd. Q Coud Repoders 1612 K, Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

269 1 the drill is only there 40 hours, but then, you can see this 2 commitment for an additional 50 years. 3 MR. MoELLER: What did you mean there that it's 4 not the maximum exposed individual? 5 MR. ESLINGER: The results we're reporting here 6 are the releases from a total systems model that was looking 7 at cumulative release and not concentration at a point in 8 space. So, we took those. You needed concentration for the 9 dosa value. There were individual doses, but there was no 10 effort to go out and find the worst sucker who got in 11 trouble, in terms of a maximum exposed individual. So, 12 there are individual doses, but this is a caveat. It's not 13 -- it's exploratory sort of information. () 14 15 (Slide.) MR. ESLINGER: People have done several paper 16 studies on Carbon-14 in the gas phase. What we did was took 17 some runs that we had done, and Sandia did. Back hore is 18 their composite porosity model, and the Weeps Model. 19 The source term that happened to be the maximum 20 over the time period we modeled, in tetus of curies por 21 year, and convert that to dose. To do that, you have to 22 make some assumptions. We were looking at cumulative flux 23 out of the mountain. You need a concentration. Well, we 24 took a mixing volume, 10 meter depth, in terms of mixing the 25 carbon in the air, 3.3 meters per second wind speed, because-O ANN RILEY & ASSOCIATES, Ltd. Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 ) (202) 293-3950

270 1 these are long-term calculations. That happens to be the 2 average wind speed at Yucca Mountain. And then we took a 3 repository size, so that it is all limited to this as it 4 moves across. Doing that, we get some doses which are down 5 in the loth of the millirem for the maximum run from TOSPAC. 6 We also made an assumption that the guy living up 7 there some how got enough water to grow a garden, has some 8 of his own vegetables and fruits, and some of the carbon got 9 into that food chain. In fact, for those doses, 90 percent 10 of it was in the ingestion pathway, and only around 10 11 percent of it, or less than 10 percent was from inhalation, 12 you know, hanging on his hammock between two_ sage brush 13 plants, breathing in contaminated air. 14 MR. MoELLER: And now why are your doses so much 15 lower than the other CCDF's we've seen, or that would be 16 Implied by the CCDF? 17 MR. ESLINGER: Some of the stuff I've seen is -- 18 you assume basically the inventory of the repository gets 19 out, in terms of carbon. And that's -- you build a 20 concentration off of that. We took predicted releases of 21 carbon to the ground surface, and did calculations based on 22 that. So, I think that's the major difference.

23 MR. MOELIER

Okay. And these doses are to the 24 bone marrow? So, are you saying maximum organ? So, the 25 effective dose would be a fraction of this?

O ^"" ai'ev a Associatesta. Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

271 1 MR. ESLINGER: I'm reporting effective dose as the 2 dose value. 3 MR. HOELLER: Oh. Okay. 4 MR. ESLINGER: But, I just pointed out the organ. 5 One of the caveats that we've kicked around 6 several times is we don't have a lot of site-specific data 7 we would like. 8 Shortly after we finished these calculations, they 9 ran a pump test in the upper saturated zone in one of the 10 wells, and it showed that our assumptions on well water 11 withdrawal were off by only a few orders of magnitude. So, 12 some of the exposure scenarios we looked at were drinking 13 water only. We were assuming there was not enough water () 14 supply that was contaminated to do either a garden or a 15 farm. So, I think we were shown wrong on that. One of the 16 test wells produced, in 36 hours, enough water to water a 17 farm for the whole irrigation season. So, we have that' sort 18 of comparison to make. In fact, it was real close, wo 19 assume for the irrigation, for the entire growing season. 20 (Slide.) 21 MR. ESLINGER: Here, just to show a few more 22 representative results. Some of the -- one of the base case 23 runs of the composite model. Sandia went ahead and ran it-24 out in time for a significant period of time to find when 25 the maximum release -- I think we ran -- the case was run C ANN .RILEY & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

I l I 272 1 for a hundred thousand years. The maximum release -- now, 2 if you take time out of the equation, and just say, well, 3 when does the maximum release occur? I said 53,000 years. 4 And the exposure scenario was drinking water, and technetium 5 and iodine were the dominant nuclidos. Neptunium had not 6 arrived because of its retardation value, even at this time 7 period. 8 one of the problems we has was we woro looking at 9 cumulative flux, with the flow tube approach, and now we had 10 a concentration. So, we made some assumptions on the 11 aquifer, and based on the ground water velocities and 12 porosities they were using, to get a dilution factor. And, 13 again, these dilution factors need to be revisited based on 14 those pump tests, because they could be quite a bit off. At (} 15 the same time, we think that the range of doses we have got 16 here is somewhere around what you would get if you look at 17 that in better detail, 18 Millirem a year run from real small, for drinking 19 water only, up to a couple of millirem a year on some of the 20 other -- in sort of the upper bound cf our range. 21 And the farm scenario here ends up you dilute the 22 amount of contaminant a lot, even though you have added 23 exposure path ways. 24 MR. POMEROY: I'm sorry to do this. But, can you. 25 finish up in a few minutes? O ^"" ni'Ev $ ^SSoci^Tes. 'td-Coud Repoders 1612 K. Street, N.W., Suite 300 l Washington, D. C. 20006 l (202) 293 3950

273 1 [ Slide.) O 2 HR. ESLINGER: Sure. No problem. 3 What did we really do? Well, we went through an 4 exorcino that turned out to be more painful than we thought 5 it would bo, even after going through PACE, and it turned 6 out more painful than wo thought it would be. 7 Really, what we did in we wont through an exercino 8 that showod that, if you are creativo onough, you can start 9 to croato the modelo you nood, you can link them together to 10 got the total system stuff. And wo unod some preliminary 11 data. 12 Sometimos -- I have got the bottom bullet on here, 13 demonstrated computer codon would run. It's not a forogono 14 conclusion you can always do that in the timeframe that you 15 wantod. And we came closo. 16 I had shortcomings and lennons learned that I 17 think have boon said throo times today already, Norm said 18 it, and Folton road from Norm'a viewgraphs. 19 I want to mako one philosophic point on the end of 20 thin analysis. How will we got the real answetr Well, one 21 of them 10 you've got to got a regulatory convergence. We 22 will just kind of leave that as a given. But, I think ono 23 thing we have to do when we talk about the real answer is, I 24 think koop the right philosophic position. We can never say 25 wo are going to predict, with precisonosa, the future of the Q ANN RlLEY & ASSOCIATES, Ltd. Court Reporters 1612 K. Street, N.W., Suite 300 l Washington, D. C. 20006 (202) 293-3950 1 l

274 1 real repository. What we are trying to do is give enough 2 range of information that seems plausible to make a-3 reasonable scientific judgment. You know, if we get 4 crossways between those two philosophies, we will end up 5 never completing the analysis. So, that's sort of a 6 philosophy position. 7 I think I hit most of my notes here, in terms of 8 our differences. Any questions? Mick? 9 MR. APTED: Mick Apted. I was interested to see 10 that you had taken your calculations out so long. You have Hil 50,000 years, and probably assuming, longer.

                                 -                                                           As we get to          ,

12 those kind of time scales, you had 10 nuclides. I guess I 13 am concerned, when I look at people who do dose based 14 calculations overseas, they have a completely different set [} 15 of key nuclides at that time period, a lot of the grow-ins, 16 the pro-techtenium, the radium. So, any thoughts on 17 including those, in terms of grow-in and'a more complete set 18 of nuclides? 19 MR. ESLINGER: Well, the stuff we did -- first of 20 all, we modeled chain decay so_we could model the grow-in. 21 _But, yes, if -- we picked the set of 10 nuclides, which had i 22 at least-one chain that could grow in if it got there. Both 23 the transport and'the. dose models for the stuff we did 24- allowed that;to happen. Ifiyou let the stuff: arrive, the 25- dose goes up tremendously, because the exposure,_or the dose-h' ANN RILEY_ & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300 , V' shington, D.~ C. 20006 (202) 293-3950 -l

275 r- 1 commitment from those nuclides is much higher than some of ( 2 these fission products sort of things. 3 So, yes. If you go to very long time periods, a 4 hundred thousand, to a million years, you have a different 5 set of -- 6 MR. POMEROY: Dave? 7 MR. OKRENT: What I am trying to understand is 8 whether the wording of the bill passed by the Congreas and 9 signed by the President, in which it is asked that -- 10 whether a health-based standard, based on doses to 11 individual members of the public from release -- whether 12 that means a farm family, as you have, or just someone who 13 is using a well. There can be quite a big difference. And, () 14 15 I must ccnfess, until I saw your analyses, I had maybe naively been assuming it was still a drinking water kind of 16 standard. I wonder what in fact it is when the ICRP sets a 17 number, or in those European-proposed regulations, where 18 they used a health-based, or a dose-based standard. Is it 19 really a self-containcd farm family, or is it a drinking 20 water limit? Does anyone know? 21 MR. ESLINGER: I'll let somebody else answer that 22 one. 23 I believe it is alll exposure pathways, and not 24 just drinking water. Maybe not with the -- I know the -- 25 Vernon Thompson, of the UK Program -- you know, they're O ^"" ni'ev a Associates. 'id. Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 l (202) 293-3950 l

276

   ,          1      looking at climate ages with pluvial climates, and then 2     drier climates, over around a million years, in terms of 3     time span, with all exposure pathways in it.                           So, at least 4     some of the programs -- in fact, I don't think I know of any 5     that are limited to drinking water only.

6 MR. OKRENT' Well, if you don't grow your own 7 vegetables, it really doesn't -- and I don't, for example - 8 - nor my chickens. It la going to be really drinking water 9 it seems to me. 10 MR. ESLINGER: It could be, unless you grow some 11 produce in the area. Then you have to worry about the 12 person who drinks or eats that produce. So, this was just 13 one way to sort of self-contain that sort of process in -- (} 14 15 from the site -- MR. OKRENT: But, that person may be further away 16 MR. ESLINGER: Could be. 17 MR. OKRENT: -- and not have the same drinking 18 water. And, especially for this particular site, where one 19 anticipates a change in dose with distance, I think. 20 MR. ESLINGER: Yes. The whole question of how you 21 define the biosphere is a very important question relative 22 to the dose. Do we use food parameters that are current 23 day, or future? 24 MR. OKRENT: Well, if I can make one other-25 comment? It seems to me, with the rather considerable O ^"" ai'ev a ^ssoci^Tes' 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

277 1 progress that the NRC and DOE have exhibited today, and in d 2 the last two days in the area of CCDP, we are lacking the 3 results of sufficiently robust analyses to look at the dose 4 question or the health-based standard that Congress has 5 asked about. 6 MR. APTED: Just a point of information for both 7 NRC and DOE. I believe the International Atomic Energy 8 Agency, has this BIO MOBS 2 project going on, that they're 9 trying to look at a number of repository groups and l 10 regulatory groups getting together to try to establish l And, 11 reference biospheres and pathway models to look into. 12 I believe somebody from the NRC has just petitioned our l 13 group to join that. 14 MR. ESLINGER: I know the same sort of thing has l 15 been done on the low-level waste. There's a serious effort 16 in the United States to define a reference biosphere, so you 17 can at least compare sites that make sense. And PNL has 18 been involved in that activity. 19 MR. APTED: The only difference there is that you 20 worry about a different set of nuclides for your low-level 21 waste. 22 MR. ESLINGER: That's correct. They are seeming 23 to reach some sort of convergence on the problem of the 24 reference biosphere though. 25 MR. POMEROY: Thank you, Paul. ANN RlLEY & ASSOCIATES, Ltd.

 -Q                             Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

1 278

1 Jerry, you are the last and the best, as always.

2 [Brief pause.). 3 MR. BOAK: Now, that we're back on schedule, I 4 want to thank llolly, and Felton, and Paul for getting us 5 back to it -- for abbreviating their talks. I am certain 6 there's a lot of interesting material in them that we could 7 have spent a good deal of time on. We will certainly try to 8 make room for some futv:re discussions on that. - 9 (Slide.) 10 MR. BOAK: I want to talk about our directions, 11 with respect to our next iteration of total system 12 performance assessment. But, to do that, I have got to talk 13 a little bit about what we are going to be doing in the 14 fiscal year 1993. Because that constrains and helps define 15 what we expect to do in the way of a next iteration of total 16 system performance assessment. So, once I have shown you 17 some of those priorities, then I can talk about the scope of 18 that 3xt TSPA -- who will be intolved in it, how we are 19 going to approach it, and some hint of a schedule for that 20 next iteration. 21 It is important to note that performance 22 assessment han-a lot of items on its plate for he next-23 fiscal year. We need to provide a number of supporting 24 calculations to ensure that many of the features of the ESF 25 design that are being done in the course of this year, are Q ANN RILEY & _ ASSOCIATES, Ltd. Court Reporters 1612 K. Street, N.W., Suite 300 l Washington, D. C. 20006 , (?O2) 293-3950 l

279 p 1 not in any way going to affect the isolation of waste, they O 2 are not going to interfere with any of the tests that are 3 going to Le done in the ESF or elsewhere. And so a fair 4 amount of our effort this year will be dedicated to 5 providing supporting analysis. Predominantly these are 6 going to be done with the lower level models from tre 7 pyramid. They are not going to be total system performance 8 assessments. Although t % L11 have to at least address 9 what we expect in the way of any alterations of performance 10 that might occur as a consequence of these activities. 11 In addition, we will have to be doing some of that 12 for the surface-based testing program. In fact, quite a 13 number of our activities will be dedicated to making sure O 14 that the tests that are going on on the surface aren't going k.) 15 to interfere with ESF or with each other -- that they're not 16 going to, in any way, impair the capability of the site to 17 isolate waste. An ing tant part of that is, in addition to 18 looking and seeing what. would cause that problem is 19 determining whether there are any controls that are not 20 already captured in our requirements documents that ought to 21 be added to those -- to the controls tnat are applied to 22 those tests. 23 That then, I think, is going to be the primary 24 task this year. Because the program priority is to get that 25 exploratory study facility started, and that has a lot to do O ^"" ni'ev a ^ssoci^Tes. 'ta-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 l (202) 293-3950

i l 280 1 with the need for the program to demonstrate the kinds of 2 progress that certain people, who decide how much money we 3 get out of the nuclear waste fund, consider progress. A lot 4 of other interesting priorities are things that we think are 5 important. For the time being, our most important priority 6 is to convince people who give us the money to do the things 7 we want to do. We are in fact making progress. And we are 8 relatively convinced, from our interaction with them, that 9 getting an ESF started is the most important thing they see 10 as indicating progress towards understanding the performance 11 of this site and progress towards characterizing that site. 12 The next major pricrity item is the one that has 13 come up repeatedly during the de , an that's what are we () 14 15 going to do about the standards / We fully expect to be spending a good deal of time thinking about possible avenues 16 that the recommendations of the NAS could drive the EPA, and 17 ultimately the NRC, in terms of regulating the work we do. 18 Interesting enough, in looking over some of the proposals 19 that our management and operating contractor has made for 20 what we might want to do to prepare ourselves for that, I 21 began to realize that, to some extent, we have to think 22 about what some of those -- some of the things might mean 23 for any repository. Although, the direction is somewhat 24 specific to Yucca Mountain, I don't think we can cocoletely l 25 divorce ourselves. For that reason, one of the things we h ANN RILEY & ASSOCIATES, Ltd.

                                 . Coud Repoders 1612 K. Street, N.W., Suite SG)

Washington, D. C. 20006 (202) 293-3950

281 ,f-- 1 were considering is looking back at some of the studies done ()/ 2 a previous NAS NRC panel, the WISP Panel, to see how our 3 evaluations of some of the scenarios that they did in some 4 of the waste sites that they looked at might be modified, in 5 light of what we have learned in the ensuing 10 years or so. 6 7 So, we expect to see a fair amount of effort 8 dedicated to making sure that the NAS gets the input they 9 think they need from us. And a lot of that has to be 10 anticipatory. We have to think about the kinds of things - 11 - what we need to look at in the way of doses, what we need 12 to look at in the way of understanding the uncertainties 13 that might apply to various kinds of standards. 14 We have looked, in a very preliminary way, at the

   }

15 question of uncertainty with respect to timeframe of 16 regulation. If we go from 10,000 to 100,000 years, what 17 does that mean in terms of understanding how we are going to 18 comply or whether we will comply with any new standard? We 19 are going to have to look again at that. The NAS was 20 relatively unsatisfied with our first attempt to talk to 21 then about that. So, we will certainly be spending some 22 time trying to refine that. We still think the points that 23 were made were valid. We are just going to have to find 24 ways to make them a little clearer. That may entail a fair 25 amount of subsequent analysis and more specific analysis. l]' ' ANN RILEY & ASSOCIATES, Ltd. Coud Reponers 1612- K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

282 1 The third major topic the performance assessment

 -O'  2  is going to have to support this year is the_ question of 3  thermal loading.       What sort of thermal loads ought to be 4  designed into a repository for Yucca Mountain?            It is being 5  driven by a wide variety of programmatic considerations.

6 But the modeling that will go into that I think may well be 7 at a lower level in the pyramid. Again, the most important 8 questions we need to answer may not be answered by the 9 production of CCDF's. It may be that some thermal 10 hydrologic calculations can provide the kind of answers that 11 are needed without having to roll it all the way up into an 12 estimate of the releases, or of the doses involved. We will 13 hopefully be scoping it out so that some of the same studies 14 that we used in creating that will then provide us valuable 15 input in our next iteration of total system performance 16 assessment, so that perhaps, in combination with other 17 things, we can begin to get some additional insights. 18 We have found in the past that total system l 19 performance assessment, in a way, has given us a goal to 20 shoot for. Along the way, we have found out lots of 21 inportant things. We will try not to wait until we get them 22 into a total system performance assessment to communicate 23 -them. When we find something new that we think is going to 24 affect the site, we intend to communicate that immediately, 25 even when we don't have an estimate of how much it might l LQ L ANN RILEY & ASSOCIATES, Ltd. Coud Reponers I 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

283

    -    1  affect the site, and how much it might show up on the total
       2  system CCDP. S.) , the diffusion of insights from performance 3  assessment, and the counter-diffusion of insights from site 4  characterization has gone on without it necessarily 5  occurring at the level of total system perforaance 6  assessments.

7 Any way, I have laid out three major tasks. And 8 then, way down here at the bottom, the tail of the dog, is u 9 the definition of our next iteration of total system 10 performance assessment. I don't think -- we had hoped to be 11 on a schedule going approximately 18 months or so between 12 major total system performance assessments. I don't think 13 we are going to make it this time. And that is constrained, 14 to some extent, by the need to do these other analyses that 15 will not lead directly to total system performance 16 assessments, and some of which may give us relatively 17 limited insight into the top level models, but will give us 18 some important opportunities to exercise some of our lower 19 level models. 20 You will also notice that missing from this, 21 again, missing from this list is, in fact, the exercises 22 that Felton talked about, the benchmarking exercises, the 23 code development exercises. Once-again, those have been, 24 indeed trimmed from the budget. 25 At some point, before we reach the point of O# ANN RlLEY & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300 NVashin@co, D. C. 20006 (202) 293-3950

284 1 sending a license application in, we think it is virtually 2 essential that we have some better, faster, more impressive 3 codes. If nothing else, we think that, if we go into a 4 license application in 2001 with codes that really represent 5 1970s technology in the way of coding, in the way of 6 modeling, we are going to look rather bad. And we have come 7 to recognize, through a variety of exercises in the past, 8 that we see as a really critical part of moving the program - 9 ahead is the building of a scientific consensus. 10 When we have done a series of decision analytic 11 exercises, including the Calico Hills risk benefit analysis, 12 the exploratory studies facility alternatives analysis, the 13 Test Plant Prioritization Task Force. All of these tasi: 14 forces have generally led to some expert based judgements 15 about what we expect in the way of performance from the 16 site. Those were driven by a great deal of effort, and they 17 were backed by a fair amount of arguing back and forth about 18 those judgments. What we find is that performance is not 19 what we expected to learn about by looking at -- by doing 20 site characterization. What we expect to do, by gathering 21 the data in site characterization is to gather the stuff we 22 would like to have in our back pockets, when people start 23 asking us questions. We want to be able to build the 24 scientific consensus, and the regulatory consensus that will 25 enable us to get a license. And those things, in many $ ANN RILEY & ASSOCIATES, Ltd. Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

285 r- 1 cases, at least in terms of our expectation, are far more

b) 2 important than-what we expect to see in the way of changed 3 estimates of the performance of the site.

4 And I guess we have been regularly criticized for 5 that judgment. In fact, it was a total shock to many people 6 in the doe when these decision-analytic exercises came to us l 7 with that result. We were no less surprised and no less ) 1 8 upset than some of the readers of these documents have been.  ; 9 on the other hand, the answer is relatively robust. And we 10 tend to feel that, until some other group of experts has 11 spent as much time and money agonizing over those issues, 12 and come up with their own judgment, ws stand by the 13 judgment of our experts. We rely admittedly on judgment in () 14 15 a great deal of those because we did not have the analyses in place. As I say, until someone else has come up with a 16 judgment that shows substantially different waits for actual 17 performance, as a driving force in site characterization, as 18 opposed to the building of consensus, the acquiring of the 19 data to ensure that your value is robust, we stand.by those. 20 There are a lot o things that we would like to 21 have in that next iteration of total system performance 22 assessment. And the other priorities will drive us to make 23 certain choices about that. One of the choices we might 24 have is to push the deadline out further. I will chow a 25 conceptual schedule later, and it is only a conceptual h ANN RILEY & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

286 1 schedule. I intend to have a few more meetings, a few more 3 2 problem definition meetings to try and lay out what it is we 3 want, and to try and decide whether it is more important for 4 us to get another iteration out, or to get another few 5 features into that next iteration. 6 But some of the things we would like to have -- 7 obviously, thermal effects on fluid flow and transport leads 8 the cast. We know that the repository has important effects 9 on ground water flow and on a variety of other parameters 10 and affects how we model, how we would describe the 11 performance of the site. 12 We have got some new site data. Some of that site 13 data is stuff that we are very excited to get. There were a l 14 variety of priorities, and a variety of inputs that went ( ) 15 into the decision to activate, for example, the neutron 16 drilling program that Alan Flint has well advanced at this 17 point. But, among the inputs that helped to drive that task 18 to the fore, were the inputs from total system performance 19 assessment that said a better understanding of-how water 20 moves from rainfall into the mountain is really critical. 21 And we think Alan Flint has provided us some useful 22 insights. And when those are incorporated, we will see some 23 changes in the way we assess the performance of the site. 24 We would like to refine and augment the scenarios 25 we are evaluating. You have heard a number of pieces of ( l O ^"" ai'ev a Associates. 'ta-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

i 287-1 'l various-portions of the large event tree that we would like

    -O      2l      to addLin here -- things-like looking at the. indirect-3       effects of both volcanism, and tectonic activities, looking.

4 at the direct effects of fault motion, on waste package , 5 environment, and how that might affect source terms and 6 transport. 7 The' extent to which we incorporate new scenarios . 8 may well be limited, again, by those decisions about whether 9 to go ahead and enhance the scope, and delay the actual 10 timing of it. 11 I think we probably won't be doing a great deal of-12 additional'modeling on volcanism in this next iteration. We

         '13        think there are activities that are going to go on over the-14        next year-or so that may help resolve some of-the issues f

15 that we have heard that are outstanding. In the meantime,.

                                                ~
       -16          we think there are more important issues to be addressed.

17 When we get some better information perhaps for the 18 iteration after this, TSPA three, perhaps we will have a 19 more refined. version of' volcanism to puttin there. But, I 20 think, for now, our inclination is to say we looked at what 21' we. considered to be the most risky features of' volcanism, 22 and they do not sure substantial effects. . We'want-to get. +

       '23          more data before we gonback andilook at--it again.-

24 ' We would like to have.some more of the sensitivity

        .25         studies that we have just begun to do from our old data as a

'h ANN RILEY & ASSOCIATES, Ltd. . Coud Repoders. 1612 K.. Street, N.W., Suite 300 Washington, . D. C. 20006 (202) 293-3950

v. - m g

                                                           . . _ , , . . - e  .,-p,#    ,

288 1 guide to the formulation of our next total system 2 performance assessment. We would like to have time to do 3 some sensitivity analyses of the results of our next 4 iteration before we actually put it out. This last one came 5 out with relatively few sensitivity studios complete. And 6 we have done a few sense then that you have new seen 7 presentations on. Before we get our next document out, we 8 would like to have a few of those done. 9 In addition, in this next iteration, we would like 10 to have some opportunity to look at the question of 11 correlation -- correlation between parameters within the 12 site, hydrologic parameters are quite commonly coordinated, 13 porosity, permeability is a fine example. 14 We would also like to look at spatial correlations ( 15 to see whether that has an effect, and what the effect is on 16 site performance. One of the curious things -- during the 17 PACE exercises, we complicated the stratigraphy to the point 18 where I think we had 19 layers. And that reflected a 19 statement about the vertical heterogeneity of the mountain. 20 But, in doing so, we said absolutely nothing about the 21 lateral heterogeneity of the mountain. In essence, we used 22 continuous, totally uniform layers,.which we know to be 23 geologically observed. And that heterogeneity has a 24 substantial effect. We have some preliminary studies, at 25 lower levels of the modeling pyramid, that give us an Q ANN RILEY & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 t l

l 289 1 indication -- some indications of what.that might do. A 2 little more work on that is something we would really like 3 to have in the next iteration. 4 We want to enhance the modeling of the saturated 5 zone. This is the first time we have done anything with the 6 saturated zone, and we were finding some interesting 7 features down there that may well be helpful, in terms of 8 understanding what -- how radionuclides might be retarded, - 9 what travel times might be. 10 Colloids have come up repeatedly over the last 11 three days. We would like to get some kind of way of 12 treating those -- some aspect of them dealt with in sur next 13 iteration, because it is a thorny issue that has been around 14 for years. 15 I would like to look at effects of water table 16 changes. Again, those have been around for a number of 17 years and trying to figure out how to incorporate them into 18 a total system perf3rmance essessment. Really, it comes 19 down to just how difficult it is to find a -- to create a 20 pathway down through one of those scenario trees that Felton 21 showed, and actually get to something that is a sensible 22 physical calculation. We found that -- I found the most 23 stunning example of that trying to pull together a scenario 24 for human intrusion -- trying to figure out what really 25 would happen if somebody put a drill bit on the ground,.and O- ^"" ai'ev a ^SSoci^Tes. 'td-Coud Repoders , 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 i (202) 293-3950 l

I J 290 1 tried to put it down there. And some of those insights that 2 came in there came from my experience in the oil industry. 3 And it was startling to me the kinds of things that people 4 were willing to postulate in the absence of any experience 5 in that field. 6 I will come back to that point in a minute. I 7 think it is an important aspect of the business of 8 predicting performance. 9 There will be some changes. We have had a 10 management and operations contractor come onboard, and they 11 will be a major player in this next iteration. Direction 12 will still come from me, as long as I am with the program, 13 or my successor, if someone else comes in. 14 But, the M&O will be doing most of the ( 15 coordination. That means, they will be the guys gathering 16 the frequent flier miles, going out to meetings, trying to 17 resolve some of the questions Norm was talking about. And, 18 if not, they'll be the ones on the video teleconference. 19 They will also be major players in the problem 20 definition, and in_ total system evaluation. They are 21 currently evaluating a golder total system performance 22- assessment code, which some of you have asked questions 23 about. And we will probably be using it in-the-next 24 iteration.- But, that is not absolutely certain. I want to 25 do some kind of comparison between the solely silicon-based O Aww ai'ev a Associates. 'ta. Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

291 1 ripped model that golder has put together, and the mixed , 2 silicon and carbon-based computer model that Mike Wilson 3 uses. 4 MR. POMEROY: Jerry, I have a couple of questions 5 with regard to that. One is, just to repeat from yesterday, 6 that the next time we do see a set of total systems 7 performance assessments, we will probably see three? Is 8 that correct? 9 MR. BOAK: No. You will see an integrated and 10 total system performance assessment the next time. 11 MR. POMEROY: Okay. 12 MR. BOAK: But, involvement in that will come, 13 both from Sandia and from the M&O. As far as I am 14 concerned, I would like to see the document that comes out ( 15 reporting on TSPA 2 to be a single DOE document. 16 MR. POMEROY: Right. 17 MR. BOAK: But, it will reflect the work of both 18 groups. It may show some analyses done on different -- with 19 different total system codes. But, I think, in the end, I 20 would like to see a choice of one or the other and.the final 21 resort being from those, and then, using the multiple models 22 as a check, in fact. 23 MR. POMEROY: Will the M&O's total system 24 evaluation be done with an independent code, or an 25 independent set of-models, or will it be done with the SNL? Q ANN RILEY & ASSOCIATES, Ltd. COud RepOners 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

292 em. 1 Do you have a good feeling for that at this point?

        )     2                                                 MR. BOAK:                                                                 We're working out the details of that 3  exactly.                                       What I would like to say is that there will be 4  aspects of total system evaluation that are done by both.

5 The final product is a DOE product, not an M&O or an SNL 6 product. 7 MR. POMEROY: The next question deals with, as you 8 know, at least we have heard, that the M&O is doing a 9 systems analysis. 10 MR. BOAK: Uh-huh. 11 MR. POMEROY: This is not a performance 12 assessment, this is a systems analysis. 13 MR. BOAK: Uh-huh. 14 MR. POMEROY: Will that systems analysis feed 15 into, in any way, or interact, in anyway, with the 16 performance assessment? That is, could you conceive of the 17 systems analysis modifying one of the trees, for example? 18 Or would you think that once the systems analysis is done it 19 might modify some aspect of what the codes are doing? 20 MR. BOAK: It may depend on timing. "There is 21 many a slip twixt the cup and the lip." 22 So, I think that, yes, in the end, if there are 23 things -- parts of that system analysis that drive the mine 24 geologic disposal' system to different choices -- obviously 25 thermal load and placement mode are important aspects of ANN RlLEY & ASSOCIATES, Ltd. f]~ Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

293

--1 .that._ And'one of the things we are doing behind the scenes, 2. is.trying to' lay out a specification for a new generation of.- 3 engineered barrier system codes -- one which may well have 4 to address'both types. 5 6 7 8

        -9 10 11 12 13

..:OL 15 ' 16 17 18 19

       ;20 21                                                                   -

22

       '23 24                                                                    -

25 'O' '- ^"" ai'av a Associates. 'id-

                                       .. Court Reporters 11612: K. Street,~ N.W., Suite 300 Washington,' D; C. 20006

(202) 293-3950 - -

294 1 EVENING SESSION 2 [6:00 p.m.) 3 MR. ESLINGER: Sandia will also be involved in the 4 problem definition. Scenario development has been 5 traditionally their task, and it will continue to be, as far 6 as I can tell, for the rest of TSPA-2. 7 Total system evaluation, which has been hitherto 8 almost solely their domain, except by the comparison of a 9 little bit of history in the Pacific Northwest Laboratories' 10 past program was originally a headquarters oversight group 11 which was at one oversight for three sites, and we have 12 been, during the past few years, working out the details of 13 how to merge those capabilities with the additional () 14 15 complication of the MNL coming in, so that PNL's role in this case will be reduced somewhat. They will primarily be 16 involved in dose modeling and support to the EPS 17 calculations and modeling efforts. 18 Lawrence Livermore will be largely involved in 19 working on the source term and near-field evaluation. A 20 number of people have referred in the last few days to Tom 21 Bushak's work, things like that. And, of course, because 22 Lawrence Livermore will still be doing a good deal of 23 development of the lowest level models in the engineered 24 barrier system, corrosion models, spent fuel models, glass 25 models, et cetera, their involvement will be-feeding into-. l O ^"" ni'ev & Associ^Tes. 'ta-Coud Repoders l 1612 K. Street, N.W., Suite 300  ; l \Nashington, D. C. 20006  ! ! (202).293 3950

295

• g- 1 that. And we will have support from a number of players c  :

2 primarily from USGS in Los Alamos. That's largely a tie, 3 one of these links that we in essence have continued to fund 4 just in order that their TPOs might not drag them off, but 5 in fact we buy a little of their time so that we can be sure 6 that we can get some time to do some of the geochemical and 7 volcanologic issues with LANL. 8 USGS -- the most fruitful interaction there has 9 been with Allen Flint. and that has led in conjunction with 10 some of the performance assessment people to the 11 specification of the Intraval test case, using actually 12 Yucca Mountain site data. 13 LBL has been a kind of research wing. They have 14 set them loose on certain topics that we think are fairly (} 15 significant, but they have tended to be research-oriented 16 issues; looking around at the data on hydrologic properties 17 to try and figure out what are the regularities that we do 18 see, what kind of things should we be capturing in our-19 models, and what sort of new ways might we go about modeling 20 hydrologic properties. 21 I'll come back to the point we have talked a fair 22 amount about, about the business of moving up this pyramid, 23 which I show as a triangle here. From process models, the 24 most detailed lowest level models, but those models which 25 represent the smallest pieces of reality, the most focused O ^"" ai'ev a Associates. 'ta-Coud Repoders S12 K. Street, N.W., Suite- 300 Washington, D. C. 20006 (202) 293-3950

296-3 1 intense pieces of reality, up to some' intermediate models l 2 that try to capture relevant parts of this -- that is to 3 say, relevant in the fact that they might affect 4 performance, and that they can't be left out without leaving 5 behind some important aspects of that performance. 6 And then finally, the highest level models which 7 are the most comprehensive in terms of trying to capture the 8 entire world of the repository. They are of necessity 9 simplified in terms of their representation of any 10 individual process, but it is the intention that that 11 simplification be done only where it can be done without 12 costing anything in terms of presenting all the details. 13 I have seen some models recently which you might 14 call not abstracted. They model individual atoms, droplets,

   }

15 colliding with solid objects. Virtually every other model I 16 have ever seen is clearly abstracted to some extent, and I 17 suppose you could say that those ones modeling clusters of 18 atoms are abstracted because they don't represent nuclear 19 processes. But hopefully those approximations that leave 20 behind certain details are adequate. 21- So this process goes on at every level which we 22 operate. When we do hydrologic _modeling, I don't think I 23 know of anybody who puts relativistic terms into his 24 representation of gravity. More or less we accept Newton on 25 the basis of expert judgment, even though theze are groups O ^"" aev a ^Ssoci^Tes' 'id-Coud Reponers 1612 K. Street, N.W., Suite 300 Washington,' D. C. 20006 (202) 293-3950 1

297 1 out there now gathering data to refine that model. 9 2 So I wanted to come back to this point that in 3 fact these system models are not necessarily simplified 4 models. They are in fact less simplified in terms of their 5 representation of the world. d Hopefully they are 6 representing relevant pieces of the world. 7 So we will be exercising all those levels of the 8 models that we have. - 9 What is the result going to look like? 10 Well, I have this feeling we haven't seen the last 11 of this guy. For one thing, if we don't crank out some 12 CCDFs in TSPA-2, it's going to be a little harder to talk 13 about how far we have come; what's the difference; what have 14 we learned; what's changed. 15 So I suspect we will be calculating CCDFs. 16 And we vill certainly be representing our results 17 probabilistically in some respects, almost certainly. 18 For those of you who are color fanatics, I do have 19 about, I think, 40 copies of this diagram in color, so if 20 you want to replace the one that's in your handout. It 21 represents something that was -- that did not exist when I 22 came into the nuclear industry three and a half years-ago. 23 There were no CCDFs available for the perfcermance of Yucca 24 Mountain, and now there are four -- take your pick. 25 The one that you haven't seen today actually is g ANN RILEY & ASSOCIATES, Ltd. Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D C. 20006 (202) 29'c 3950

298 1 the one that was produced by EPRI, and they-have a report O 2 out, I recommend it, it has some interesting different 3 approaches. I am surprised that it comes out quite as close 4 to Sandia's as it does. Because of those differences, there 5 are some really interesting phenomena that are modeled in 6 there that are not modeled here, and we have in fact 7 provided some support to Robin McGuire, who works for Risk 8 Engineering, who generated the top level part of that model. 9 So we feel a certain proprietary interest in that. 10 Which brings me to the question of how does NRC 11 evaluate this wealth, this embarrassment of riches. I'm a 12 little unsure how to answer that, except to say that, golly, 13 we'd love to have you read it all. We'd take any comments 14 you make. obviously I think Paul has addressed some aspects { 15 in which the PNL model did not get a chance to model the 16 high flux cases, so some parts of the aqueous modeling may 17 be of less interest and yet the way in which they approached 18 the aqueous phase releases is of interest in that it is 19 somewhat different, it has a different parentage than the 20 codes that Sandia was using, or the codes that your part of 21 Sandia has.used in the past. So there are some interesting 22 things to learn from that, I think. 23 The two-dimensional modeling is significant. I

   '24 think the source term may well be a relevant chapter to look

25 at, simply because for the first time we have a total system C ANN RlLEY & ASSOCIATES, Ltd.

Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

1 l 299 1 performance that to some extent takes account of waste 2 glass. 3 I would not have thought it was necessary to read 4 both chapters on volcanism before, but then this got so 5 contentious here that I realized that it may well be worth 6 it, although the results are quite similar, and again they 7 are conditioned by the same probabilities. That may be a 8 matter for the interested reader. - 9 The approaches to combination of CCDFs obviously 10 is something that we have talked a great deal about, and so 11 will be of interest to the NRC. 12 Paul has said that he is comfortable with all 13 three approaches from a statisical viewpoint, which I take (} 14 15 as a good sign. But hopefully next time we won't have this problem quite so much. We hope to have it consolidated and 16 present you with a summary volume that presents everything 17 that we have done or most-of what we have done. There still. 18 will be, I am certain, various pieces of all these reports 19 that will come out in the high level waste conference in 20 various scientific publications. We have been advised that 21 we ought to be submitting these things to more peer review 22 journals, although I am not sure where there is a peer 23 review journal that would review all 600 pages of these two 24 documents if we produced a total system performance 25 assessment as a single thing and tried to get it published p Q ANN RILEY & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l

i 300 f-w 1 as a journal issue.

  '~   2            MR. APTED:   I guess I'm wondering-the next time we 3 reconvene, that all these models are going to have bolted 4 onto them a dose model and maybe four different dose models 5 yet again, I mean, and the situation will actually be more 6 complex, or that maybe they will all now lie on top of each 7 other because the dose models are different.          Is that a 8 worry that -- there is certainly an advantage of different 9 conceptual models, but there is also some disadvantages of 10 everybody sort of independently deciding reality _is 11 different. Especially when it comes to this dose question-12 and future populations and that kind of question.

13 Would it not be better to have a more standard 14 approach? 15 MR. ESLINGER: Well, DOE will have a standard 16 approach. I hope. I mean I hope by that time we can come 37 up with something that is uniform as a presentation on what 18 DOE thinks. 19 But as we said repeatedly in-the discussion of 20 CCDFs, alternative views will have to be expressed in order 21 to show the range of things. You know, we haven't reduced 22 by getting a new standard in the Energy Act, we haven't 23 reduced the paperwork any, I don't think. We will still be 24 killing many trees in trying to communicate the details of 25 this result. O ^"" ni'av a Associates. 'id. Court Reporters i- 1612 K. Street, N.W., Suite 300 Washington, D. -C. 20006 (202) 293-3950

1

                                                                                                                             )

i 301 1 It may be hoped that we will-get some single 2 representation that we can take to the public and explain in 3 a relatively straightforward manner. I don't know. I'm 4 still not sure how I would explain this standard to the 5 public, or how I will explain why it changed. It is a 6 difficult question. I mean I have taken a lot of people, 7 several hundred people now, Nevadans primarily, to the Yucca 8 Mountain site, and nobody has ever asked a really tough 9 question about, well, what is the standard. So I have been 10 spared the agony. But I don't think that will go on 11 forever, and I am not really sure how we are going to 12 surmount the problem of communicating that we have done 13 here, and assuring the people that what we have done is () 14 something intelligent. 15 (Slide.) 16 MR. ESLINGER: The conceptual schedule was created 17 on Saturday afternoon with no input, so it is entirely 18 conceptual. If we are really lucky, and we can lay out a 19 scope that is reasonable, :;e might get to the point of' 20 having some kind of internal review in the first quarter of 21 1994. 22 How long that will take, and exactly the path it 23 will go through, is unclear, but once that document is out 24 -- and I would prefer it to be a DOE document -- at least it 25 will be a single document going out under my name, with a () ANN RILEY & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

302 1 request for review. It may go out as a draft document, 2 actually, with a request for public review by bodies like 3 this one, by the ftate of Nevada, the Nuclear-Waste 4 Technical Review Board, and anyone else who would happen to 5 request it. We have sent out, I think it's close to 350, or 6 is it 450 copies of the Sandia Total System Performance 7 Assessment. It's not in Barnes & Noble yet, but I am about 8 to run out of the second printing, so those of you who 9 haven't gotten one yet, you may be out of luck for a while 10 till we get that third printing done. But we are sending 11 the PNL document out to the same list of people. 12 MR. POMEROY: You might note for those who don't 13 read the bottom of the slide -- 14 MR. ESLINGER: Oh, I'm sorry, I forgot about that. 15 Recently the OECD/NEA completed a review of a 16 document prepared by SKI, the Swedish regulatory agency, in 17 which they did a conceptual performance assessment. It was 18 not in fact using real site-data, it was a fabricated site 19 that they did a performance assessment of. Seth Coplan of 20 the NRC participated in that, as did Scott Sineck of our M&O 21 contractor. 22 One of the things that I would like to do with 23 this next generation of total system performance assessment 24 is to submit it to the OECD and ask them to review it-and 25 give us comments. I haven't figured out exactly whether O ^"" ai'ev * ^Sso c'ATesid-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

                                .~      -  -        -.

1 303 11 that's going to be a review of the draft version, or will 2 DOE issue revisions to that as a consequence of those 3 comments. We certainly want to get as wide'and external a 4 review of that as we can, so that either we can issue a 5 revision of it, or we can make a commitment to incorporate 6 those inputs into our next iteration. 7 MR. MOELLER: I had a question on that. You had 8 said, I believe, yesterday that during the work that you had 9 been doing on the TSPA, you have not attempted, at least at 10 this stage, to provide all the quality assurance 11 documentation and so forth that may be ultimately needed. 12 And perhaps this question is to the NRC Staff, but to what 13 degree would a review by OECD or NEA help to compensate, or () 14 15 does it compensate at all, for QA documentation? I am curious. I don't know, and I would like to 16 have some idea. Do they relate at all? I mean that's 17 certainly international peer review, I presume. 18 MR. FEDERLINE: Just from my limited knowledge of 19 our QA requirements, that would not totally satisfy the QA 20 requirements. 21 MR. BOAK: It might be sufficient to cite that 22 peer review as qualifying some aspects, some pieces of.that 23 document, so that-they could be quoted in-license 24 applications. It's possible we could use that, with some-25 additional work, to qualify some piece of it. I would be Q ANN RILEY &- ASSOCIATES, Ltd. Coud _ Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

304 1 surprised if it would be sufficient in and of itself. It O 2 would be an important piece, if we chose after the fact to 3 take some piece of that and put it into our license 4 application, I think it would be an important piece of that. 5 I wanted to touch quickly on something, an effort 6 that DOE is involved in, and John mentioned it briefly, and 7 that is there has been an effort in the last year to submit 8 several additions of an outline as well as a couple of issue 9 resolution documents. 10 What we want to do with those issue resolution 11 documents is present to the NRC a comprehensive review of 12 the data surrounding some relevant issue, and the one that 13 has been submitted is on erosion. We are planning one in (~h 14 June, planning a technical exchange on an issue resolution d 15 paper with respect to volcanism. 16 The object would be to get some kind -- to give to 17 NRC a suite of data that we say represents a reasonable case 18 for -- that this information is sufficient for the license 19 application, and barring some new discovery. 20 That is to say it might enable us to narrow the 21 scope of some parts of our site characterization activity if 22 we could get some agreement that at least for the time-23 being, this data looks sufficient to support license 24 applications, and we wish to close these issues. We know 25 that a great deal of data will be' collected in the next 10 0 ^"" aev a ^SSoc'ATES. 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

305

 - f--

1 years, and some of it may well open some old wounds. But we ( 2 think that there are some issues that are sufficiently 3 advanced that we can say we think no further data is needed, 4 and until such time as new things come up, say that that 5 issue doesn't need to be talked about a whole lot more and 6 we don't need to spend a great deal of effort on it. 7 I wanted to say a few things, because Paul had 8 mentioned a philosophical point, and it hit home to some 9 issues I had been thinking about in the course of this 10 presentation. 11 Actually, one other point about site interaction, 12 site characterization and performance assessment, how are we 13 interacting. I have had a request from our hydrology () 14 15 integration task force and our geochemistry integration' team to present the results of our total system performance to 16 present the results of our total system performance 17 assessment to groups of principal investigators on site at 18 Los Alamos and at USGS, and I think also at Livermore. 19 So we will basically be taking this show on the 20 road one more time to try and see if there aren't some 21 interesting interactions we can get in the way of direct 22 contact with the PIs doing site characterization. That 23 should be going on, and hopefully in the next month or so. 24 want to say a few things about uncertainty 25 because it came up so often, and'about maxium enterpy O ^"" ai'ev a Associates. 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

306 1 approach that Paul Caplan had used to elicit judgments about 2 certain parameters. 3 The criticism was leveled against it yesterday 4 that it is not a robust technique, in that if you change the 5 variable, if you essentially transform the variable, you 6 don't got the name distribution by that convention. 7 However, in the clicitation process, one of the 8 things that Paul has begun to do -- and it la largely 9 parallel to the way in which judgments and probability 10 distributions are elicited by decision analysto -- is to 11 look at his rangen no representing physical limita. 12 For example, if you wish to clicit porosity, 13 obviously the limits are zero and 1, and so those are his 14 rar. gen for the clicitation of poroalty. And then after 15 that, what he is looking for in a mean value and some idea 16 of the coefficient of variation. 17 once he has generated a preliminary distribution 18 for that, then he looks to coe what other information the 39 analyst might have that makes him uncomfortable or 20 comfortable with the form of that distribution. 21 If you transform flux to one over flux, first of 22 all you are going to have a very difficult time getting a 23 reasonable distribution out of it because, in fact, most 24 experts will tell you that the probability distribution 25 function should be weighted heavily at negative infinity and O' ^"" ni'ev & Associ^Tes. 'ta. Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 ) (202) 293 3950

307 1 positive infinity. So it would be very hard to get a O 2 distribution that you are comfortable looking at, and saying 3 that represents your vievooint, and I think that is exactly 4 why it is that you go about thinking about how to transform 5 the variable so that it can adequately represent what you 6 are doing. 7 If you properly elicited, I think, the flux 8 values, you might actually come up with the same value. 9 one of the things that Paul talks about doing in 10 that, for example, if you want to choose a log uniform 11 versus a regular uniform, what you are saying is that the 12 probability distribution function should be equally weighted 13 for each decade. That is to say, the probability of values 14 from 1 to 10 is roughly the same as probabilitics of values

}

15 from 10 to 100. 16 That is very different from saying that the 17 probability of values from 1 to 5 is the same as the 18 probability of values from 5 to 10. And so if that expert 19 has a reason for wanting his distribution to be log uniform 20 versus uniform, it probably reflects something that you can 21 actually get out of him. You can provide a justification 22 for providing, for using a log uniform distribution rather 23 than a plain uniform distribution. And the process of doing 24 that, I think, is exactly parallel to the process that 25 decision analysts go through when they try to elicit say C ANN RILEY & ASSOCIATES, Ltd. Coud Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

308 gs 1 quartile percentages or something like that. They are 2 trying to give them information, post choices, Bo make it 3 casier for the expert to capture what he knows, so that he 4 can then provide a reason for it. 5 The only piece of evloence I have to say that that 6 technique is as valid for getting those probability 7 distributions as those that the decision analysts use is 8 that in taking an expert judgment class with Steve Hora, who 9 has worked for the NRC and many others, he performed one of 10 the standard techniques to dctermine biases in expert 11 judgments, which is the Almanac Test, and gives a list of 12 questions derived from values in almanacs, and asks you to 13 provide a probability, an estimate in your probabilitle:s for 14 your estimate of the values of the answers to those 15 questions. And they are usually questions you don't have 16 any really good clue for. 17 In that class he demonstrated very nicely the 18 tendency that is observed in experts overywhere for people 19 to be overconfident, and he actually had measures by summing 20 up the values that people had, had a nearure of the degree 21 of overconfidence of the people in that class. 22 Two people in this class were underconfident. 23 They were Paul Kaplan, who originated this techn'que, and 24 myself, who used it in solving -- in answering the almanac 25 questions. ANN RlLEY & ASSOCIATES, Ltd. Q' Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

l 309 1 Now that isn't, of course, a rigorous test of the O 2 validity of this approach to eliciting probability 3 distributions, bat it does pose interesting research 4 questions. 5 It leads me to another thought I had about 6 uncertainties, which relates to uncertainties in engineered 7 b3rrier systems and uncertainties in natural systems. It a has to do with the fantasy of engineered systems. 9 I didn't realize this phenomenon until I sat 1r through a meeting about characterization of the waste form, 11 and the experts there, all of whom knew a great deal about 12 spent fuel, really despaired of fully characterizing our 13 uncertainties about spent fuel, even to the point of not 14 really ever really knowing just how much was in -- what part 15 of the inventory was on the grain boundarios or in the spent 16 fuel gap, the fast release fraction. 17 They constantly cited just how uncertain various 18 things were, the grain size distrib*1 tion, the porosity, the 19 chemistry, what was in there, and it was truly astounding to 20 me as a geologist, because I look at spent fuel, and I see 21 an incredible monomineralic rock with a really narrow range 22 of porosity, and a grain size distribution that is trivial 23 compared to any sediment you ever want to look at, and so it 24 suddenly dawned on me that uncertainty about engineered 25 systems -- that what their uncertainty was about was the O' ^"" ni'ev & Associates. 'id. Coud Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l

r 310 1 fact that this stuff was manufactured to very tight 0 2 specifications, but the specifications weren't for waste i 3 performance. 4 I then realized that an engineered system is one 5 that you haven't built yet, and one need only look at the 6 space shuttle, Titan Dam, and any number of nuclear waste  ; 7 shots at the Nevada test site, to realize that in fact this 8 fantasy that engineered barrier systems are so well 9 characterized and so easy to characterize is widespread, but 10 not necessarily valid. 11 I get many more questions about the public about 12 how we can possibly demonstrate that a manmade system is 13 going to last for 10,000 yearn than I do about geologic 14 systems. Their uncertainties, contrary to what some people ( 15 think, are really more about the mannade systems. 16 That leads to a question sbout prediction. It 17 goes back to what Paul said, and %:at is our predictions -- 18 what we need to predict is not intimate details of the 19 future state of the site, and for that I use an example that 20 is drawn from something that I think still everybody agrees 21 is even more uncertain than either engineered barriers or 22 natural barriers, and that is human activities. 23 I quote from a performance assessment that I 24 recently got ahold of and read, and it is a performance-25 assessment made of the Soviet Empire, and I will just quote O ^"" ai'ev a Assoc'ATes' 'id-Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202)_293-3950

311 1 one relatively precise prediction made in there, and ask the 2 question does this confirm the assertion made by politicians 3 and newsmen and other people with very short time horizons, l 4 that no one could have predicted the fall of the Berlin Wall 5 or the distintegration of the Soviet Empire, even five years 6 ago. 7 It says here: 8 "Who can say whether in these circumstances the 9 eventual rejuvenation of the higher spheres of authority, 10 which can only be a matter of time" -- this is addressing 11 the question of what happens when people like Andropov go 12 away - "can take place smoothly and peacefully, or whether 13 rivals in the quest for higher power will not eventually 14 reach down into these politically immature ar.d inexperienced 15 masses in order to find support for their respective 16 claims?" 17 He's referring to the younger members of the 18 Communist Party who had no experience of really running 19 things. 20 "If this were ever to h=ppen, strange consequences 21 could flow for the Communist Party, for the membership at 22 large has been exercised only in the practices of iron 23 discipline and obedience, and not in the arts of compromise 24 and accommodation." 25 And now here is the prediction: O ^"" ai'ev a ^ssoci^Tes. 'id-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

312 1 "And if disunity were ever to seize and paralyze O 2 the party, the chaos and weakness of Russian society would 3 be revealed in forms beyond description." 4 I wnuld suggest that that is a fine description of 5 what is going on in the Congress of People's Deputies over 6 the past three years. 7 Now I cannot assert fully that predicting human 8 futures 45 years in advance is exactly equivalent to 9 predicting geological and engineered systems 10,000 years in 10 advance, but it is certainly pretty flaky. 11 That prediction was, in fact, done 47 years ago by 12 X in Foreign Policy, June 1947. Luckily, George Kennan has 13 lived to see his prediction come true, and I would say given {} 14 15 that this was a performance assessment conducted solely on carbon-based computers, and that on the basis of this risks 16 were taken with the lives of 100 million or more, probably 17 300 million by the time you add up all the people who have 18 died in the subsequent time -- 300 million lives, without 19 their informed consent, as it was implemented in the policy 20 of containment, that we actually are already doing a better 21 job of trying to communicate and provide informed consent 22 about the risks involved in repository performance 23 assessment. We have a long way to go to getting truly 24 informed consent, but the precedent is there. 25 Thank you. O ANN RlLEY & ASSOCIATES, Ltd. Coud depoders 1 1612 K. Street, N.W., Suite 300 L Washington, D. C. 20006 (202) 293-3950 1

, . . .                                                                                                        l l

313 1 MR. POMEROY: Thank you, Jerry. 2 on behalf of the committee, I would like to thank 3 doe, but oppecially you, Felton, Holly, and Paul, for 4 bearing with us through the third day of this. We are 5 deeply appreciative of the insights that you have given us, 6 and we look forward to seeing you with TSPA. 7 MR. DOAK: I forgot my last viewgraph, which is 8 the kind of prediction that we don't need to do. 9 MR. PoMEROY: I'm not going to try to do at this 10 late hour a full scale round table discussion, obviously. 11 I would like to give Steve Frishman an opportunity 12 to give us his views with regard to some of the things we 13 have heard today. ( ) 14 We will then go around the table quickly, if 15 anybody wants to make any last comments. 16 MR. FRISHMAN: Thank you, Paul. 17 I am not going to take the time to repeat what we 18 talked about yesterday. Now there will be a report of that 19 technical exchange, and I am sure it can be available for 20 your records. 21 What I did want to do was open up what I believe 22 is probably a new topic of consideration, but I think is 23 very much related to the discussions and the work that is 24 being discussed here relative to the configuration of the 25 new legislation, and that is that because the new { ANN RILEY & ASSOCIATES, Ltd. Coud . Reponers 1612 K. Street, N.W., Suite 300 Washington, D. C 20006 (202) 293-3950

314 1 regulations are mandated to protect the public from 2 increased exposure tha. would result from a Yucca Mountain 3 repository, I think what that means is that when it is a 4 dose standard, then the dose standard is going to have to be 5 for the collective dose to the individual. 6 How look where Yucca Mountain is. Yucca Mountain 7 is adjacent and, in fact, in part on the Nevada Test Site. 8 I have not yet heard any discussion of how the Nevada Test 9 Site as a source is going to be accounted for within this 10 program, and I think now that a dose standard is mandated, 11 it is going to have to be accounted for. 12 So we know at least something about the types and 13 extent of contamination in the area. It's being much (} 14 15 further characterized now because it has to be relative to the clean-up program. But we do know of a number of 16 different types of events that have in fact first placed and 17 just through time have entrained radionuclides in the Yucca 18 Mountain system, and adjacent to the Yucca Mountain system. 19 We know there's carbon 14 in interstitial waters 20 in the rock there. We know that the earlier rocket programs 21 blow plutonium all over area 25. We know that from 22 atmospheric testing, there's some extent of strontium and 23 cesium in the area, and there is probably a lot more. And 24 if there is going to be a dose standard, we're going to have 25 to start looking into pathways, and_there's material Q ANN RILEY & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293 3950

315 1 entrained in those pathways right now that is all ultimately 2 going to result in some exposure calculation that is going l 3 to have to include Yucca hountain. No longer is it in the 4 original EPA rule. There in fact was the allowance for what 5 we called a double exposure. There was 25 millirem to the 6 public from a repository and another 25 millirem permissible 7 from an adjacent or coincident defense facility. 8 No longer is that possible, the way this 9 legislation is written. And because you are using -- you 10 have a mandated dose standard. 11 There is another part that is at least equally 12 interesting, and that is that there is at least the 13 possibility -- and this has been reported in the literature

    'L 4        -- the possibility that the repository itself will result in 15         another release that has been on and off considered in 16         different documents of the program, but never looked at 17         seriously, and we can't find where there is an intent to.

18 And that's that it seems like there's at least a strong 19 possibility that an increased thermal load or an increased 20 temperature of the repository will increase _the radon 21 emissions from Yucca Mountain by, it's estimated, a factor 22 of four to four and a half. And it's estimated in that case 23 that we are looking at on the order of hundreds to_maybe a - 24 thousand curies of radon and daughters annually as a result 25 of the repository. O ANN RILEY & ASSOCIATES, Ltd. Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 . l (202) 293-3950 u

 .-      . , ~ . = . . - . . . - , . .

_ _ . _ . _ _ _ _ _ . _ _ . . _ _ _ . _ _ _. _ .__ - _ . . . . _ . . ~ . _ . . _ _ _ __ i l 316 l 1 So we have another release from the repository 2 that hasn't been considered, and I don't see it considered  ; 3 in any of the performance assessments. I don't see it even 4 discussed in terms of thermal loading. 5 So these are two factors that add up to the 6 thought that, first of all, it is a collective dose we are 7 going to have to come up with, and performance assessment is 8 going to have to deal with all of the radionuclides in the 9 system, not just releases from a repository that are then 10 looked at as concentrations, and then 17tked at as dores. 11 Because you have built into the systein radionuclides already 12 that are in excess of what you would find maybe at other 13 places. (} 14 So that's a thought that I think at this point 15 needs to be factored in, especially because of the very 16 strong interest I hear on the part of both the NRC Staff and 17 doe to interact with the Na ional Academy when they start 18 thinking about what to recommend to EPA, 7.nd once again, as 19 I said yesterday, I certainly hope it we are going to have 20 assistance and recommer,dation- it's as well informed as we

21 can possibly get it, i

j 22 MR. OKRENT: Where in the law do you think it says 23 that? Because I don't see-anything about collective dose in 24 the request as to what the National Academy should do. l l 25 MR. FRISHMAN: What I am looking at is, first of 1 l ANN RILEY & ASSOCIATES, Ltd. O- Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950 l

317 1 all, the assumption that the NRC is directed to make when 2 it's revisf rig its rule, and then the requirement on the l 1 3 Secretary regarding post-closure oversight, and I see the l 4 direction to both the NRC to assume and to the Secretary to 5 -- well, we'll take the Secretary's, because it's the easier 6 one to follow. 7 "Following repository closure, the Secretary of 8 Energy shall continue to oversee the Yucca Mountain site to 9 prevent any activity at the site that poses an unreasonable 10 risk of (2) increasing the exposure of individual members of 11 the public to radiation beyond allowable limits." i 12 I take. that allowable limit to be the tota.' that 13 the individual -- the limit of individual exposure including () 14 15 Yucca Mountain and all other sources. MR. OKRENT: Well, I guess I see where you're 16 drawing the conclusion. I must say I couldn't decide on 17 reading this that there. was a clear direction from the 18 Congress in this regard. That's why -- 19 MR. FRISHMAN: Well, I'm not sure it was intended 20 to be clear. I think it may have been inadvertent in their 21 zeal to get something down on paper. 22 MR. POMEROY: I don't think we are going to solve 23 that problem tonight. 24 MR. MOELLER: A quick comment, though. I think, 25 Steve, when you were-using the word " collective dose," you O ANN FilLEY & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

318 l 1 were --  ! 2 MR. FRISHMAN: I probably meant cumulative. 3 MR. MOELLIR: -- you meant from several sources? 4 MR. FRISHMAN: Yes. Collected. 5 MR. MOELLER: Okay, that's helpful. 6 But just another quick comment. In the -- well, I 7 guess I'm just not sure what EPA's limit would be for 1 8 members of the public right now. I l 9 MR. FRISHMAN: Well, I'm not sure either, and I  ; 10 don't think any of us could really guess what that number 11 would be. But what I'm saying is that considerations of 12 performance assessment, first of all, need to be broader 13 than what they are now, and if there is going to be 14 interaction that ultimately brings a rulemaking onck to NRC, ( 15 then that consideration has to start now, and I believe 16 should consider much more than just the releases from the 17 repository itself. There are at least two other factors 18 that have to go into it. 19 MR. BOAK: Would that include the carbon 14 20 releases from wood fire in Clark County which comes to about 21 a quarter of the EPA limit over 10,000 years?- 22 MR. FRISHMAN: We have a source existing at the 23 same place as Yucca Mountain. We also have a source that is 24 a manufactured source that is essentially -- or that is 25 coincident with this project, and it's a source that Q ANN RILEY & ASSOCIATES, Ltd. Coud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C, 20006 (202) 293 3950

319 1 contributes not globally to the individual exposured at 2 Yucca Mountain, it contributes in addition to globally, and 3 I would say that the Las Vegas firewood source probably is 4 part of the global contribution. We know that there is 5 something special about the Nevada Test Site; otherwise, 6 there wouldn't be guards and there wouldn't be EPA 7 monitoring, and there wouldn't be concern about it. 8 I don't quite know what the committee, if it is so 9 inclined, could suggest be done about this, but I think it 10 is certainly worthy of thinking about and thinking about 11 pretty quickly, if there is going to be communication that 12 ultimately is going to come back to the Commission for 13 rulemaking. 14 MR. POMEROY: Okay, thank you, Steve. 15 MR. FRISHMAN: Thank you. 16 MR. POMEROY: I would like to give everybody 17 around the table a chance to make any last comments if 18 they'd like to. I am going to have a necting right here 19 after this meeting with the committee's consultants, to try 20 to pull some of these comments together. So if you have any 21 last-minute things that you would like to leave freshly 22 impressed on our minds, please do it now. 23 Would you start, Margaret. 24 MS. FEDERLINE: Okay, fine. Very, very briefly. l 25 I know the hour is late. O ^"" ai'ev a Associates' 'ta-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

320 1 We just wanted to reinforce in the previous 2 comments on the importance of iterative performance 3 assessment for DOE to continue to perform this iterative 4 performance assessment as a continuing tool to ensure that 5 site charactorization produces the data necessary to measure 6 performance. 7 We agree with the need for more detailed models. 8 However, we want to weigh in with a caution that the total 9 system analysis should be used as a tool to prioritize and 10 guide the way, as well as not only TSPA, but the insights 11 from modelers should also be used as sort of a joint 12 framework to sort of guide that. 13 We also want to recognize the importance -- wo 14 have sort of focused on is this code right, is that code 15 right. We want to focus on the icportance of a process 16 which produces skilled analysts to look at an integrated 17 system. I think that is what we should all be focused on, la and we are looking at very long time periods, and we can 19 argue today whether our code or DOE's code is right, but I 20 think the long term goal is to produce these knowledgeable 21 analysts who are looking at the integrated system. 22 So those are pretty global comments, but we have 23 gone through many of the others, like the need for technical 24 discussions between NRC and DOE, so given the hour, we won't 25 go into those. O ^w" ai'ev a Associates. 'ta. Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 (202) 293-3950

321 1 MR. POMEROY: Thank you, Margaret. 2 Norm, do you have anything to add to that? 3 MR. EISENBERG: Let me just add one thing to what l 4 Margaret, is that I believe we are looking at a suite of i 5 codes and tools, not a single code. And we have to keep 6 that in mind, that wo need to do the analysis at several 7 different levels, as many people have said, and that you are 8 going to need -- the critical issues for determining 9 licensability will be argued at potentially several 10 different levels, and everybody needs to be equipped to be 11 able to argue at those levels and make a case at those 12 levels. 13 MR. POMEROY: Thank you. 14 Members of the committee? Dade?

 }

15 MR. MOELLER: I think I'll hold mine until 16 tomorrow or the next day. 17 MR. POMEROY: Martin? You don't have to. 18 MR. STEINDLER: I'm tempted to -- yes, I realize 19 that. 20 Let me raise simply three points: l 21 I think an unresolved issue that is going to_come l 22 back to bite us, which we need to work on in some fashion or 23 other collectively or individually, is this whole question 24 of how do we safely exerciso expert judgment and incorporate 25 it into the processes. 'O ^"" ai'ev $ ^ssoci^Tes' 'id-COud Repoders 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 l (202) 293-3950

l l 322 1 Secondly, I am troubled in generic terms by the 2 drive toward mechanistic approaches to codes, because I 3 think that is a nonclosurable, if you will pardon that 4 terrible word, issue in the kind of time schedule that we 5 are normally talking about. 6 Thirdly, I would say we collectively need to make 7 sure that we can defend down to the very detailed algorithms 8 used by folks who want to speed up the process in machines 9 the assumptions that are made to put a code together and 10 defend those against the question do you have any idea what 11 the limitotions of those codes are in the real world. And 12 equally important, since presumably we are gathering some 13 information from the folks in laboratories somewhere, 14 whether or not the data that are being obtained in 15 laboratories have any relationship to the repository, the 16 real world in the repository. 17 Some of those have been issues that the committee 18 has raised before, and has transmitted to the chairman in 19 response to some questions that he has, and some of them 20 have simply been brought out in this very intere ting 21 session that we have had all day, 22 MR. POMEROY: Thank you, Marty. 23 MR. IIINZE: I think I'll pass. 24 MR. POMEROY: llolly? No? 25 Well, the hour is late. O ^"' aiu< * ^SSoci^Tes. 'td-Court Reporters 1612 K. Street, N.W., Suite 300 Washington, D. C. 20006 l (202) 293-3950

323 1 Paul? Jerry, do you want to have any last words? 2 MR. BOAK: I wanted to thank Dr. Steindler for 3 urging us to get out into the field. It does good to a 4 geologist to hear somebody suggesting that we ought to plug 5 into reality, that modelers like to go and get bigger 6 computers and the geologists say let's go to the field. 7 The last thing I was going to say was I wanted to 8 add a laudatcry note because I have been sitting on this end 9 of the table, I have noticed what a phenomenal deal of 10 stamina it took to be a stenographer for this effort, 11 particularly today, and I thought I'd add a laudatory note 12 to this persor; over here who has worked so hard. 13 (Applause.) 14 MR. POMEROY: I'd like to thank everybody for 15 participating today, and say that I enjoyed, and the people 16 who are here certainly enjoyed the past two days as well. 17 Thank you all, and we will see you soon. 18 (Whereupon, at 6:48 p.m., the committee was 19 adjourned.] 20 21 22 23 24 l 25 Q ANN RILEY & ASSOCIATES, Ltd. Court Reporters 1612 K. Street, f(W., Suite 300

                                           . Washington, D. C. 20006 (202) 293 3950

REPORTER'S CERTIFICATE This is to certify that the attached proceedings before the United States Nuclear Regulatory Commission In the Matter oft NAME OF PROCEEDING: ACNW Working Group DOCKET NUMBER: PLACE OF PROCEEDING: liethesda, Maryland were held as herein appears, and that this is the original transcript thereof for the file of the United States Nuclear Regulatory Commission taken by me and thereafter reduced to typewriting by me or under the direction of the court reporting company, and that the O- transcript is a true and accurate record of the foregoing proceedings, m')

                                                         / ~;1s.<     , C, ,  ;-/ : xlx ,_)

{/ Official Reporter Ann Riley & Associates, Ltd. O 3

1 l VERSION 2 12/03/92 INTRODUCTORY STATEMENT BY THE WORKING GROUP CHAIRMAN OF THE l ACNW WORKING GROUP MEETING ON PERFORMANCE ASSESSMENT IN THE HLW PROGRAM DECEMBER 16, 1992 TH2 MEETING WILL COME TO ORDER. THIS IS THE WORKING GROUP MEETING OF THE ADVISORY COMMITTEE ON NUCLEAR WASTE ON PERFORMANCE ASSESSMENT IN THE HLW PROGRAM. I AM PAUL POMEROY, CHAIRMAN OF THIS WORKING GROUP AND A MEMBER OF THE ADVISORY COMMITTEE ON NUCLEAR WASTE. OTHER MEMBERS PRESENT TODAY ARE: DADE MOELLER, CHAIRMAN OF THE ACNW MARTIN STEINDLER, VICE-CHAIRMAN OF THE ACNW AND BILL HINZE AS CONSULTANTS AND EXPERTS ASSISTING THE WORKING GROUP, WE HAVE WITH US TODAY: DAVID OKRENT, PROFESSOR EMERITUS, UCLA MICK APTED, INTERA. SCIENCES, INC. PAUL DAVIS, SANDIA NATIONAL LABORATORIES OUR SUBJECT TODAY IS THE CURRENT STATUS OF THE PERFORMANCE ASSESSMENT -- BOTH IPA AND TSPA -- EFFORTS WITHIN THE HIGH-LEVEL RADIOACTIVE WASTE (HLW) DISPOSAL PROGRAM.  ; IN ORDER TO CLEARLY UNDERSTAND HOW RESEARCH, LABORATORY AND FIELD WORK, AND OTHER ANALYTICAL PROCESSES FACTOR INTO.THE METHODOLOGY OF PERFORMANCE ASSESSMENT, THE COMMITTEE HAS REQUESTED THAT BOTH NRC AND DOE STAFFS PRESENT THEIR RESPECTIVE APPROACHES IN INTEGRATING -i MODELING AND INFORMATION ACQUISITION. - IT IS ALSO EXPECTED THAT THE NPC AND DOE STAFFS ADDRESS HOW THEIR MODELING AND PA ' ACTIVITIES-INFLUENCE THEIR--INFORMATION-ACQUISITION EFFORTS. THIS MORNING WE WILL. HEAR- FROM THE NRC STAFF AND,- THIS AFTERNOON, FROM d REPRESENTATIVES-OF THE U.-5,. DEPARTMENT OF ENERGY (DOE).

   ;O 1

a

  'a BOTH THE NRC AND DOE STAFF HAVE BEEN REQUESTED TO BRIEF Tl!E ACNW REGARDING THEIR PROGRESS IN TOTAL SYSTEMS PERFORMANCE- ANALYSIS (TSPA). ISSUES IMPORTANT TO THIS ACNW WORKING GROUP INCLUDE:

A. WilAT IS THE CURRENT STATUS OF TIIE PA PROGRAMS? i B. HOW HAVE OTHER EFFORTS IN THE HLW PROGRAM AFFECTED THE EVOLUTION OF THE PA AND TSPA CAPABILITIES (E.G., RESEARCH)? C. HOW HAVE PA ACCOMPLISHMENTS AFFECTED THE OTHER EFFORTS = IN THE HLW PROGRAM (SITE CHARACTERIZATION, TEST PRIORITIZATION, FIELD WORK, RESEARCll, ETC....)? D. WHAT ARE THE LESSONS LEARNED IN THE PROGRAM UP TO THIS POINT; IN PARTICULAR, ARE THERE ANY IDENTIFIABLE POTENTIALLY FATAL FLAWS IN THE PA AREA? r E. FINALLY, WHAT ARE THE ' FUTURE PLANS, IN THE SHORT TERM AND

• THE LONG TERM, FOR PA WITHIN THE NRC AND DOE?-

THE NRC STAFF IS IN PHASE 2 OF THE IPA,. WHICH IS FOCUSED ON , ASSEMBLING AN EXPANDED, MORE ROBUST TOTAL SYSTEM . CODE, WHICH-INCLUDES CONSEQUENCE MODULES. DOSE MODULES ARE ALSO BEING INCORPORATED. SIGNIFICANT PROGRESG HAS BEEN MADE:BY THE NRC STAFF IN DEVELOPING THE SECOND PHASE OF IPA. THE C-14 MODULE HAS SEEN SIGNIFICANT PROGRESS, AND A COMPLETE RUN-THROUGH OF THE ENTIRE l SERIES OF MODELS IS NEARING COMPLETION. THE STAFF HAS ALSO BEEN PURSUING AN APPROACH TO DEAL WITH AND USE EXPERT JUDGMENT IN PHASE 2.5. > THE DOE HAS MADE SIGNIFICANT PROGRESS IN THEIR TRANSITION FROM-THE: PERFORMANCE ASSESSMENT CALCULATIONAL EXERCISB- (PACE) . TO THE TSPA EFFORT IN 1991:AND.1992. - EFFORTS AT BATTELLE PACIFIC NORTHWEST LABORATORY AND AT SANDIA NATIONAL LABORATORIES HAVE BEEN INTEGRATED 2 L

                                                                 -e n sr..-a~--      r.

1 IN DEVELOPING A TSPA, PRESENTERS WILL INCLUDE MEMBERS OF THE NRC STAFF IN THE NUCLEAR MATERIAL SAFETY AND SAFEGUARDS, NUCLEAR REGULATORY RESEARCH; THE CENTER FOR NUCLEAR WASTE REGULATORY ANALYSES, THE DOE'S OFFICE OF CIVILIAN NUCLEAR WASTE MANAGEMENT (OCRWM); CONSULTANTS FROM NATIONAL LABORATORIES, AND OTHER DOE CONTRACTORS. THE WORKING GROUP FORMAT IS INFORMAL AND I ENCOURAGE PERTINENT QUESTIONS FROM OUR CONSULTANTS AND EXPERTS, AS WELL AS FROM THE

                                                                                   ]

COMMITTEE MEMBERS. WE DO NOT INTEND TO LIMIT DISCUSSION IN ANY WAY. AT THE SAME TIME, WE HAVE A LARGE BODY OF MATERIAL TO COVER BEFORE THE END OF THE DAY AND I KNOW EVERYONE WILL COOPERATE IN ' ATTEMPTING, AT LEAST, TO STAY ON SCHEDULE. -- I WOULD ALSO REQUEST THAT ANY OF THE INVITED EXPERTS PARTICIPATING IN THE WORKING GROUP TODAY RECUSE THEMSELVES FROM OR LIMIT THEIR PARTICIPATION IN ANY PORTION OF THE MEETING, WHERE THEY FEEL THAT THEIR OPINIONS OR ADVICE WOULD CONSTITUTE-A CONFLICT OF INTEREST WITH THE GENERIC GOALS AND PURPOSES OF THIS MEETING. IF TIME PERMITS, A FOLLOW UP DISCUSSION WILL BE HELD AT THE END OF TODAY'S PRESENTATIONS. THIS DISCUSSION WILL BE IN THE FORM OF A ROUND-TABLE ELICITATION, WHERE THE WORKING GROUP- MEMBERS, CONSULTANTS AND EXPERTS-WILL HAVE AN OPPORTUNITY TO COMMENT, AS WELL-AS PROVIDE OBSERVATIONS AND RECOMMENDATIONS. THIS MEETING IS BEING CONDUCTED IN ACCORDANCE WITH THE PROVISIONS OF THE FEDERAL ADVISORY COMMITTEE ACT. GIORGIO GNUGNOLI WILL SERVE AS DESIGNATED FEDERAL OFFICIALS THROUGHOUT THE MEETING. THE RULES FOR PARTICIPATION IN THIS MEETING HAVE BEEN' ANNOUNCED AS PART OF THE NOTICE OF THIS MEETING THAT WAS PUBLISHED IN THE FEDERAL REGISTER. O  ; 3  ;

                                 .    - _   _ . ~ . . _ _ . - - - _  . _ _ _ _ ._

O WE Have RECE1vso NO WR1TTeN Sr^rzxENrS OR R8oU8SrS TO xAx8 ORAL STATEMENTS FROM MEMBERS OF THE PUBLIC REGARDING THIS MEETING. A TRANSCRIPT OF PORTIONS OF THE MEETING WILL BE KEPT, AND IT IS REQUESTED THAT ALL SPEAKERS USE ONE OF THE MICROPHONES, IDENTIFY HIMSELF OR liERSELF, AND SPEAK WITH SUFFICIENT CLARITY AND VOLUME SO THAT HE OR SHE CAN BE HEARD. SHOULD ANY MEMBER OF THE PUBLIC OR ANY ORGANIZATION DESIRE TO MAKE ANY COMMENTS RELATIVE TO THE SUBJECT OF THIS MEETING, PLEASE MAKE YOUR INTENTIONS KNOWN TO GIORGIO GNUGNOLI, AND WE SHALL MAKE EVERY-EFFORT TO FIT YOU INTO Ti!E MEETING SCHEDULE, IF POSSIBLE. I WOULD LIKE ALSO LIKE TO ASK OUR INVITED EXPERTS TO PROVIDE, IF YOU FEEL IT WOULD BE USEFUL TO THE COMMITTEE, ANY WRITTEN COMMENTS ON YOUR OVERALL IMPRESSIONS OF THE NRC AND DOE STAFF'S PROGRESS AND DIRECTION IN THE AREA OF HLW DISPOSAL PERFORMANCE ASSESSMENT STRATEGIES. IF YOU WISH TO PROVIDE COMMENTS, PLEASE FORWARD THEM O TO GIORG10 GNUGNOL1, OF OUR STAFF. I ASSURE v0U THAT YOuR THOUGHrS WILL BE FACTORED INTO OUR FUTURE DELIEERATIONS. UNLESS OTHERWISE SPECIFIED BY YOU, YOUR REMARKS WILL BE - INCLUDED AS PART OF THE RECORD OF THIS PROCEEDING. BEFORE BEGINNING WITH THE FIRST PRESENTATION, I WOULD LIKE TO TAKE THIS OPPORTUNITY TO ASK WHETHER ANY OF THE ACNW MEMBERS OR-THE INVITED CONSULTANTS OR EXPERTS HAVE OPENING REMARKS TO MAKE OR QUESTIONS TO ASK? ,

   .O                                                                                                                                                                       .
                                                                                                       -4
   ,----,_.,r-.-
                      #--r...- . , _ , , ,      , ,..._--e-,,,--,-.,w.rw-- - - . - - , , . . . , , , ,    p    .-.,r.,-,,-.,.-my,m...-,g ,-w,w-,. --,,..--y,,   , , - , , ,

O O Os-

ACNW: WORKING GROUP MEETING ON PERFORMANCE ASSESSMENT NRC's. Iterative Performance Assessment Program i

Margaret V. Federline, Chief . Hydrology and Systems Performance Branch Division of High-Level Waste Management 301/504-3455 4 e l D!V CF 4.hVEL G ACNW L1 )

AGENDA

Introduction:

NRC's Performance Assessment Program for a Geologic Repository

• Overview of NRC's Performance Assessment Activities During IPA Phase 2
• Detailed Example of the IPA Process in a Specific Discipline: " Carbon
• Use of Expert Judgment in NRC's Licensing Process
• Considerations for Future NRC IPA Activities 2

ACNW WORKMG GaOUP MEETMG ON PEfWOffAANCE ASSESSMENT OfVf5KJN OF HtGH LEVEL WASTE MANAGEMENT DECEMBER 16.1992 ACNW - 2 g .

O O. O

Introduction:

NRC's Performance Assessment Program . for a Geologic Repository r H EVEL A M G T DECEMBER 16.1992 ACNW .3

ATTRIBUTES OF A PERFORMANCE ASSESSMENT

• Quantitative
• Magnitude and Likelihood Considered
• Directed to Regulatory Requirements ACNW WORKNG GROUP MEETNG ON PERFC PMANCE ASSESSMENT OfVtS3ON OF HIG41EVEL WASTE MANAGEMENT DECEMBER 16,1992 ACNW - 4 G G G

O O O TWO COMPONENTS OF PERFORMANCE ASSESSMENT

• Quantitative Estimates of Performance

                  *'                      Auxiliary Analyses o Evaluate assumptions o Determine parameter values o  Detailed models to support simplified performance assessment models l

ACNW WORKNG GROUP MEETNG ON PERFODMANCE ASSESSMENT OrvtsaON OF HlGH-EEVEL WASTE MANAGEMENT DECEMBER 16.1992 ACNW - 5

                             ~   '

s h - i Mission Differences Cause D?.erences in Scope of Respective Performance Assessment Frograms Between NRC and DOE

         *   ' DOE responsible for determining if the site is suitable for licensing through adequate -

characterization I a DOE develops tools and techniques necessary for demonstrating compliance with NRC 1 regulations !

• NRC and DOE consult during pre-licensino abase l

!.

• NRC reviews DOE's performance assessment iterations during pre-licensing consultations
• NRC evaluates DOE's demonstration of compliance in license applications
• NRC needs an independent understanding of repository performance
• NRC Research Program provides needed understanding ACNW WORKtJG GROUP MEETrJG ON PERrOfVAANCE ASSESSMENT DIVIS3ON OF HtGH-LEVEL WASTE MANAGEMENT DECEMBER 16.1992 ACNW - 6 I

O O. .O t. l ,

                                                                   ..._m               ,

Oi .O O' OVERALL OBJECTIVE NRC's PERFORMANCE ASSESSMENT PROGRAM

                       ....'to develop, maintain, and enhance the NRC staff capability to review effectively.                       i,
                    . performance assessment in the DOE License Application for HLW[high-level waste]
                    . repository. "

I i i a 4' i? ACNW WORKNG GROUP MEETNG ON PERFOf9AANCE ASSESSMENT DIVts:ON Of HIGH-LEVEL WASTE MANAGEMENT DECEMBER 16,1992', , ACNW - 7

i l

     , .:                                .                     ,                   -  .a                       . ..         ...

m W

                                                   -NRCL APPROACH TO THE REVIEW OF PA IN DOE's LICENSE APPLICATION SUBMITTAL                                                                  2 t

• NRC.will review the performance assessment in DOE's license application o Review entire performance assessment at a broad level

                                   - o ~ More detailed review of significant areas                                                                           ,

q o Iterative Performance Assessment (IPA) and the Systematic Regulatory . Analysis (SRA) l will indicate areas for verificationty independent NRC quantitative analysis o' At least one rough quantitative check on overall performance estimate i i 5 I I-ACNW WORKING GROUP MEETNG ON PENOPMANCE ASSESSMENT OtytSION OF HO+LEvil WASTE MANAGEMENT

                                                                                                                                   - DECEh*BER 16,1992 ACNW - 8 i,                                                                                                                                                           i
       - - _ . i--     - - - - - - -        . - -  -             -                  +.                       \c'.        a ~            ,

O' O O ADDITIONAL IPA BENEFITS

• Ongoing NRC evaluation of DOE site characterization
• Ongoing NRC evaluation of HLW regulations l
• Technical input to development of NRC staff positions and other regulatory products

               *     'antribute to the integration of multi-disciplinary perspectives in NRC HLW programmatic activities and research                                                                                                  i

• Identification of technical uncertainties helps to define research needs and establish programmatic priorities ACNW WORKNG GROUP MEETNG ON PERFORMANCE ASSESSMENT l DMSION OF HIGH-LEVEL WASTE MANAGtMENT DECEMBER 16,1992 ACNW - 9

STAFF PARTICIPATION IN IPA PHASE 2-i I ACTIVITY / TASK STAFF TASK 1 - TOTAL SYSTEM DEVELOPMENT CODE DEVELOPMENT LEAD: RBaca/CNWRA AND INTEGRATION Overall Structure of the Total System Code BSagar/CNWRA, RJanetzke/CNWRA, RBaca/CNWRA,- NEisenberg/NMSS Dose Modeling :: RNeel/NMSS .

              ' TASK 2.iSCENARIO' ANALYSIS '                                                       LEAD: NEisenberg/NMSS Evaluation of SNL Scenario Methodology                                             JPark/NMSS, NEisenberg/NMSS
               ' Seismic Analysis .                                                                AB!brahim/NMSS, NEisenberg/NMSS Climatic ' Analysis -                                                              MMiklas/CNWRA .

Magmatic Analysis LAbramson/RES, Lancaster/RES, LKovach/RES, JTrapp/NMSS ' f Human Intrusion ' Analysis NEisenberg/NMSS

TASK 3 - FLOW AND TRANSPORT . LEAD: TMcCartin/RES Unsaturated Flow Module and Data TMcCartin/RES Saturated Flow Module . WFord/NMSS, TMcCartin/RES -

               . Gas Module -                                                                      RWescott/NMSS
                                                                                                            . ACNW WORKP4G GFtOUP MEETING ON PERFOfvAANCE ASSESSMENT DIVISION OF HtGH-LEVEL WASTE MANAGEMENT
            '                                                                                                                                         . DECEMBER 18.1992 f ACNW -f10 h                                                                      h:                                                                     k

__m_._

                                        .m__ _ _

y _

                                                              ,3.'=. ,        e     n ,

s

.,: 14, , . 0- O O STAFF PARTICIPATION IN IPA PHASE 2 (continued) ACTIVITY / TASK STAFF TASK 3 - FLOW AND TRANSPORT (continued!

       " Carbon Analysis                                          RCodell/NMSS, WMurphy/CNWRA K, Approxirnation Modeling                                 JBradbury/NMSS Effects of layering and Dipping                            ABagtzoglou/CFAVRA USGS Regional Modeling                                     NColeman/NMSS Saturated Regional Modeling                                MAhola/CNWRA Two-Dimensional Unsaturated Modeling                       JPohle/NMSS, TMcCartin/CNWRA TASK 4 - SOURCE TERM                                       LEAD: JWalton/CNWRA Corrosion Models                                           BSagar/CNWRA Structural Failure Models                                  KChang/NMSS, DDancer/NMSS Waste Dissolution and Mass Transfer Modeling .             RCodeft/NMSS Cottoid Formation and Modelling                            TAhn/NMSS ACNW WORKPdG GROUP MEETPdG ON PERFORMANCE ASSESSMENT   '

DtVIS80N OF HIGH-LEVR WASTE MANAGEMENT oECEMBER 16.1992 ACNW.- 11 ,

STAFF PARTICIPATION IN IPA PHASE 2 (continued) ACTIVITY / TASK STAFF TASK 5 - CISRUPTIVE CONSEQUENCE ANALYSIS LEAD: PEisenberg/NMSS Human Intrusion Model NEisenberg/NMSS, CFrietas/CNWRA Magmatic Consequences Mode . RBaca/CPN/RA, RDrake/CNWRA, LLancaster,HES, JTrapp/NMSS, TMarguties/RES, NEisenberg/NMSS Climate Consequences Model MMiktas/CFAVRA, BSagar/CNWRA, Seismic Consequences Modet NEisenberg!NMSS, CFrietas/CNWRA, PNair/CPAVRA, AChoWdhury!CNWRA, A!brahim/NMSS, KChang,WMSS Auxiliary Analysis for Source Term under Magmatic Events RCode!!!NMSS TASK 6 -- SENSITI'/ITY AND UNCERTAINTY ANALYSIS LEAD: RCodeUNMSS Sensitivity and Uncertainty Analysis RByrne/NMSS, VColten-Brad:ey/NMSS Comparison of Various Sensitivity Analysis Methods BGureghian/CNWRA, RCodetLHMSS ACNW WO8ErdG GROUP V'ETPJG ON PERFOf%1ANCE ASSESSMENT DTVtsf0N OF HMH LEVEL WASTE MANAGEMD4T DECEMBER 16.1992 ACNW - 12 O O O

O O O NRC's Performance Assessment Activities During IPA Phase 2 Overview Changes in the IPA Methodology Resulting from Research and other HLW investigations Changes in the HLW Program Resulting from the IPA Process Lessons Learned D ISi i OF 46 -t VE EM T DECEMBER 16,1992 ACNW - 13 L

NRC's Performance Assessment Activities During IPA Phase 2: Overview Norman A. Eisenberg, Senior Operations Analyst Repository Performance Assessment Section Hydrology and Systems Performance Branch Division of High-Level Waste Management 301/504-2324 i j f l

                                                                         *~

l$Io'r~n$?$a$s~?u$$a$$$ l DECEMBER 16.1992 1 ACNW - 14 l O O O t . . . . , . . .

~ ~ pU n n V __ V-lPA PHASE 2 ADDITIONAL OBJECTIVES

• ONGOING NRC EVALUATION OF L'OE SITE CHARACTERIZATION PROGRAM a ONGOING NRC EVALUATION OF HLW-RELATED REGULATIONS
• TECHNICAL INPUT TO THE DEVELOPMENT OF:

o NRC STAFF TECHNICAL POSITIONS o OTHER NRC REGULATORY PRODUCTS (LARP) o NRC HLW RESEARCH PROGRAM

                                                                            " " "' " T E S 'oU J$t EE 7 7[E dSI$ ",u dr DECEMBER 16 1992' ACNW - 15

I u, ! IPA PHASE 2 SPECIAL OBJECTIVES *

    -*       . PROVIDE FOR- A SMOG'H TRANSITION OF CONTRACTUAL SUPPORT FROM SANDIA NATIONAL L AMOHATOb'r (SNL) TO THE CENTER FOR NUCLEAR WASTE REGULATORY-                                          ;

ANALYSES-(CNWRA)

• EVALUATE THE TUFF METHODOLOGY PREPARED BY SNL ,

i _l 9

                                                                                                                               ?

1 ACNW WORKWG GROUP MEETING ON PEfW0%1ANCE MSSZSSMENT , DMSO4 OF HIGi+LEVR WASTE MANAGEME*8T. DECEMBER 16.1992

    ~ ACNW .16 _

1 l  ;-

   "*'3                             eg         h                     4-    -    3                      1 g       a"w-TP-

l; y .o o; - k IPA ORGANIZATION , MANAGEMENT BOARD M. FEDERUNE. MMSS M. SILBER 9 ERG, RES B. SAGAR, CNWRA TECHNICAL-COORDINATION

                                                                                                          ~

N. EISENBERG, NMSS T.McCARTIN RES R. BACA, CNWRA j' Team 1 Team 2 .- . Team 3 Team 4 Team 5 Team 6 . i System Code Scenerlo Analysis Flow ent! ' Source Term Disruptuo Sens. and ' '

                                                         - Transport                                Cm z _ _ . -                 Uncert. Ant.

R M CNWRA N NB8SS ' J Ai.TINf,CNWRA T ted:AfmN.MES EMiel8 ERG,98088 000EU.10 MSS , - ACNW WORKMG GROUP MEETNG ON PERFORMANCE ASSESSMENT DfV!SION OF HIGH4EVEL WASTE MANAGEMENT

                                                                                                                                       - DECEMBER 18,1992 ACNW'        _
                             .-               -                                                      -. .                        - -        ,               .a

ll COMPONENiS OF TOTAL SYSTEM PERFORMANCE ASSESSMENT: PHASE 1 MINED GEOLOGIC REPOstTORY SYSTEM DESCRIPTION ENGW8EEMED I BARRIER "e=yg

                                                   #"        SYSTEM r

1 r SCENAtHO -  ;  ; CONSEQUEEE ANALYSIS ANALYSIS

• DESCIUPTION

                  '

• SCREENING MRCE TERM e PROBAttuTIES PERFORMANCE CALCULATION FLOW & TRANSPORT

                                         *CCDF

• SYSTEM CODE

                      ~

I _ l _ SENSITIVITY & COMPARlSON - UNCERTAINTY *

TO REGut.ATORY STANDARO  : ANALY*s48 ACNW WORKNG GROUP MEETNG ON PERFORMANCE ASSESSMENT ONIS80N OF HNLEVEL WASTE MANAGEMENT DECEMBER 16.1992 ACNW'- 18 k

MINED GEOLOGIC REPOSITORY SYSTEM ENC PARAMETER WASTE SITE 310 SPHERE ('s s FORM R SAMPLING l SYSTEM (LHS) 1f V V - SCEfMRIO ANALYGls OONSEQUENCE MODEUNG

• IDEPRIFICATIOPV SV1 < SOURCE TERM ENUMERATION SV2 UQUID l GAS l DIRECT
• SCRE.'.NING SV3 4 FLOW
• PROBABluTY GROUNDWATER l GAS ESTIMATION i SV4 -< TRANSPORT UQUlO l GAS BIOSPHERE TRANSPORT SV5 < DOSE TO MAN (3 CUMMULATIVE *t-V r CCDF

                                                     ! ',                            RELEASE            -q
                        ^

CONSTRUCTION ,

                                                                                          /

( \ y V Y AUXILIARY m COMPARISON TO SENSITIVITY & ANAYLSES ^ REGULATORY 4- UNCERTAINTY < STANDARD , ANALYSES V TO BCENARIO ANALYSIS TO CONSEQUENCE MOD 3UNG SVi- Surrogate for Waste Packago Ufetirne SV2 - Surrogate for Fractional Re6 ease Rate SV3 Surrogate for GWTT SV4 - Su.Togate for Concentratlon in groundwater SV5- Surrogata for IndMdual dose O Phase 2 Components of Total System Performance Assessment - ACNW WORKING GROUP MEETING ON PERFORMANCE ASSESSMENT ACNW - 19 orvision or nics trytt wAsrc uAnActucuT

• DECEMBER 16.1992

AUXILIARY ANALYSES IN IPA PHASE 2

      =   Evaluctions of DCM3D and NEFTRAN 11
      =   Effects of Layering, Faulting and Dipping on Fluid Flow
      =   Review U.S. Geological Survey Regional Modeling Approach
      =   Saturated Regional Modeling
      =   Applicability of Ko Approximations
      =   Retardation of "C in Geosphere
      =   Colloid Formation Modeling
      =   Source-term for Magmatic Events
      =   Comparison of Sensitivity and Uncertainty Methods ACNW WORKING GROUP MEETING ON PERFsRMANCE ASSESSMENT OfVISION OF H6GH-LEVEL WASTE MANAGEMENT DECEMBER 18,1992 ACNW - 20 0                                           0                                                              0

y 9

                                       ;9
                                                          ! ENHANCEMENTS ADDED' DURING IPA PHASE'2 GL   " Internal Mode" system code

. G- Dose assessm'ent capability added

                                      'e    Evaluation of.SNL Scenario Methodology G   '16 scenario classes e    More' refined modeling of flow and transport in unsaturated fractured rock
                                      .O    Gaseous transport pathway added S    Saturated zone transport added ACNW WORKWG GROUP MEETING ON P'at+0HMAreti ASSESSMENT DIVISION OF HO4-JVR WASTE MANAGEMENT '

DECEMBER 16,'1992 a ACNW - 21. i

                                                                                                                 <w -

ENHANCEMENTS ADDED DURING IPA PHASE 2 (continued) 9 Improved treatment of source term o Mechanistic waste package failure l o improved waste dissolution / transport model o Gaseous phase source term 9 Treatment of several disruptive scenario classes 9 " Turn-key" sensitivity and uncertainty analysis ) l 9 Higher spatial resolution: 7 repository regions and corresponding geosphere transport legs l l- , l l ACNW WORKNG GROUP MEETNG ON PERFORMANCE ASSESSMENT OfVWON OF HIGHtEVn. WASTE MANAGDAENT DECEMBER 16,1992 ACNW - 22 O O O

~ O O O IPA PHASE 2: COMMUNICATION OF RESULTS

         -First:    .NUREG-1464 (in preparation)

Second:. NUREG/CR's: e.g. User's Manuals (TPA, SOTEC); Auxiliary Ana!yses (Effects of

                    -Dipping, Faulting, etc.; Comparison of Uncertainty & Sensitivity Analysis Methods; Saturated Flow; Evaluation of DCM3D)

Third: Professional Papers: e.g. " Carbon, Magmatic Probability

         - Fourth: Meetings: NRC/ DOE Technical Exchanges, ACNW, NWTRB, International High-Level Radioactive Waste Management Conference These various means of communication provide'for peer review and critical evaluation of results.

ACf#J ERKPgG GROUP MEETP4G ON PERFOfLMAhtCE ASSESSMEN*

                                                                                       . OfvtStO*4 OF HtGH-LEVEL WASTE MANAGEA*ENT DECEMBER 16,1992 ACNW.- 23

u . . . i s THINGS TO BE DONE TO COMPLETE IPA PHASE 2

       =        Continue Testing all Modules in TPA
       =     . Teat Sensitivity ' cod Uncertainty Analysis " Turn-key" System
                                                                                                                                                                                                                           .i a        Document Auxiliary Analyses-
       =        Conduct Production Runs
       =        Conduct Sensitivity and Uncertainty Analysis l       =-       Document.Results -- NUREG-1464 plus other Reports and Papers h

ACNW WO?iKWG GROUP MEETNG ON PERFOAMANCE ASSESSMENT-DNISIC*i OF HKJi-1EVEL WASTE MANAGEMENT - DECEMBER 16.1992 ACNW - 24 h h .,k

                                                                                                                         -w<a
                                                                                                                           -   m.. -.-- . ,... -- ~,
                                                               .a-a:-...--- ...- .<.-.,, .-< . --.--r ........ i ... .
          .......u.in.
           -- .        i--i,...r..,,..c-

O O O Overview of NRC's Performance Assessment Activities . During IPA Phase 2: Changes in (and contributions to) the IPA Methodology from. Research and Other HLW Investigations Rex G. Wescott, Senior Hydrologist Hydrologic Transport Section Hydrology and' Systems Performance Branch

                                ' Division of High-Level Waste Management 301/504-2167
                                                                ^'""'" "" "7Js"JI"" c""Ja"""Ei',OH**"f0 DECEMBER 10,1992 ACNW - 25

CHANGES IN.(AND CONTRIBUTIONS TO) THE IPA METHODOLOGY RESULTING FROM RESEARCH AND OTHER HLW INVESTIGATIONS CONTRACTORS CONTRIBUTING TO IPA

• SANDIA' NATIONAL LABORATORIES o DCM3D,1991 o NEFTRAN 11,1991 ,

o SCENARIO METHODOLOGY,1990

• CENTER FOR NUCLEAR tNASTE REGULATORY ANALYSES

                            - o GEOCHEMISTRY RESEARCH PROJECT

o' INTEGRATED WASTE PACKAGE EXPERIMENT o STOCHASTIC RESEARCH PROJECT o PERFORMANCE' ASSESSMENT RESEARCH PROJECT.

                                                                     ~~" 7Js"o' fos $t8a7EllSEEE[r
                                                                                           - DECEMBER 16.1992 ACNW - 26 9.

O O.

   -_ . . _ _ _ '     -  'e ( _         1-'-                                             m

10 O Ol CONTRACTORS CONTRIBUTING TO INDIRECTLY TO IPA

• LAWRENCE BERKELEY LABORATORY o TOUGH,1987
• UNIVERSITY OF ARIZONA o APACHE LEAP TUFF STUDIES ACNW WORKNG GROUP MEETING ON PERTORMANCE ASSESSMENT .

Ofv!SION OF HIGH-LEVEL WASTE MANAGEMENT DECEMBER 16,1992 ACNW SANDIA' NATIONAL LABORATORIES

• DCM3D o A dual continuum,3D ground-water flow code for unsaturat'ed, fractured, porous media o . Represents fractures an'd matrix as two separate continua interconnected through a transfer term o Used in IPA Phase 2:

to determine matrix and fracture flow in the unsaturated zone by a series of simulations

               --    in an. auxiliary analysis to examine fracture and matrix flow in detail

• NEFTRAN 11 o ' Network Flow and Transport in Time Dependent Velocity Fields i

o ' One dimensional network model

j. o Transport of. decay chains of radionuclides ACNW WORKWG GROUP MEETING ON PEfEORMANCE ASSESSMENT-OfVtS10N OF HIGl+ LEVEL WASTE MANAGEMENT -

DECEMBER 16.1992

     ,   ' ACNW 128.

y o o; SANDIA NATIONAL LABORATOR!ES

• NEFTRAN 11- (continued)-

o Time varying flow fields (improvement over NEFTRAN)

               -o    Used in' IPA phase 2:
                     - for. transport of radionuclides using velocities. calculated by' DCM3D
                     - for calculating time integrated discharge (EPA release limits) and concentrations
                   - (dose) .

• SANDIA SCENARIO METHODOLOGY-o Modified' version used'in IPA phase 2 o . Screening of events and processes-o Scenario construction methodology ]

o Methodology for estimation 'of,scensrio; probabilities 1 ACNW WORKNG GRO'P MEETNG ON PERFORMANCE ASSESSMENT ' OTVISION OF HIGH-LEVEL WASTE MANAGEMENT -

                                                                                                                   . DECEMBER 16,1992 ACNW - 29.

l t

CENTER FOR NUCLEAR WASTE REGULATORY ANALYSES

• GEOCHEMISTRY RESEARCH PROJECT o Provided insights used in the auxiliary analysis related to "C transport
• INTEGRATED WASTE PROJECT EXPERIMENT o Experimental results were basis for crevice corrosion algorithm in IPA source term module (SOTEC)
• STOCHASTIC RESEARCH PROJECT o Refinement of BIGFLOW code o Used in an auxiliary analysis related to the effects of tilting stratigraphy and presence of a fault zone un groundwater velocity vectors in the unsaturated zone PERFORMANCE ASSESSMENT RESEARCH PROJECT o Refinement of PORFLOW code o Used in an auxiliary analysis related to regional modeling of the saturated zone ACNW WORKWG GROUP MEETNG ON PEMOfWAANCE ASSESSMENT OfVtSON OF HIGH-LEVEL WASTE MANAGEMENT OtCEMBER 16,1992 ACNW - 30 0 0 0

m che p4 / / NNNg <.. X:@ IMAGE EVALUATION t -  ;, 4' g# O l 7/ [9@',N #g

  /'    v  '      4                     TEST TARGET (MT-3)                                         s 3:         r, %

\///r/ - -

                                                                                              %     5l    k      '?

47 v

    \h                 %@\

f k 1 1.0 "2 i]:

9' U2 E i ~. p 2.0 lj uma W l.8 ll!km I.25 W i.6

                                                  -        l ._1.4_iu     =

• 150mm - >

4 -- 6" x+,"%f 4 's /y s /,*x A k %),,,y <g, g e4; .

                                                                                   ~
                                                                                                   ,1 Q f Op             >

y (- y a x ;- _ pgl QQ ik f[3 A - 8t%\ 4 ,, ,

A yn,% //.O

        ,c                                                                                          /

r. e q O* , ob f

        'tt,   ~

g IMAGE EV Al.U ATION /gj g e, f,ff)' (g [o//p i Qf TEST TARGET (MT-3) 0 '9 J [f

    '\# $sk(f@                                                                                  //g         p
        #                                                                                              'k I.0     F ", "'    S'
                                                           ;g m

u= l,l t tu EL2aL == yL=ts e F l.25 1.4 i.6

                                                           -       j[imm

• 150mm >

d 6" - > g%  %$p sAo if

 %ams hp 3 '

Oy [

                                                                                     ~

4%+ f;

.1

                                                                                                'klpsy q k

go i

f. .

                                                                                ~
                                                                                       -j

[".'~, ' -

                                                                           < tOlO*WLdN ,.. ,,,, ,,,
                                                                            ~
                                                                                                                 .,,.,,.C
 -       _____               _1

m w ___

                       'S                                                                                                 /
                     ,1 *:
              *'                                                                                                       /

\

             ,,   w     . 4 '
                                                                                                                                   ,t ~ %

IMAGE EV Al.U ATION

     ,[O/ 't [^.'M
                   ,             g TEST TARGET (MT-3)
                                                                                                                       </
                                                                                                                                ,6        %        pj
                                                                                                                             .c
                                     %                                                                               3
                                                                                                                                   ~g y;se N.\                     '97                                                                                                           ~

ig) 4[

              *t$9                                                                                                          4g 1.0      m 42.=__
                                                                                ? >.

nq e ss lj .. SJ1 m I.25 I W i.6

                                                                    ===       !!le.4 L e        we 4          -                       -

150mm >

                          <__          .__ __ _ ___..- -- -                     6"                                              >

dh 9% \ 4/

         'Ygy Q,:nU 4
                                           >;-                                                                  +<p                                   '

am% ////4o s m ;O Q y N..' . sjjk

                                                     ~

o

                                                                                            -e    5
                                                                                                    ' N AN - b                           ..

L

                       '                                                                                                    ~
    .-     - :i
                                                                                                                                 ;9
                . Q) -                                                        ;                                                 Q ,I-

1. AWRENCE BERKELEY LABORATORY -

t

                *'     ' TOUGH o ' Code for transport of unsaturated groundwater and heat o  Developed under funds from DOE geothermal o User's manual funded by NRC o Used.to.model University of Arizona core heating experiments f.; L*s! TRAVAL o Provided limited cunfirmation of gas flow modeling assumptions o  Limited attempt was made in IPA Phase 2 to analyze heat induced moisture redistribution o  Use expected in. future IPA phases to analyze near field drying and resaturation and' verify aspects of gas transport models ACNW WORKt4G GROUP MEETr4G ON PERFORMANCE ASSESSMENT DIVtSION OF HIGH-LEVEL WASTE MANAGEMENT DECEMBER 16,1992 ACNW - 31
 .w      3          e                            . _ _ . . . -. . . _ ,                                                   -             -

                       .----._-.c..,.                                     -                                              __ __

UNIVERSITY OF ARIZONA

• APACHE LEAP TUFF STUDIES o IPA contribution expected from data collection from the inclined borehole experiment ,

and the pneumatic studies in the multiple borehole field site. o The CNWRA will use the data for: l

            - development of alternative conceptual models
            - investigation of validation strategies ACNW WORKPJG GROUP MEETirJG OfJ PEREOPMANCE ASSES 9*ENT DivtSION OF HIGH LEVEL WASTE MANAGEMENT OECE VRER 10.1992 ACNW - 32 O                                                            @

O

~ O O Ol Overview of NRC's Performance Assessment Activities During IPA Phase 2: Changes in the HLW Program Resulting from the IPA Process

                                    **~~""'=====;r,==";

DECEMBER 16.1992 ACNW - 33

a i CHANGES IN THE HLW PROGRAM RESULTING FROM THE IPA PROCESS: j SOME EXAMPLES l Development of SOTEC in IPA Phase 2 led to significant changes in the development of. - EBSPAC ' 4

• Development'of SOTEC led to changes in the IWPE program in regard to modeling pitting  :

4 ard nrevice corrosion  ! Based on IPA Phase 2 activities, it was decided that global climate research should have a' .i

lower priority than research..in volcanism and tectonics

            *      " Carbon transport analysis provided insights which were useful in interacting with EPA and DOE l
            *     . Problems in modeling sorption provided incentive for the SORPTION research project                                ,

The need to define realistic. failure modes ~of waste packages under seismic loading has led .;

l. . to the inclusion of structural analysis under repeated loading into CNWRA RDCO Task 3 i

' 1

• Difficulties in running a total system performance code within a reasonable expenditure.of: 1 3 ' computer time led to research on improving the efficiency of computational models under' j the'CNWRA PA Research Project ACNW WORKING GROUP MEETPJG ON PEnrORMANCE ASSESSMENT DIVISION OF HtGMEVEL W MANATI T ACNW - 34 e O O

O O O Overview of NRC's Performance Assessment Activities During IPA Pha.se 2: Preliminary " Lessons Learned" for IPA Phase 2  ; Norman A. Eisenberg, Senior Operations Analyst u L Repository Performance Assessment Section Hydrology and Systems Performance Branch Division of High-Level Waste Management 301/504-2324

                                                   ^""**"""? s%" ?:D"?" L'L%;

DECEMBER 16.1992 ACNW - 35

PRELIMINARY " LESSONS LEARNED" FOR IPA PHASE 2

   =   These " lessons learned" are based on work in-progress -- they are tentative!
   =   These " lessons-learned" are derived from developing methods and models in Phase 2.
   =   Additional tentative conclusions expected from production runs of the system code:

o system performance o methods o models

   =   Conclusions about system performance, as in Phase 1, are limited by:

o the current lack of mature site characterization data o the use of models and codes which are only partially tested and supported o limited peer review ACNW WORKPJG GROUP MEETFJG ON PERFORMANCE ASSESSMENT OfVISON OF H8GH-LEVEL WASTE MANAGEMENT DECEMBER 16.1992 ACNW - 36 9 O O

~ O O O LESSONS LEARNED -- OVERALL , i

1. Computer Needs:

a. Need more effort to simplify and to speed up computational modules. l

b. Need more effort and time for code planning and development.

c Software quality assurance (SOA) o . Structured programming a Data control and documentation re Dedicated staff is required for code development and maintenane: this staff must work close!y with scientific and technical analysts.

c. Phase 2 computer facilities inadequate -- computer upgrade expected to rectify (currently approved & being implemented).

ACNW WORKriG GROUP MEETNG ON PERrON1Af4CE ASSESSMENT DNtSON OF Hei--LEVEL WASTE MANAGEMENT DECEMBER 16,1992 ACNW - 37

LESSONS LEARNED -- OVERALL (continued)

1. Computer Needs:

d. Rigid specification for code requirements will be necessary.

       =    Motivation:

o Expanded effort o Higher complexity a More players, o Examples:

             - Variables required by succeeding modules
              - Format of data transfers among modules                                                                 ,

ACNW WORKING GROUP MEET!NG ON PERFORMANCE ASSESSMENT DIVISION OF HIGH-LEVEL WASTE MANAGEMENT

                                               .                                                      DECEMBER 16.1992 ACNW - 38 0                                            O                                                             O
                                                                                                                       .-..a

O O O LESSONS LEARNED -- OVERALL (continued)

2. Transportsbility of Computer Codes:

       =   Code performance varied significantly when codes were moved to different computers and different computer environments.

o not planned for in Phase 2 o not prevented by use of standard programming language

3. " Carbon:

a. Difficult and costly to model.

b. 'More modeling realism will be even mcre costly.

c. May be a problem for demonstrating compliance with EPA and NRC regulations.

4. Waste Package Design:

       =   There does not appear to be any analysis of post-closure repository performance for seismic' disturbances.

ACNW WORKP4G GROUP MEETPdG ON PERFOfB4AhfCE ASSESSMENT DN!S3ON OF HIGH LEVEL WASTE MANAGEMENT DECEMBER 16.1992 ACNW - 39

                                                                                                                    . .                               v A
                                                                                                                                             ~ i --

l LESSONS LEARNED -- OVERALL (continued) .

i. 5. Technical Coordination:-

                      =   ' Coordination of the various technical activities is a full-time activity (at least).                                            "
                          'o     IPA is;. inherently interdisciplinary 4

o Exacerbated by separation by distance of NMSS, RES, and CNWRA

           -6.     . Project Management:

a. -Improved strategy for management of this type of long-term activity:

4 o Closer tracking of resource expenditures vs. work accomplished

o Closer'. tracking of work accomplished vs. schedule

                     -b.  . Closer interaction and coordination of IPA with other technical and regulatory activities i                           has been beneficial,'but has extended schedule.

i l' ACNW WORKMG GOUP MEETING ON PERFORMANCE ASSESSMENT . f 8- OfvtS80N OF HIGMEVEL WASTE MANAGEMENT - DECEMBER 16.1992i , ACNW 40 5 _ - .__ ._ = . ._ - . _. . ._ . a

      -O                                              O                                                                                              OL LESSONS LEARNED -- BY TASK Task 1 -- System Code:

1. A complex code; further efforts require more effort on:

a. SOA

b. Computational efficiency

c. The process of simplifying complex codes

d. Th'e design of the system code. ,

2. Additional support staff will be required to develop and maintain the' system code effectively.

3. Overall design an'd structure of the system code may need revision to treat initiation of

various scenarios during the performance period.

ACNW WOREP4G GROL*P MEE7PdG ON PERF08WAANCE ASSESSMENT DfViseON OF HIGl+ LEVEL WASTE MANAGEMENT DECEMBER 16.1992-ACNW - 41

LESSONS LEARNED -- BY TASK (continued)

Task 2 - Scenario Methodology:

1. SNL methodology needed modification to work.

= Limitation' on nature of events considered - external only.

           =    Eventification of processes, e.g. climate change.

2. Sampling certain variables will ch'ange the probability of occurrence of a scenario - points to effectiveness.of WIPP approach.

3. Evaluation of scenario' probabilities is deemed inadequate.

                                                                                                                               ]

l = One problem is the dearth of site-specific' data.

           =    Another problem is that more precise estimates of probabilities requires advances in t

the state-of-the-art of earth and planetary' sciences (e.g. an enhanced understanding of the causes of magmatism in the basin and range j l and the relationship of tectonic forces, magma genesis, and control by geologic - structures.)  ; ACNW WORKNG GRO'JP MEETNG ON PERFORMANCE ASSESSMENT - , DfVtS80N OF H6GHLEVEL WASTE MANAGEMENT .

                                                                                                          . DECEMBER 16,1992 ACNW - 42.

l l . e to 10

LOL O QL q 1 LESSONS LEARNED -- BY TASK (continued) Task'3 -- Flow and Transport:

1. . -

is the flow paradigm acceptable?

              =    Real problem:

o Transient o dimensional a: .. Partially' saturated flow - o: . Potentially significant air and. water vapor movement o - ; Fractured porous medium ,

              =   Which simplifying assumptions provide an adequate representation of. transport?
     .2.

Thermal effects in the.early phase of repository development could be profound.

              = This problem and its analytical methods difficult.
              = Substantial effort and additional code development may be required.                                                         ,
                     ,-                                                                                                                'i
                                                                          - ACNW WCRKr4G GROUP MEEirdG ON FERFORMANCE ASSESSMENT -        ,

CIVIS10N OF filG+ LEVEL WASTE MANAGEMENT - DECEMBER 16,1997: ACNW - 43  ;

                                                                                                                                      's   l

LESSONS LEARNED -- BY TASK Task 3 -- Flow and Transport:

3. Changes in parameters and boundary conditions due to initiation of previously inactive scenarios could be very important.

    =   " Restart" capability of SNL flow and transport codes not used in Phase 2.
    =   May be difficult to address.

4. The current gas flow analysis makes simplifying assumptions which need further evaluation (e.g. simplified stratigraphy, 2-D, not two-phase fiaw).

5. Even the simplified steady-state, two dimensional solution of unsaturated fracture-flow problem is computationally intensive and required simplification for system code use.

6. New sets of technical issues arose from the need to calculate concentration.

ACNW WORKING GROtJP MEETING ON PERFORMANCE ASSESSMENT OfVIS10N OF HIGH-LEVEL WASTE MANAGEMENT DECEMBE R 16.1992 ACN'N - 44 9 O O

r. .

Of

              .O                                                    O LESSONS LEARNED - 'BY TASK icontinued)
        . Task 4 -- Source Termt
         ' 1. -

Source, term code design made difficult because releases occur on different' spatial scales. (e.g.. release of gases on a' drift scale versus release in liquid phase on a package scale). !~

2. Mechanical and seismic waste package failure requires more study.

                      =   Includes coupling of mechanical failure m' odes with degradation by corrosion.

3. . Source ferm models 'and codes'are'atmost as. complex as the system code.

                      =   Near-field and heating effects 'make prob'.em harder.
                      =-  Simplification of, more co'mplex models-and codes is a substantial' effort.
                      =   EBSPAC program recently refocussed for this purpose.

] f'.

                                                                                  ' ACNW WORKW4G GROUP MEETING ON PERFORMANCE ASSESSMENT DNtSKPd Or HIGH EVEL WASTE MANAGEMD4T -

DECEMBER 16,1992~

        , - ACNW. '45 '

e- c- y , +r o , e r e, - + 1 v c 4-+-< ,-vw- .'-s

                                                                                                                                    ~
                                                                                                                                          ;w    ,

1'i 1

LESSONS LEARNED - BY TASKL(continued)

TasV4 ~-- Source Term:

4. - Technical areas requiring.more focus in future work include:

a. How water gets into the waste package, if.at all.

b. How. water contacts the waste within the waste package.

c. What near-field hydrologic processes produce wetting.

1

                                                                                                                                                 'j ACNW WORKNG GROUP MEETNG ON PERF0F94ANCe, hggggggggy :           .

DfVISION OF HIGH-LEVEL WASTE MANAGEMENT

                                                                                                                         ~ DECEMBCR 16.1912 i-            ACNW '- 46 i

e

l. .

e. .e

   , . .                                      e , . . . ,  ,         s              , , _ . + .             n.       . _        .      ,,

' ~ O O O LESSONS LEARNED -- BY TASK (continued) Task 5 -- Disruptive Consequences:

1. The interactions and coupling among the various disruptive consequence models and codes is important.

           = Driven by fundamental technical concerns, such as conservation laws.
           = Must be considered at a high level and factored into:

o .the architecture of the system code a the " base case" consequence module chain a the various disruptive consequence modules

2. Disruptive consequence modeling is very labor intensive because:

a. A wide variety of different outcomes must be modeled (e.g. intrusive and extrusive magmatism).

b. Coupling among several different phenomena and/or events is required.

c. Good data, especially for rare events, may be difficult to obtain.

d. Some very energetic disruptive events are inherently difficult to describe mechanistically with outputs useful in PA. <

ACNW WO6KNG GROUP MEETr4G ON PERFORMANCE ASSESSM&JT OTtSt0N OF H'GH-tEVEL WASTE MANAGUAENT DECEMBER 10,1992 ACNW - 47 _ _ _ _ _ _ _ - _ - _ _ _ = _ _ _

9 r x - J. A LESSONS -LEARNED -- BY TASK (continued) Task .6 - Sensitivity and Uncertainty Analysas L . .

1. Sensitivity and uncertainty analysis methods currently available appear to be adequate for  !

the 'needs of performance assessment. But, some issues remain: 4 . .

a. Evaluate additional methods.

b. ' Use o.ther sampling schemes (e.g. FPPA, importance sampl:ng).

4

                          - c. Evaluate adequacy of step-wise regression for PA'with variables ranging over several orders of magnitude and operatin'g within highly nonlinear mo'dels.

2. Need to develop methods for determining and using variables synthesized from aggregates

                          ~of random' variables.

l- = Some-sampled variables representing stochastic occurrences provide little or no insight' 1 -.into system perfonnance.

                           =      These variables are sampled to provide random inputs to various consequen'ce models                                         '

i (e.g. location of exploratory boreholes). L l

                          -=      :Thp      elementary variable's ma'y be synthesized into other variables (e.g. the centroid i                                   of '.ne exploratory boreholes),' which may provide insights inio system performance.-

ACNW WORKING GROUP MEETrJG ON PEP 50feAANCE ASSESSMENT .[ j DfVISION OF HG441 VEL WASTE MANAGEMENT: DECEMBER 16,1992q ji  ; ACNW -- 48..: 4- . i j 'I  :.

O - O LCr Detailed Example of the IPA Process in a Specific Discipline: '4 Carbon Richard B. Codell, Senior Hydraulic Engineer Hydrologic Transport Section Hydrology and Systems Performance Branch Division of High-Level Waste Management 301/504-2408 ,

                     ' Rex G. Wescott, Senior Hydrologist Hydrologic Transport Section Hydrology and Systems: Performance Branch Division of High-Level Waste Management 301/504-2167 l                                                 ^'""'"*""'"O5"J%2EJuTflu^211"oM DECEMBER 16,1992 i

ACNW .49 i

r

                                                                                                                                  .}.

PURPOSE OF PRESENTATION: A Comprehensive Demonstration of an' IPA Phase 2 Process L . l-

      -*      Source term'for C l      -*      Gas flow through Yucca Mountain

• Trarisport of. C l

L * - Auxiliary Analysis in support of transport assumptions

• Dose r,ssessment and cumulative releases
• impact of IPA on regulatory process

                                                                      ; ACNW WOAKNG GROUP MEETNG ON PEPIFOfWAANCE ASSESSMENT DMS!ON OF HIGREVEL WASTE MANAGEMENT ..
                                                                                                               ;DECDABER16,1992
      , ACNW: 50 -

$i

i. .

gf

  ~ ~
            !O                                            O                                                                    .01             ~

14 CARBON RELEASES FROM YUCCA MOUNTAIN REPOSITORY

            *   "C is associated with spent fuel

-

• Oxidized to carbon dioxide gas
• Transported as gas in unsaturated rock
• Realistic assessment of impacts of "C requires coupling of source release, heat transfer, fluid flow, and distribution-of carbon among solid, liquid, and gas phases o Assessment of dose. impacts from'"C ,

t i t.

                                                                                                                                                   ..3 ACNW WORKING GROUP MEETING ON PERFOPMANCE ASSESSMENT '

OfVtSION OF H1GRLEVEL WASTE MANAGEMENT DECEMBER 16,1992 ! ACNW - 51 r l

                                                                                                             .___:-______-__=____-_- - _ - - .

LOCATION OF RADIONUCLIDES IN SPENT FUEL AND POTENTIAL RELEASES OF "C (from Apted et al.,1989)

                                                          =

TappedVxas Cracks (Gap)

                                                                                             \

(Q /hc l l l I. 3

                                                                                            .s
                                                                                                           .x-
                                                                                                            . N.-

i l m must l l L . ). I

                                                             ~

i, .. > ,

                                                             -             l                                  /       .

I i

                                                                                      -     .   .             i.               !
                                                             ~

I h/. i

                                                                         / Gap /            \

Grart mounomy

                                                                                                                \

UO MsMr

                                                                                                                             \

W _ Oud "C "C "C Acstdoes "C

                                                              -              *i                 *t
                                                       ~-
                                                           ~-
                                                                               ~,..                 ,.
                                                                               "1t:               "11:

S0 g 90 ,r ACNW WORKWG GROUP MIETING ON I'iRFOFWAANCE AsstSSMENT . 2 OfVISION OF HG+tEVEL W MmAGEMENT ACNW - 52

                        .e                                                                         e                                                                      .e.

n y to: o. - RELEASES OF C FROM SPENT FUEL e . Initial cladding oxide and crud

         '-*. Grain' boundary and cladding gap e  . Zircaloy oxidation

• Oxidation of' fuel-i i

T t

ACNW WORKMG GROUP MEETNG ON PERFORMANCE ASSESSMENT.

DIVISION OF HG+1EVEL WASTE MANAGEMENT.;. DECEMBER 16,1992-TACNW - 53.l-b p-_cg e' u e # y

                  /

ADJUSTED'4 CARBON CONTENT IN SPENT FUEL (Ci/MTHM? (after Park,1992? TYPE BURNUP UO, ZlRCALOY HARDWARE TOTAL

                      ' Mwd /MTHM BWR       35,000. 0.69       0.48         0.13           1.3 PWR       40,000      0.73       0.22         0.26          1.21 Average      --        0.72       0.31         0.21          1.24
                                                     ~"*" 78$fc7nE[E' tE"$5'O"tu  o     d DECEMBER 16.1992 ACNW.- 54 l

O O O

O O O MODEL FOR RELEASE OF " CARBON

• Calculations demonstrate that:

f o '#C released quickly from grain boundaries, c.ladding/ fuel gap, and initial zirconium oxide o Minor releases from oxidation of cladding and other metals o Major releasable inventory from UO 2 l i i ACNW WORKF4G GROUP MEETrJG ON PERFOPMANCE ASSESSMENT DNtsaON OF HG4-LEVEL WASTE MANAGEVENT DECEMBER 16.1992 ACNW - 55

4

                      ' SHRINKING' CORE MODEL FOR FUEL OXIDATION s

Low conversion High conversion Oxide Layer Unoxided Fuel v v TR

                                             ~'

Tm, _. ((

                                                                      .       ~1                    9
                                      ~'"

i I -. j 1,Gk l i II I l l 'l 1 Reaction i i l ll l' ll l zone 8 I ll 11 g g I i I i I: 1 -i ll

                                                                                                                                                           ~

1 1 I I  : ll l l

                                                            ^                   I         ^ l       l i A 11                      11           I  I       I I             I 1

I 1 11 1 I I li ,

                            -ll l                    ll         ::   1       I l-I l_t          l I                      l          l               l            ,                   i l                     i          l               I            g.                                                                 ~!
                            -l 1                     1          l               1            3                   !

I I I I l' l

                            'l l'         l               I            i                   I L            i       J    m       I       i       ! -          e        i              -
                           'R             0         R.        R.        O      R            .R      0           R
                                                                      ~

Radial position l' l ACNW WORKNG GROUP Mt ETWG ON PERFOfWIANCE ASSESSMENT ~ OfVtssON OF IDGH LEVEL WASTE MANAGEMENT : l" DECEMBER 16,1992

      --. ACNW - 56 '
                                                                                                                                                                .. {

O O O

                          " CARBON GASEOUS RELEASE MODEL s

oxygen fuel .

                                                    ?CO,             >

oxide

                                             ~

grain boundary layer

                                                        ^'"* * "*""?Js"J"l!"a =f=;TO =;"

DECEMBER 16,1992 ACNW - 57

FUEL OXIDATION MODEL ASSUMPTIONS: o No oxidation until canister fails

• No protection of fuel by cladding e Oxygen diffuses througi1 two layers:

o outer layer representing grain boundaries o inner layer representing oxidized fuei e Oxide is U3 07 stoichiometrically l e Oxygen concentration zero at inner boundary e Oxygen concentration in layers are at steady state l ACNW WORKWG GROUP MEETNG ON PEPIOPMANCE ASSESSMENT DNISION OF MfGHtEVEL WASTE MINAGEMINT DECEMBrR 18.1992 ACNW - 58 e o e

                                       ==mmmmmmmmmmmmmmmmmerummmmmmmumusammmmu

" ~

O O O 0

                           ' PARAMETER IDENTIFICATION AND MODEL VERIFICATION

• Grain diameter and oxide' growth from micrographs of fuel (20 microns)  ;

4

• - Outer layer diameter taken as fragment size (2mm) l'

• Weight gain from thermal gravimetric analysis and dry bath experiments between 110 C and 250 C
• Activation energy and diffusim coefficients adjusted for best fit to oxidation data on fuel fragments in 6. temperature ranges
• Little data found on "C releases from fuel oxidation r

4 i 1 + I p-ACNW WofENG GROUP MEETNG ON PEftFORMANCE ASSESGMENT i' OfVISION OF HKMtEVEL WASTE MANAGEMENT occt. m te.tas

      ~ACNW                                                                                                                                    t I

i n - y y , e re w w c- r--- , - - - :r, v- --- -- -u ~-- - - - - - -<

i! > !

                                                                                                                                                                                              - E

_ # ^

                                                                                                                                           -                                                  I g n=g
                                                                                                                                               )       o I 2o k$
                                                                                                                                                                                            *E rs sh da o        o                3 5                 Bb o                                                         u*                     y ?,

o 0 0 e 1 P $ 0 0 5 4 0 0

                                       %                                                                                      n g                                        N     a y                                                                    v m-o                       .        5                ~m~o                    -

Q

                                                                                                      =        =                      n
                                                                                                                                      ,                                                E
                                                                                                                                                               .            v  31 8

00 .O e 0 j g e9 g :I - l . g,

                                                                                                                                                                                   \.             o w
                                                                                                        - - -                     .            4 O

5 ~ 3 5 4 9 l

   ,         >    .[ !       l-       -     3It,'                  ,;t     , :ll!!(          ,         .           , ' ,        ![      .;   t e

w w TT2 o NNs E

                                                                                                                           *J SE SGe.

EA S Na E 9 M1 t

• w S

A At e E MM CEt T c N St y A Ao A RW R OL F E i RV Et Pt . NH OG

• GH I

NF TO w E E N + l mao PI S e UVf O O d R w e S o G G N T m K R y C O W E n W F F i o N C A ** t E . s a - i N .di c O x o I -

                                                                                                                                               +

T' y A w

                                                                                             . o s       l v

D t e r

 -o            I w

e n u e y X 4 f - O n 9 s i T 0 d ,., N o e E o I o .d . S - - - - o u N = 5 1 0 1 5 0 l c A' 0 c 0 0 .n R eo33>su i T . t 3 o N 1 .

6. #

o W  % e N - C. A

                                                                                                                                                  +
                                                                                                                                               + -

- w n. L i! .^!! I , (. . 4  :!'  ;:

                                                        ?*

e  ::o .

;~ {lLiii ;1 i!k < 1 '; i

i t. o r 0 n i e _ ers

                                                                                           ,0
                                                                                              - 0 c                 r r                                     iema                                 0 i

w a - t e y = T T 2 NN9 E E 9 O_ f B i0 MM1 s l SE 6, L 0 nos t. SG 1

a. E A d r3 l e o mu u u

SNR S A AEB w.mr, e ' - t i s n0 r Co

                                                                                          ~J t

E CEE MM T C N SE ro, O e a _ A AD I r a0 et A aA taL[ e C0f 1 0. S AW P OL n 2 A+ = 0: 1 FEV tl i - 0A F mu;g

      - =
          #.;nf. C                                                                  -            0 E 0Y EE PL Ni
        = -    .
                                                                                 =                           OG
     <=*_E
      .= . ;

i t E

                                                                                                           . GH NF l

O m 0. M f E m i L - r 0l 0T MO O N o r A 0 4 P fS UN f P OD R E , G G e 0 N s I K i0 H a 0 O 0 W l e 2 W e  : N C A R C

                                                                                         ;0 0

0 0 0

                 =                                                  m
"*               o c

5 0 1 0 0 1 s 0 o MME* 47U gQ $E r e gy c

                                                                                           ,00 0

O i f es 0 r r _ C iamr . 0 a I J t e = B _. Y

                                 $                          i f

L0 ey 0

            -          -               s 3,

e0. d (f tis r0 =o 0 e - t re t e1 l o 0 1 r o s2G 1 e m. R. f g. r p i n0 re m e - LAs i o a0 C2 HA 0 i n g A 7PCA 7(A T Ll LL t *

                                                                                        ~00 :SA      1 n                              ;u*:                                                         0E
             ,                     . : * :                                                     GY E                       .

cc*

                                    .=*
                                       =*

m .._ 0. M E

                                                                  -                     7 00l        T or                                                         t i-4 0

ff// 1 f m . i L-e A-0 s n a P E-p 0 0 l e - 0 2 e fi , 2 6 R l - ~ -

 ~               -          -
                                                                  -                   '; -      0                   W O

' o o 0 N

                 =                                                                     o                            C 2

o e o 5 1 s

c. o o A M$gu 4 7O , g 5$

i ,

a o o i l- GAS FLOW AND " CARBON TRANSPORT MODEL

• Based on model of Ross et al (TGIF) i i- *- 2-D Finite difference model-

, o Quasi steady state; transient temperature field but steady gas flow i-i

                                                               *   - Equivalent porous medium' combines frEtaras and matrix c'   Single phase -- water-saturated gas only

! * "C transport by advection with retardation factor i I' iI 1 ACNW WORKNG GROUP MEETWG ON PERFO8E4ANCE ASSESSMENT OfvtS3ON OF HX334tEVEL WASTE MANAGEMENT DECEMBER 16.1992

                                                              - ACNW - 63 Y' g                   4'                                           [   'N              N -
                                                                                                 %' ' ' " "r tFre- N-- w"- 7T          'P*Y r--T-+ --

9T *M I E~*?**- m 9* '~vm't- --'--A E* tam- -M--** *

• J REPOSITORY CROSS SECTION FOR GAS FLOW MODEL J=;

J=tg J:16 _ TCw

                         ,P T9 .'
                                                      ~ % JT o p(g)

Tsw J=5g J=42 J:62 i=2 THETA =.10472 I=Nc.y ACNW WOMNG GROUP MIETNG ON PERFOWANCE ASSESSMENT Ofv1SION OF HG4(EVEL WASTE MANAGEMENT DECEMBER te.1932 ACNW - 64 O O O _ _ _ - -. ___m.__ .

i+'

O O O 1

REPRESENTATION OF REPOSITORY CROSS SECTION FOR HEAT TRANSFER-i 1 I

                                           .i:                            REPOSITORY FLUX                                            g g

I ~~ i l i I l- . t

                                              '----                                                   o                n T--      n     n            n          n 4                                                                     GEOTHERWAL FLUX l

ACNW WOPutMG GWNP MEETNG OP8 PENOWAMCE ASSESSMDsT DfV.SION OF **KMLEVEL WASTE MANAGEMEMT - DECEMBEM 16.1992 ACNW - 65..

                                    *4      g           .

n y-$f wr sa _ ~  % w- -

g. 6-- -,- n, - , - , , , -+m-,,---_4 , ,w - s tm.. m n ,

y b h E 3

=4.........A

            .     .  ...........o                                                                               b w

3 ..... ... 11 {.c Y

            *                                     ... ii                                                     ,5 g -

o. 4(

KB g m u . . . . 4. I j 7 m . . f e;u 4 l w . . . .. . . . . ..n:ll  : .

                                                                                                           . e, p

f O :5

l:I l

O ..r , . . , , A /l:f 1 . 1.0 2 O

                  ~'k*'" ll
                 .4h,A666                           6      .n*6 6 0

w e. ..rr g-M  : .. ~ ,i # '

      .g M        - >l
                                       - a                  ~.2, m                   f ..                          ..y                            -

O a

r. ..

                                                                          ,i g g             -,>...;9  ,
               . .M N

k. n

                                           ~ ~ ~ . .,                          ,
                                                                        'i, c

4 . . . s . - w; g :j q

                                    .. ... J p,j k                        zz
                                    - - -     4. - - :545
                                      . .             . - - s ,, ,                   '                                  w.
                                                                                                                      .c.

g [: 44 l;

 ,, ' 1                                                                                                                                       +

d'-ete ,..

DEFICIENCIES OF" CARBON FLOW AND TRANSPORT MODEL

• Does not include 2-phase flow
• 2-D Steady state only -- Cannot include third dimension or time
• Retardation coefficient may not adequately capture interaction of "C with water and rock ACNW WORKING GROUP MEETNG ON PENOPAtANCE ASSESSMENT '

OfVAYJN OF NG+1EVEL WASTE MANAGEMENT OECEMBER 16.1992 ACNW 67 L ..j.-- - -. . . . . . . . . . . . . . . . _ _ _ - . - - - - . _ __

TT2 NN5 EE9 t e MM1 SE 6. SG EA 1t S Nf O a S A AE B n E M+ A o CEE NTC br y A AD A SE a l c 5W c e i t 8 CL i b i v i s NE E V t l PU

               ,       a          a                            N+

e r e OO C t a i e t r r GH N5 ' n y o T 0 EN o r f E M80 R b r t s e r Pt Uv S O a c i m u A OO f F e t a G G ht h r n G i c e p t 9 N f w O n m W I L C o e V n E " br a d t O A D f o c n O l a n a , M o i t n i t n o c e T i t a s a R r e s e r u O t n O t r a P i o s S n f s , N i l a n wyr A p o o o x i t l f t R e a t s i T o u no t q ep L l e i ser e A d e c n ad C o n tr te I m a ara l M E h c l b a nu r s ot a H a e e i s s C s v nn O e l o eu r s inm E a , da G f o i t n en t r o p ni osn xe et t a h-c s ei o nw o a

                          'e.

di ud t 8 cd c n l 6 i n na t A icn o - V N 9 'e e C A

                                                                           .O Illt                        (Il           1i !lIL !ll l l.

l~O o O O l 4

i.  ;

L ESE*,NTIAL CARBON SYSTEM < L i CHEMICAL REACTIONS , i i- i aH aHCO3-i- H200) + CO2(aq) = H+ + HCO3- Encos-= g+ g4 (1)  : HCO3 co H+ + COS2-Ecome.= % aH+ acC$~ (2) i P aH20(D  ! H++OH o H2O Enzo(0 = 8a+ aou- (3)  ; P aH2O(0 H2O(gs a M M = fHzo(8)/ f H2O ( 1

                                                                                                                    ~

W

• WW " M /Pcon ,

t hM

CACO 3(ce) + H+ o Ca2+ + HCO3- Eoc = g) y (6) -

i

4.  ;

1 i J ACNW WOfENG GROUP MEETNG ON PEMORMANCE ASSESSMENT OfvtS80N Of HKiH(EVEL WASTE MANAGEMENT l

                                                                                                ' DECEMEER 16.1992 5 t

ACNW - 69 I

L EEE h gli w 33 35 FLOW AND TRANSPORT MODEL  !! 55 DEPTH SELOW SURFACE O~ VOLUME 0.0 Motors PO M  : CONSTANT

                 - 145 -                      TEMPERATURE                $

8OUNDARY N/  : LOCAL CHEn# CAL $ EQUIUSRIUM IN EACH 3 h/ REPRESENTATIVE 5 VoLUMa k UNIPORM GAS PLOW IN ONE DtWENSION h

            ) f-                              REPOSITORY:

480 Motors - - SS -  : NEATSOURCE AND Q

                                              "C SOURCE N  w /           GASINPUT
            ) f  -1    -

725 Moders y EN R 5. O

                                                                              *I O                                                                                               -

I jii o E! 150- Il-Es 140- I b 130- )'\ l' 120- - to years 110- h - 100 500

                                                                            'g y

g - 1000 '5< o 100- N

                                                .3000 90-         '

a 80- ' O g 70 -

  .j   60-b                              -

40- [ 30-20-10- ( 0- - 280 300 . 320 340 360 380 Temp.erature, Deg K p -Q k 4

                                                                                                     ......mm&

un I!5 0

                                                                                                   !<;s ps 11 is es se ex sa l

il i 18 i ix - 5

                                                                                            =
                                                                                            ~

g I o C,m

• x.. a:

                                                                                             ,2-                 e

_00 Ce NC-r C _O O ,, O 1 i i I i i ~i C O C O C O C O

n. n. N. N. . . O. O.

O O O O O O O O g ulo/affaeq!I 'xnId sed N b a 5 9

                                                                                                                                                                  -N A                                                                                                                                                         gzg b N'    -

yNU i 70 s a 140l' . 5 I ,,,, 140] ts 130 : 130 : g j 120l. 120i. o+

                              . g ,,,....e. :o ,... .

l t- E i i 110 *- i

                           - a re.                                                              110 q .                                                      t; g
         .3 100 -                                                                             ~1004            ... ;. i ...- s:o ... ,                       I 50         !
           -    90                                                                                     ~ ~ ~ " ' ' '                                       g

80 -

C 80 *) ' kn I

         ,     70 1                                                            ,            _ 70 t                                                              -

f- 60 d e0 4

         .d
         -     60 4 i%s
                             )              ',
                                                                                             -   5 04I,'                       '... -

w #p .. & f R 3 40 <:

                                                                                             40 4:                                    .

5

         #                                                              5                                                                                    <

30 ] i 30 j7 . 20 - 20 - 10 - ', 10 - 0 0 0.000 0.001 0.002 0 000 0 001 0 002 Moles C per Cell Moles C per Cell f~

    \

140-  : 140 - , 130 - 130- 1 T 120 - ,- 120 - f' 180 - f 110 - f l 100- . . . - ges

                                    ,,ua- 2o00 rwe                 ,.                    -o 100 - I' --       C .e.os - acoo , ,
                                                                                                            ...c o

90 -

                                                                                        'O 90 - i - C ***d'                                     I
        's                                                         -                    'E                                                     ,

o 80 - ,' S

• 80 -
• 70 - / 70 - I e > e /

        ,u,, 6 0 -                                        /                             ,o 60 -                                         l E    $0 -                                    I'                                 Y                                            l, o SO -

i ' v *

        > 40 -                                           ;                              '* 4 0 -

1 i 30 -  ; 30 - '1 20 -  ; 20 -  ; 10 -p: i 10 - I O ~ '. 0 d ', 0.002 0.000 0.001 0.002 0.000 0.001 Moles C per Cell M0les C per Cell Carbon,contant of gas, 11guld and solid phasee M 5 i i z y s 4 l 4 e

EEN

                                                                                                              --            h,!,-

8$" 4 u$a n 140J 140 4 $8 l  ! Is 1304 1304 5 t 120) 120) .;... 3:e ,,,, g

                                                                                                <p, i10                                                              110l         260         .' 2 '                     5

100 $ ~100-0g 3 - - . c'*

i e p us s ':0 ..r. i Eg 5 90 - $ 90 '.,, C

                                                                                                                       E
   .$ 80 a                                                          .$ 80 L   '                               .

_ 70 * . 70 ,.

                                                                                            - "*~~...

s s i O a 60 .. ' - '- v 80 - N

Q w 50 - )-- -

a',- --.........,,* a v 50 y g

  #     40 -                                                       #

40 4 g 30 - 30 - l 20 - 20 - 10 - 10 - 0 0 0 DE+000 4 OE-008 8.0E-008 1.2E-007 0.0 E + 000 4.0E-008 8OE-008 C-14 on Solid, Liq and Gas C-14 on Solid. Liq. and Gas

                 ,                                                              i 140-      ,                                                      140-     ,

130- ,' I30 -l 120 d 120-l l 110 - i 110 -'i

l: -- c 's ges soco , m o 100-i 4

C .is

                                 -'*  9a N ra'                      9  800 * ,

• i ,

g " ;" a ,e - - - c i . .. e 1 5m 90 - i 5m 90 - i e

                                                                                     ~ C - ' ' C"'

I L 3 80 -i A o 80 - , i i i

 ,     70 -                                                        ,,   70 - ,
  $ 80 -                                                            O 80 -

50 - 50 o o

 > 40 -                                                            > 40 -

30 - 30 - 20 - 20 - 10 - 10 - 0 . 0 0.0E+000 8.0E-009 1.8E-008 0.0 E + 000 4.0E-009 8.0E-009 1.2E-008 C-14 on Solid, Liq. and Gas C-14 on Solid, Liq. and Gas I c content of gas, liquid and solid phases I (NOTts scales cAange) m I b l 3 5

                                                                                                                                  <  e
                                                                                                                                          .A

U O OL p ,

                                         - lMPACT OF " CARBON RELEASE                                                                                    ,

• Cumulative Release at Accessible Environment

-* Lifetime dose to individuals in local population o Near field concentrations in air 4

o. Water from Wells - .

4

• Average dose to individuals'in affected population 7.

I [ I [ i ACNW WCPK78G GROUP MEETNG ON PEBFOPMANCE ASSESSMENT DfVtSION OF HKMIEVEL WASTE MANAGEMENT ' ,

• CECEMBER 16,1992

           ' ACNW ~- 74'.

b

  ,v<+       ,,,         e      .,w-n .-     ~

w. n - +-,,~m ,ee c. e =-t ..-e- -

n-. e ,w-- ~,,-nsnn,- - e.m.er- ,,+

TT2 s NN9 EE 3 n O i t o MM1 SE 61 SG 4A i a eE AN R A E B r E MM a C *c E NTC v SE A AD r M P W t e OL f E e t f EE V P1 m N4 ar OG , GH a MF C p T O E3 E*

     '                ht           V051 f                             P1 o

i w J% J W t G r o n G p i o N K R s t a O n t W a r n W t e N C A d m n i r e a p e x Y s e a R e w A l e r l o O M h e l a M t r t u U o b S f , t r y s o r l e a d p s C

                  '   i n

o n ' m m a f i o l r e f t r o m r e n p e o v o f i i d s i t u e t c t c a ne ai s t s e r i t v e d v a a d t ei t g w a i n i f i i s n m d l l pne i n a p o o r a i c i ms l e f w m se v e s o e ei n e l h r amr d a f c e s o se s l e a e l i C R G G ee t 5 do d 7 R N o o o Mto - W N O

• e C

A

I. TENTATIVE CONCLUSIONS i 4

                . Significant concerns exist about likely size of gaseous "C releases and the uncertainty in 4

. estimates of those releases.

, i

• Even after substantial research and analysis, the remaining uncertainties may make it ,

difficult to evaluate compliance with EPA's "C release limit e

i i i

1 t 1 je i  ! t w ACNW WORRING GROUP ATETF3G ON PERF09MANCE ASSESSMENT DIVtS80N CF H8GH1 Evil Wr_STE 1WLeJ4ACEMENT DECEV8EM iS.1992

                     ~

ACNW . O O O

      -       .         .  .    -   .     -    ?   --           -.-                             -.         .      -         ._       .- _

O O O O 1

Use of Expert Judgment in NRC's Licensing Piecess 1 4 Daniel J. Fehringer, Senior Systems Performance Analyst Repository Performance Assessment Section Hydrology and Systems Performance Branch Division of High-Level Waste Managemert 301/504-1426 F OF D.EL AST MAN T DECi? M A 16.1992 ACNW - 77

INTRODUCTION

• A licensing decision will be based on a combination of fact and opinion.
• NRC's licensing process provides a forum (the hearing) for evaluating facts and opinions.
• This talk discusses:

o NRC's licensing process. o The role of expert judgment in the process. o Compatibility between formal methods for obtaining expert judgment and NRC hearings. ACYN WORKF4G GAOUP MCETF4G ON FENORMANCE ASSES % TENT DIVISION OF HIGH LEVEL WASTE MANAGtVENT CECIMBtR 16,1992 ACNW - 78 O O O Y

8 . O O O PHASES IN LICENSING PROCESS

• Pre-licensing consultation.
• NRC staff review of licensee application.
• Hearing.
• Appeal / Commission review.

l ACNW WOPKWG GROUP MEETNG Oed PEMOTTAANCE ASMSGMENT ON1SION OF H'GH LEVEL WASTE MANAGEA%NT DECEMet R 16.1992 ACNW - 79

ROLES OF EXPERT JUDGMENT

• Pre-licensing consultation.

o Ensure adequacy of data collection.

• NRC staff review of license application.
• Hearing.

o Written evidence followed o' y sworn testimony. o Cross-examination of witnesses.

• Appeal / Commission review.

o Previously established record. ACNW WORKNG GROUP MEETNG ON PERFORMANCE ASSES *mT OfVTS3CN OF HtGH LEVEL WASTE MANAGEMENT DECIVBER 16.1992

   ' ACNW - 80 0                                            0                                                          0

 ~ ~

O O O COMPATIBILITY WITH LICENSING PROCESS

             *    . Two criteria for admitting evidence:

o Establish expertise of expert (s). o :Show relevance of judgment (s).

• Federal Rules of Evidence do not apply.
• Experts need not have first-hand knowledge of facts.
• Excert judgments limited to area of expertise.

o Decision-makers, rather than technical experts, decide admissibility and relevance of evidence and overall acceptability.of a facility.

             *:     Licensing board and Commission not bound by judgments'of technical experts.

ACNW WORfuNG GROUD MEETNG ON PERFORMANCE ASSESSMENT DMSaON OF HGH-tEVEL WASTE flANAGEMENT DECEMBER 16.1992 ACNW - 81 ~

COMPATIBILITY WITH LICENSING PROCESS (continued)

• Combining multiple judgments.

o Judgments of panels are admissible. o Potential problem if weighing method becomes controversial. o Potential problem if members of panel are unavailable to provide sworn testimony.

• Combining multiple models.

o Assigning weights seems OK as expression of confidence. o Weighing and combining projections of models will be more controversial. o Hearing process tries to identify " correct" model for evaluating safety. o if projections are combined, result should be physically meaningful. o Should not discard or down-play relevant information. ACNW WORKNG GROUP MEETNG CN PEPEORMANCE ASSESSVENT OfVtS80N OF HfG+1EVEL WASTE MANAGEMENT DECEMBER 16.1997 ACNW - 82 O O O

O O O BOTTOM LINE Q: How can DOE be assured that its judgments will prevail during licensing? A: There can be no such assurance: o- Judgments must meet a threshold of " goodness." o No protection against new information. o No protection against new interpretations. o Alternative judgements of other parties cannot be excluded from the hearing. ACNW WORKNG GROUP MEEiNG ON PENOfMANCE Asses MENT DNtSION OF HIGH LEVEL WASTE MANAGRNT DECFa8"Ci %.1992 ACNW - 83

l PROGRAM FOR THE IDENTIFICATION OF NEEDS FOR GUIDANCE ON THE USE OF EXPERT JUDGEMENT IN PA Vinklerplora Study Phase 7.5 Elfettation Phase 7FA Effort i

   .                             FretIcut Def inILlen              NorsetIve eapert deveimps                   De seeIfsIs eI l                                 Meetfag I/74                      optfens and recemumendatfees                scenerte plenned

• Mow does espert opfalen for en example elicitation for Fhese 7.5 represent truth 7 factedfe , expert selectten esIng secbentstic
• Mew could dettsfon theory techafgoes, etc. analysts and process be compettble 4 "Impilef t* judgruent with Itcensing process 7 j *Dther issues for report Enamyte Elicitatten 4 Enemine strengths l and Ifattattees m e de Report en DTP as
• W tending of defensible inpet elicitation to licensteg process
• Lessees learned

                                                                      " Issues Geotechatcal                 Strategy for revfew of Deciston-Mating              DOU s ese of expert Stedy                        jedgement 19 FA fa seppert of Ifeensfag Input to FA                          identify needs revfew plan                          for goldence e.g. selectfen                                                           I of experts.

minfefring blas,

                                                                                           .etc.

h Develey GeIdowce ACNW WORKNG GROUP MEETFtG ON FEPWORMANCE ASSES 9MNT DrN OF HO4t(VEL WASTE MANAGEMENT DECEUBER 16.1992 ACNW - 84 O O O ____m

OL 'O O Preliminary Plan for IPA Phase 2.5 Daniel J. Fehringer, Senior Systems Performance Analyst Repository Performance ' Assessment Section Hydrology and Systems Performance Branch Division.of High-Level Waste Management 301/504-1426 Aaron.R. DeWispelare, Prinicpal Engineer Waste. Systems Engineering and integration Element Center for Nuclear Waste Regulatory' Analyses

                                       .210/522-6072 OF (na      asA DECE*ABER16.1992 ACNW -' 85
                                   '                                                          .I

PRELIMINARY PLAN FOR IPA PHASE 2.5

OBJECTIVE AND SCOPE-

• Apply Expert Elicitation Techniques to develop information for climatic scenarios.
• Form expert panel to elicit expert judgment on' characteristics and probabilities for NTS climate' scenario.
• Provide data and documentation for use in IPA Phase 3 analysis.
• Acquire expertise to' aid in reviews of DOE's future use of expert judgment and possibly contribute to the development of guidance a

TECHNICAL TEAM NRC:- Daniel Fehringer CNWRA: - Aaron DeWispelare (principal investigator), Tandy Herren, Robert Baca Consultants: - Robert Winkler (Duke University), Robert Clemen (University of Oregon) Expert Pane /: :To be selected by reputation survey. ACNW WORKFJG GROUP MEETNG ON FERFCfvAANCE ASSESSMENT DIVISION OF HIGH-LEVEL WASTE MNJAGEMENT DECEMBER 10.1992

            'ACNW - 86 ~
                                                        -          ~          .  .                   -             -       -.         . .

O O O PRELIMINARY PLAN FOR lPA' PHASE 2.5 (continued) , SCHEDULE l(approximate) ..F October 1992 Select elicitation approach

. Dec.1992L- Jan.1993 Conduct reputation survey Feb. .May 1993 Conduct elicitation' June 1993

Prepare initial draft report p July 1993 Finalize and submit report to NRC e

I 4 J is T 4 ACNW WORKING GROUP MEETNG ON PERFORMANCE ASSESSMENT 3 DIVtS ON OF HtGH-LEVEL WASTE MANAGEMENT DECEMBER 16,1992

               -ACNW - 8.7J

!~

          ,                   u.                        .           .    .   - -           _    _      . _ . . . ..     ..     .
                                                                                                                                 .=

Considerations for Future NRC Performance Assessment (PA) Activities Margaret V. Federline, Chief Hydrology and Systems Performance Branch Division of High-Level Waste Management 301/504-3396

                                              ~ " "*""""5lII"t%[l1""Oi'JO"l"J m

DECEMSER 16.1992 ACN'N - 88 O O O

  ~

LO .

                                                     -O                                                  LOL LONG-TERM GOALS OF NRC PERFORMANCE ASSESSMENT PROGRAM                                                       '

• STRATEGIC PLAN FOR *A o DEVELOP-AN ' INDEPENDENT UNDERSTANDING OF THE SIGNIFICANCE OFTHE KEY l PROCESSES CONTROLLING REPOSITORY PERFORMANCE - ,

, o DEVELOP AND DOCUMENT'AN INDEPENDENY CAPABILITY FOR REVIEWING DOE'S-TOTAL SYSTEM AND SUBSYSTEM PERFORMANCE' ASSESSMENT FOR THE PROPOSED

            - GEOLOGIC REPOSITORY
               ~
      -o         ASSIST-WITH THE TECHNICAL INTEGRATION OF NRC RESEARCH PROJECTS AND TECHNICAL ASSISTANCE o       CONTRIBUTE TO THE DEVELOPMENT OF COMPLIANCE DETERMINATION METHODS
            -FOR THE LICENSE APPLICATION REVIEW PLAN (LARP) o       REVIEW AND COMMENT ON DOE'S TOTAL SYSTEM PERFORMANCE' ASSESSMENTS'
            -(TSPAs) AND' IPA RESULTS o       PROVIDE GUIDANCE.TO DOE ON THE ADEQUACY OF SITE CHARACTERIZATION DATA                           -

AND REPOSITORY; DESIGN i

                                                             ^'"* *
                                                                       "'""""r#Jfa=0"T=0%**=f" DECEMBER 16.1992
    - ACNW - 89 y    %            .   ,r           . , ~ i.    w
                                                                                                   .. ,- .u.

3 STRATEGIC PLAN FOR PA , LEXAMPLES OF FUNDAMENTAL PA ISSUES WHICH WE EX' ECT TO ADDRESS.IN OUR

PLANNING PROCESS:

o VAllDATION

o. . DETERMINING SCENARIO PROBABILITIES ,..

l 'o SCENARIO ANALYSIS /CCDF CONSTRUCTION o' USE OF' EXPERT JUDGEMENT o FLOW IN PARTIALLY SATURATED FRACTURED MEDIA Fracture / Matrix Interaction

                                   . Steady State vs. Transient
                             - .   . Tw'o-phase Flow

( AC. # WORKING GROUP MCETNG ON PERFORMANCE ASSESSMENT  ! DfVISION OF HIGH LEVEL W E AG ENT

                   'ACNW .- 90                                                                                                     4 e                                             e                                                     e
                                                                                                                                     ~
              -w..

y, -y , b + i,~;. , w ,+

O O O SHORT-TERM GOALS OF PA -- FISCAL YEAR 1993

 -*   COMPLETE PHASE 2 OF IPA

• CONTINUE WITH THE REVIEW OF DOE'S TSPAs (SANDIA NATIONAL LABORATORY AND PACIFIC NORTHWEST LABORATORY)
• CONDUCT PHASE .2.5 OF IPA
• PREPARE STRATEGIC PLAN FOR PA

  =   SUPPORT NRC INTERACTIONS WITH THE NATIONAL ACADEMY OF SCIENCES (NAS)

AND THE ENVIRONMENTAL PROTECTION AGENCY (EPA)

• PLAN FOR PHASE 3 OF IPA
• INPUT TO SYSTEMATIC REGULATORY ANALYSIS

                                                      ^'" " ""'" sis fo7 Ed[E' TEE"sfe'u^uS E,'-   r DECEMBCP 18,1992 ACNW - 91

  - .-.    .~     -           -.  .              .

PL.AN FOR PHASE 3 OF IPA -- CONSIDERATIONS - <- c

- *' 'PRIORITIZE ADDITIONS AND REFINEMENTS OF EXISTING NRC IPA MODELS AND CODES-lDENTIFIED UNDER PHASE 2

• INSIGHTS FROM DOE'S TSPA's; NEED FOR DETAILED REVIEW IN SPECIFIC AREAS USING'NRC MODELS AND CODES AND/OR DOE CODES
• CONSIDER SCOPE-OF NAS STUDY REGARDING EPA' STANDARD' a- . TECHNICAL UNCERTAINTIES IDENTIFIED THROUGH THE SYSTEMATIC REGULATORY ANALYSIS ,

t i  : i

                                                             ^"~~~ 75"o"" UE3;$;3 1 ACNW.- 92 O

1

IO O [OJ CONSIDERATIONS FOR ADDITIONS  : AND ENHANCEMENTS FOR WORK IN PHASE 3 e: , Task 1 .: SYSTEM CODE , e improve'the handling of data in the system code . Task 2 SCENARIO ANALYSIS-

• Develop a better method for estimating the probability of magmatic events. .

Task 3 FLOW- AND TRANSPORT

• Enhance our understanding of transient versus steady state approximation for liquid.  :

flow  :

e. Enhance our understa'nding of two phase. flow at Yucca Mountain L

l ACNW WORKNG GROUP MEETNG ON PERFORMANCE ASSESWENT L . OfVfSION OF HIGH1EVEL WASTE MANAGEMENT DECEMBER 16,1992-ACNW - 93 ~ , b

CONSIDERATIONS FOR ADDITIONS AND ENHANCEMENTS FOR WORK IN PHASE 3 (continued) Task 4 SOURCE TERM , e Refine SOTEC Module , e Develop model for release of colloids from waste e improve modeling of near-field involving the thermal effects for hydrology and geochemistry Task 5 DISRUPTIVE CONSEQUENCE ANALYSIS e Perform simulation.of waste package / rock response to thermal and seismic loads site. e improve consequence modeling for magmatic events , 1 Task 6 SENSITIVITY AND UNCERTAINTY ANALYSIS e Explore alternative methods to multiple linear regression ACNW WORKWG GfOJP MEETNG ON PERFORMANCE ASSESSMcNT OfVISION OF HG& LEVEL WASTE MANAGEMENT DECEMBER 16.1992 ACNW - 94 e e. e

  ~

O O O j CONSIDERATIONS FOR IPA FROM DOE'S TSPA'S t Examples of TSPA issues to be Considered for Analysis in Subsequent IPA:

• Comparison of NRC and DOE Gas Flow Models
• Comparison of NRC and DOE Source Term Model
• CCDF Combination and Construction
• Comparison of NRC and DOE Unsaturated Flow Models
• Comparison of NRC and DOE Scenario Methodologies ACHW WORK?JG GROUP MEET NG ON PERFOWANCE ASSESS */ENT DritS3ON OF HIGH LEVEL WASTE MANAGEMENT DEG MBER 16,1992 ACNW - 95

I 1 T I POTENTIAL IMPACTS OF NAS STUDY ON NRC'S PERFORMANCE ASSESSMENT PROGRAM

• DOSE-BASED STANDARDS o Need for site-specific pathways models o lNeed for pathways an'd nuclide data o Need for. demographic projections o' Need for Land Use projections

        -*      INDIVIDUAL PROTECTION STANDARDS o Need to'know c~oncentrations'of releases o Need to identify appropriate individual (max, average, adult, child, etc.)

• RELIANCE ON' ACTIVE INSTITUTIONAL CONTROLS o Eliminate human intrusion scenarios o Emphasis will.be.on Design Basis release estimates to ensure adecuacy of remediation plans .

o LSignificant differences.in type of. modeling required LSUPPORTABILITY'OF PROBABILITY ESTIMATES o Potential for deterministic standards

                                                                          ^ " * " " " ##5"o7nE E7a 7E//u^$$$$ E DECEMBER 16.1992
        . 'ACNM/' 96 -
  . _    k k                                      .

o o o -Q CONSIDERATIONS FOR IPA RELATED TO DEVELOPMENT-OF LICENSE APPLICATION REVIEW PLAN (LARP) LARP provides_ guidance to NRC staff.on how to review DOE's license application and how to determine if DOE has demonstrated compliance with 10 CFR Part 60. .

                 -e      Performance Assessment is being used to help' focus review plans on the technical areas-that are most important to performance.                                                                                               ,

e ~ 102 Review Plans will be developed which correspond to 10 CFR Part 60 regulatory requirement topics and associated sections of the Format and Content Regulatory Guide

(DG-3003) i

• Initial efforts have been directed toward developing strategies for compliance '

determination for regulatory' requirement topics f i ACNW WORKWG GMOUP MEETNG ON PERFORMANCS ASSESSMENT ' OfvtSION OF filGH4EVEL WASTE MANAGEMEPET . DECEMBER 16.1992

                 - ACNW - 97.
                                   \

9 5- Au=- =- L ._.- my.._wm _.m,.. __Em- _.-m. - - - - - - - - -- 1 _. _- u

                                       -EXAMPLES OF CONSIDERATIONS FOR IPA RESULTING FROM LARP

• Potentially adverse conditions

                     -. ? Ability to predict the likelihood of earthquake recurrence

Adverse effects of future groundwater withdrawals on the groundwater flow system

                *  : Assessment of compliance with the engineered barrier system performance objectives
                     - Prediction of th'ermomechanical effects on waste package and engineered barrier L                        system
                     - Prediction of release rates of radionuclides from partially failed waste packages e   Assessment for Compliance with the individual Protection Requirements
                     - Conceptual model representations of the natural and engineered systems
                     - Validation of mathematical models-ACNW WORKNG GROUP MEETNG ON PERFORMANCE ASSESSMO4T DNISION OF HIG44.EVEL WASTE MANAGEMENT DECEMBER 16.1992
              ' ACNW - 98 0                                                ie.                                                           -.e     g

. I                                          L ,

TOTAL-SYSTEM PERFORMANCE ASSESSMENT EXERCISE (TSPA-1) OVERV!EW l PRESENTED TO THE ACNW WORKING GROUP ON PA PRESENTED BY DR. HOLLY A. DOCKERY SANDIA NATIONAL LABORATORIES DECEMBER 16,1992 - -' BETHESDA, MARYLAND

  ]                                           l l

l

PURPOSES OF THIS TSPA , l

             - To start the series of iterative total-system assessments
             - To review system performance under conditions thought reasonably worthy of modeling before site characterization
             - To help in guiding site characterization
             - To learn by doing the interim steps in the SCP l               process
             - To learn how to reduce detailed models to the simplified models needed for CCDF construction TSPA91HD1P.125 ACNW1121692

PARTICIPANTS Sandia National Laboratories:

                                 - Coordinated inital stages of the TSPA
                                 - Performed calculations using abstract models Pacific Northwest Laboratory:
                                 - Performed calculations
                                 - Performed dose calculations Los Alamos National Laboratory:
                                 - Provided information used in volcanic analysis
                                 - Defined parameter distributions for retardation Lawrence Livermore National Laboratory:
                                 - Specified the source term                                               o TSPA91KDdP.125 ACNW/12-16-92
                 .O                                    O                                            O  . .

                                                                                                                                              ~

O O O SCHEDULE b E W W Q- Z O y)

     >- O                        <C        cn Io                          n          w
 .                                                                                 m 87      88       89       90        9               92          9 Construct.

Data Set P liminary SNL

    \        Begin             resentation                                  Report ~

Calculations YMP - Released PNL Draft to . B in Dose YMP- Policy Report Rev.,iew. Released-alculatioris

                                                             /

l _ i i Y. i l l r; iY l i I I I i 7/91 10/91 1/92 L/92 7/92 12/92 l T5PA91HD5P.125 ACNW/12-16 92 l

EXPANSIONS ON PREVIOUS EXERCISES More phenomena modeled Releases calculated to AE along 2 paths , More sophisticated source term used .

              - Better understanding of water contact modes                                 j
              - Computationally simpler
              - Larger suite of nuclides incorporated Stochastic : simulations; performed Some sensitivity studies performed Results used in dose calculations l

TSPA91HO7P.125 ACNW/12-16-92 . e o e 1

O

                                                                                                                                                     ~

O OL: SCOPE OF THETSPA

                                                                                                                                                           -t

Groundwater Flowc Conditional CCDF ,

i UZ Composite-ModelTransport  ! UZ Weeps-ModelTransport [ t , UZTransport

                                                                                                                                                           .t s

LGas Flow:; __ Conditiont. CCDF +

Surface Release l

                                                                                                                   + Total-S         Ystem
                                                               ~

L CCDF Human Intrusion _ Conditional CCDF- + l .. T Surface Release - l Direct SZ Release"

Basaltic Volcanism? . Conditional CCDF +

l ' Surface Release. i i

           " ~

L. Tectonism- _ Conditional CCDF + l l: -WaterTable Rise . !' Gas Flow in Fractures-TSPA91HD2P.125 ACNW/12-16 92 o .; r e ,, -- s r  %. 1m. -s - , ew

an. w- .

               -   ~-_u. s.. , a    n      -                                     ~sw.   .
                                                                                                                   ,. x u - a n . .a      .~w~~~...-,. s    . . - . ~ ~ . , - + -            u.. - . . - -_..-+.~~..~.~;...~.--...~.~-..-

t g l

                                                                     &_                               ~ %$

l 69 M Q '4;;u" "l:fi um , en

                                                                     ".mgk       G                     45h                                                j;%y
                                                                                                                                                                              ;c[% cfgkf/ppy?                                       __                Q   T
                                                                                                                                                                                                                                                          ~
                                                                    ]u gry j p .{W;@1                     N         !

jEfi.4; q ui c f

                                                                                                                                                          =
                                                                                                                                                               &,~ -' u
                                                                                                                                                                              mm  ; 3ggct g 'u'a                                                     ,_

Q u o I u- : h,, ipr O< W [ US '

                                                                                                                                                                                                 ~
                                                                                                ;     ~

k $$

                                                                                                                               ""**                                                                                                  '            *~

C@L $ $ Ty 5 M((*[i$$l3 & -g-i y m mityst e4%%mj k,3;ygf .. g C O  ; tap -Q- &= QL? b gj g a -- --: 9y_,p/.. lk$pch$1gbljss;g Q)g @{w $ i Q +& 0

                                                                                                        %                         g                                                                                                                       e        l V                          fQll @ghy "

OB >E #y#g s $gg y ,

E um

                                                                                                                                  .                                                                                                                   u.           :

W yR"g _

                                                                                                                                                          ? 'it y ~ '                                                                    .

m  : om , e 70 Cl) ea u . C g < twa h a- 4 Q @n: a_. n

                                                                                      'y u
                                                                                                            'le_.~~.

nh t

                                                                                                                                                                .=
                                                                                                                                                          %d ).{f                        ,_ ,

E

                                                                                                                                                                                                                                         =

Nk,y s . bi kh&$% k

                                                                               &2NhW l',3yg Nsp$td@-

L(L. g Z O As ,_ . 4 ti s4, b;. =4 x

                                                     +            e-               _

O

                                                                                                   +

a4

                                                     ..l<

9  ? .g .

                                                                    *-J'                           Y
                                       .}

. Q V :

                                        # U%

n k

                                                                                                  $i
                                                                                                                                         .x

4,:l? >

                                                                                 ?             y                                    _

u-

                                       ~

g ign m @

p .' LO

$h k ln l, e amene .

                                                                >>,                             y s

q '~! ;"' d .i:* 'e$ J b N' .m O.

                                           $                             ~

2 w L

                                         ,              c 4                                                                               +-s                                                                                                          a 4.

4% N Ih . y} O IN.. NA g ' Q f[,-W. . ~y- '> :q

                                        ]             s e

Ji s. 2

• k E. g i #

                                        .![s.

4li ff. ' y.

                                       ====          C'( W 4

4 '

     ..~.        - ,              ,.              m,

                                                                                          ~ ~

O. . O. O - i MODELED TRANSECT . i a r L TSPA Transect j r H-5 l G-4 l Outline of UE-25 a#1 N Potential i Repository- ] Ghost- . Dance e N i Fault 4

O .5- 1 2

(km) 1 8

T5PA91HD8P.125 ACNW/124692 ' I

a.d . 4 1- m . _a a e h

                                                                                                                    *4

( e -Q %a

                                                                                                                                                               %                         mD c; %
                                                                                                                                                                                                     '              90 9

mm 33 e hh h &- 5 o'%D a% e - 5 a mN *

                                                                                                                                      \                                                 O\                          k
                                                                                                                                          'l             n i   %jkBM%t. . ..

( !s

                                                                . -?                                              m r                                       'c@yg.-nh h N ,'

N

                                                                                                                                        ,                                          j)'g,g' ' T v;:p:g.f :i%
                                                                                                ;;';                         ' ~
                                                                                                                                                                                    ;%.w t ;l:n. .%:l2fC??l k
                                                                                                        >e                                 >

9%.qp?%nu4MgW i, vism n

                                                                            <                                 >           , am                                                    ,+ m q gtBf,%:
                                                                              >>\

• 3.y - .

N 2 hdk & W_ *t& Y. d5 y'

                                                                                           ,,'s
                                                                    , ,                                                                                                                            t          d
                                                                                        ; <: D
                                                                                                            ;j @%:.""+e.         ' <                                            $M @f'
                                                                                                                                                                                              '3

=

k f'. $$ 1.'fh;..} If3[x PSY; fth 3 $FA * '

                                                                                                                            ~&                                                g p%A W g                                             4o
                                                                                                -          e:    ^ w-u99           de np a

4.+'<

                                             ,.,e                                                                                                                             jijg, y                        5..
                                           ,j-l' ^ ' sgg

( 3 -ggy] s gf; }, ggj Q) u og.

                                                                                            % R 3 lNGy3 s                                                                       sg :&@gQ n                                                                              <
                                                                                                                            -,, .                                            ga n.          #ys -
                                           'e w
                                               ,t                                                                         ': p~)           ,                                                                  ce

.} gg<.

                                                                                                                                                                                  , - e Qg)
                                                    ,,.                                              v-m              1         . . .

i M 5... SG /g 6:p % '- y% 9 g < - -y. r , .. .ug

4p ;

s' ' + rm ww-ss p y f'r y wa z~ >>a agg wm y;;;qy ww ym y  %

                                                                  > k                      a         'ASMQs                  .3N                         =
                                                                                                                                                                           .G M*kydes.$W 4

e .m N@ + 7 , 04:4 'g. j .s O [ ' _[^4*' . c; dhh , g.

  <                                                                                                  n")}*hk^                 <4* &d5 >                        ' gq
                      . O.,.,,
                                               '~
                                                                                                    v:cw*                              9                     -         aw %,                   l
                                                                                                                                                         'g'

• jW < g4 I"gg$g; 4 Od Zp j L 3w r T; M

e: e

• Qj Si lg 1 3 <

   ]        gg                                                                                                                                          w- l! b.g
                                             .            1
                                                                                                      =Qt WQ^ j                                                                                                 l
                          ,                                     w                                    eA-                                                  CL i + in kl                                  e:                        : L                                  4%                                             $- i- t~~                                       g i

g .O .C -  ; y,. p g'" o7g E$5E

           *- C
                                            *%=,M 13:: g - gp@.g im E $gjg$$gdM j

h

      .m gO o                           ;                ,                                                 ;                                          mogggggg gp:

o; C eA32 g p-E o .c gpu gm sg. g*@ .c e g g chaw nM: o e3 ]n *-c c' e M ,MprR i%y: _ n, O x' , ^ . . - . g~

. I e _m.- --  ;'

r

                                                                                                                          ^- . '

1 ,%%fX.:%wMMkf?fy? PM,v eyp%,uemg

                                                                                                        *>               s v              >  -

Q.. am. OfM M 6,MM.4

                                                                                                                                                                                  .n.,

s w w g,- kN ! fNhh. h ff5fDhWW)5 .fk

                                                                                                                                               >=*'              6,,    f' m%?rM                 Nv 5 W N #1;X k_iPs

N+ :: #nurenft s 3

                                                                                                                                                                 /^ibs' nib $9shdf0D.y $F +

v m, Nh . kshkEllb.

p. m e w s d W;f:hpdkIbt g.

O O O' TSPA STRATIGRAPHIC CROSS-SECTION USW H-5 -

                   \    Ghost Dance Fault           UE-25 a#1                   -

USW G-4 welded x Potentjal x vitrophyre Repository

    !!$$$ vitric zeolitic partiali                        Water welded                          Table-TSPA9tHOtop 125 ACt.W/12-1692

p. -

     -                                                                                                                   p GEOHYDROLOGIC DATA SET l                                 DEVELOPMENT                                                                           l 4

Derived from site and analog data , L

              - Matrix values from' Peters et al., PACE, and analog sites i

(Apache Leap) 1 L - Fracture properties from Spengler et al., Zimmerman, i and Carsel and Parrish

              - Distributions developed for each parameter 1

Providedia long-needed tool for this and subsequent i analyses o TSPA91HD11P 125 ACNW/12-16 92 O O' O . .i

                                                                           ~ '

O' O .O GEOLOGIC DATA SET APPLICATIONS 1 r Flow and transport calculations: . SNL . - Unsaturated groundwater scenarios

          - Saturated groundwater scenarios t

[ PNL

          - All scenarios                                                        j i

TSPA91HO12P.125 ACNW/121G92 t

TUFF RETARDATION COEFFICIENTS

• E l Geochemical information interpreted by Meijer (LANL)  :

Rocks subdivided into 3 types: vitric, nonvitric, and zeolitic

- Nuclides with. retardation = 0:-Tc, I, C

                 - Nuclides with retardation = 100: Am, Pu, Sn
                 - Nuclides with PDFs: U, Se, Cs, Np Ranges.of retardation values established for the range of.pH values in J-13 water                                              j

- Oxidizing conditions assumed TSPA91HD13P,125 ACNW/1216 92 e e .e .

                                                                                              .l

                                                                                        ~

O O O CARBONATE RETARDATION COEFFICIENTS Carbonate-aquifer retardation values based on data from WIPP (Culebra Dolomite) Matrix values only Water chemistry

                    - Oxidizing conditions assumed
                    - Chlorides assumed to have no effect on Kos Kos for only Pu, Am, U l

TSPA91HD14P 125 ACNW/1216 92 s ..__ - ....

BOUNDARY CONDITIONS Lateral. boundaries - no flow for 2D

                - Run from ininal saturation and flux to steady-state

for specified percolation i

Range for flux at the repository horizon =  : 0.0 - 39.0 mm/yr.

                  - Range of values allows for climate change
                  - Range ensures some calculations exceed threshold for1 fracture-dominated flow                                                  .

Shape of distribution weighted to low flux values y J TSPA91HD15P.125 ACNW/12,1692 -

e e gj

                                                                                      ~

O O O DISTRIBUTION OF PERCOLATION FLUXES AT REPOSITORY l:

           ~

8 8- - e LL I I I I I i I i i I 0.0 1.0 2.0 3.0 4.0 5.0 Percolation Flux (mm/yr) TSPA91 hot 6P 125 ACtav.12 ?6 92

1. _

CAVEATS TSPA-91 reflects our current understanding of the site. It Is expected to ultimately contribute estimates of the ability of a potential repository system at Yucca Mountain to comply with regulations; however, It is not comprehensive in terms of modeled components, The data and models used are not validated,

    -   And the ranges of values very broad Therefore, the " answers" should not be used as                                  ,

baseline values for licensing. They may be useful for guiding near-term site characterization activities TSPA91HD17P 125 ACNW!121692

9. O O .

L O o o ~!

SPECIFIC RESULTS TO BE PRESENTED I  ;

l Comparison of calculations based on: [ - Detailed models l- - Abstracted models i Construction of preliminary CCDFs n Comparison of results for dose based standards l Guidance for next TSPA iteration i 1 4 TSPA9tHD6P 125 ACNW11216 92 f 4 r - s

1 O O O V' i I - I

       . J                                                           L   .

i i

SUMMARY

OF TOTAL-SYSTEM PERFORMANCE ASSESSMENT i PERFORMED BY SNL l i PRESENTED TO THE ACNW WORKING GROUP ON PA i PRESENTED BY DR. FELTON BINGHAM SANDIA NATIONAL LABORATORIES DECEMBER 16,1992 BETHESDA, MARYLAND f I i l ___-__ w

                                                                                              ~

O O O ~l L PURPOSES OF THE 1991 TSPA i t i Development of the abstraction process l - The process in the site-characterization plan (SCP) 1

                  - Implications of alternative conceptual models
                  - NRCinfluence 4

l Assembly of complex combinations of

probabilistic data

- Use of field data, as in the SCP i

l Guidance for site characterization [ Guidance for future total-system analyses !' [ TSPAFBGt 125 ACNWN2-16 92

MODEL HIERARCHY: "THE PYRAMID"

                                                                                                                                          '                 ,/
                                                                                                                                         .l.=*-
                                                                                                                        +                                                                system.

4, 'qgodegng:

                                                                                                                                                              'se          '4P.                      (
                                                                                                                                                                         ',G
                                                                                                                                     -N~ :

hk ) _' + .. phE"h

,e -

                                                                                                                                                                  -rp l

l TSPAFBG2125 ACNWzt21G92 O O O . .

O O O FACTS ABOUT THE SCHEDULE i l August - October 1991: calculations November, December 1991: preliminary writing, , presentations . January - March 1992: writing of draft, . recalculating L ! March-May 1992: revision of draft presentations L TSPAFBG3125 ACt&t/1216 92 ~. -_

ASSESSING COMPLIANCE WITH THE EPA STANDARD

 . Scenario 1:.

f3 9 Complementary f

           ,e              Scenario 4:                     N             cumulative
        '8l                             ,e                                distribution M                                    function:
                                  ,r'        %.

Scenario 2: b%7_> iffg ,'( EPA

                  #                                                                                     Hm s Scenario 5:

c' P

      'uh'                 4#

f J e >

                         =           g L           -

e

                                ,,','                                                        R Scenario 3:-                     b                 :
                  ,s ,                                              R = normalized value of

• radionuclide release f
• over 10,000 years

         '                    Scenario N:                    j      P = probability that release exceeds R TSFAFBG4125 ACNW'121692 O                                           O                                                              O . .

li\i 'l!l\iI ll

 ~
 ~

O N 2 9 6 I , A 1

                                                                                           .'.                                   2 1
                                                                                          .,                                     /

T s y .-- W N C N m -- A 5 2 1 U s _e g 5 G O O i F p n n A P

                              . .p a                                             _a S

T M  :

g.

l c o = h A V C e C - i c t a C t n a l r U . s a i i C Y B -

                                                                                                ~

R ( . L f fn'

                                                                                             .. _ ~ y O                                                     -

y;x 's y,

                                                                                                       .r _-
                                                                                                              +

y(T F , O D E .. L .y , E ng s n D ng i o - O - g-t i n e d o M n i s o u S .vg a C t r n N ,a d i e n O I

                                 ~

br u s a T . t  % m I s h ll u D H i d n N U O '5, C

                                                                                        ,          ,)z u

O. f llll,i 1ilil l lIl ll i

THE SOURCE MATERIAL FOR THIS TALK i TSPA 1991: An Initial Total-System Performance Assessment for Yucca Mountain R. W. Barnard, M. L. Wilson, H. A. Dockery,  ; J. H. Gauthier, P. G. Kaplan, R. R. Eaton, F. W. Bingham, T. H. Robey e t 1 TSPAFBGG 125 ACNW/1216 92 l O O O .

          ~

O O O '! i SCENARIO CATEGORIES  : i Groundwater Flow Conditional CCDF _y l. ' l UZComposite-ModelTransport , l UZ Weeps-ModelTransport - UZTransport I 4 Gaseous Flow _ Conditional CCDF -y Surface Release l Total-SYstem .

• CCDF

Human Intrusion _

Conditional CCDF -> , l Surface Release - s i Direct SZ Release i . Basaltic Volcanism Conditional CCDF -> - l Surface Reiease s. Tectonism -

                                                                          .g Conditional CCDF _y l        .WaterTable Rise               -
                               . Gas Flow in Fractures TSPAFBG7.I25 AC*r#!1216 92 1                                                                                                                                                                                           i
                                                                   . ~   . . _ . .        .     ._     _ _ _ , . _ . . _ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

TREE FOR VOLCANISM Basaltic Volcanism i I I I intrusion Acts Directly intrusion ActsIndirectly on Repository on Repository __I I l , , Dike Forms Surface Drainage Subsurta m A!!ered Drainage Aftered

                                 'I        I'               i                        ;                              .

I i I I; Transport No Waste Magmatic Basaltic Cone of Magma Alteration of po,ms Siti Forms Waste intact Contact Waste i i i i i i i l I Waste Fragmented and Entrained Waste Entraned in Ash Ptume f I I I Waste is Expelled with Cinders and Flow Direct Exposure Surface Weathenng i I l  ! I I l Direct Exposure l l SurtaceWeathenng l Waste is Altered Chernicany/ Thermany and Entraned I Waste is Expened in Cinde< Cs..e l 1 I I I I Dkect Expmure l l SurfaceWeathenng l i TSPAFBGB 125 ACtW1216 92 O O O .

l; ! z\ \ \ \ \ s \\ \ \\ x z ' t  ! t

 ~

O 2 9 s i 2 e _ u _ c E A s e - T t m e S d r a A s _ T _ W  % p \g G . N I N I . w '

                                                                >,r-
                                                                 ~
                                                                  ~
                                                                     ~       7 A

R a\ .~t8 T . O N

                                      \

kh E E K g.-

• u l

. D - CI

                                                       / -

T L ( A S A B O

                      ,-                                               \   '           'iii

i : t! t .!  !!ll:;!  ;; !t'!! t!!ir!Li{ l 2 9 G 1 e 2 e v O A 5 2 1 c t G B F A P S T 0 4 2-u 1 4 . P 2-

                                     /                               m r                                                A e

h t O e

                       /[

7 8 3 3 1 2-s u C P > 9 3 2-u P

                                                                            . e.

4 r -  ! }! ; . ,ltal!

I

                                               ; - :>';'.,                                                    i ,'             f        ;:i l   .     ,

e v T - O 2 2 9 6 1

                                      , ,                                                                                             0           2
                   //                        '/- /                                                                                    1 1
                                                                                                                                                  /

W M /# .

                                                                                    -                                                1 N

C A 5 w S -f. '- /

                                                                                                                    -                             2
                                             /             .
                                                               ...           .         :.       ^

0 1 1

                  /                 .

1 I . - 1 G - O N

                                    .                      .                                                                                      F A
                                    .                                                                                                             P A
                                    .                      .                                                                         0            S T
                               .           . : ;.                    .             :          .~           -         -       -

0 C w 1 L - . - - O

                               .f
                                                                     . . . .              . .~                                         0 1

V - 2 R 0 m

.;.:.;.:N.i.

.. i  ;-

u O - 1 s F 3

                                                                                                                                       -      A P
                                                                                     - .-                                              0 F      ::.              . . ..       . : . :..                   : .. l                           i I-            . :

E O

                         -          .                      .                                                                           1 D                  -

4-C

!: . . : .  :  : ..- l .  : - -

0 C 1 s-L - A 0

.j:. . : ;.- . . l .  ::  :- -

1 N 4 O

                    .i -:. . .:.:..:.                      .           :..:5              .- ;-

0 - I - . - 1 T I -

                                    .                                                                                                7-D 0

_ N 0 1 3- 4_ 5

                                                                                  -             4            7-                 8 1

O 0 1 0 9 0 0 0 0 0 C t 1 3 1 1 1 1 1

          > =**O.~

j Oud q)>_ sa3E3obg

                                                                   -                      o   5E9.oEO O                                                                                                                                                      w
    +                                        i

SENSITIVITY STUDY: VOLCANISM

   >        0                                                               .. -- . --

2 10 .

                                                          -       -=

10-1

                                                                                         //

e / Q- 10-2 .

                                                    ~

O 5 10-3 i - {-

                                                    ~
   ~5 E    10-4    .         -.         .                  ' % . . . ..                    -

3

    >   10-5
                                     ~
                                                  .;.     :         N..                               ~

E - c -6 b' - U CZ3 EPA limit Base Case -k

                                                                             -j ~
   -[   10-7            Larger waii-nock Fraction
                  ---- Be*3*I E[ypj!on,ypjge,,
                                                    -                          j.

E :I o 10-8 O 0 10 1 10 2 10-7 10-6 10-5 10-4 10-3 10-2 10-1 10 EPA sum l TSPAFBGt2125 ANw'1216 92 O O O .

O O O l GROUNDWATER INTERACTIONS WITH CONTAINER Backfilled Drift C ap Ground

                                          /                 /          Water Air Gap
      /                                              C
                                                \ ontainer f         Rubble-Filled                          Failure Gap

% Waste i

! Ground Container Water Fuel ' Rods y Temporary Dry-Out t

                                         \                         Zone TSPAFBG13125 ACNW:12 ?6 92
,                                                                                                             i

m u d .c t r s e e en i i t v er d r i r l u pe py l ey e l o eye l bm 2 9 l A Ul a Wla Zl a i sn sor 6 1 2

                                                                                               ~                                                              ei     1 cv     W' N
                                                                                           '                                                                  c n    C
                                                                                                                                              -             Ae       A 5

2 1 t - -

                                                                    ~~. N~~.

l - 4 u ~- 1 L a -

                                                                                   ~
                                                                                                                                           -                  =      G 8     .

F ~ F A . E j ~

                                                                                                                                         ~
                                                                                                                                         ~
                                                                                                                                          ~                          p S

T D ~

                                                   ~
                                                              ~
                                                                ~           .~                                                          ~
                                                                                                                                        ~
                                                                                                                                        ~

O

                                                  ~
                                                                         -                                                             ~
                                                ~
                                                ~          -                                                                           ~                -!
                                                                                                                                       ~

M

                                                            ~
                                                          / .
                                                            -             .                                                            ~
                                                                          .,/                                                                         '

x.f

                                                           ~
                                                                                                   /                                  ~

Y

                                                          ~               .
                                                          ~               .
                                                          ~
                                                          ~

f f ~ T

                                                          ~~
                                                                                                                                /T' /

I ,~,~~~ _ _ / S /

                                                                    ~
                                                                  ~

y #,

                                                   ~                                           R O    n          i}
                                                  ~

T O R e~~re

                                                         ~

i o - r ,~

                                                        ,                                   s t

r t ~ O a P e O

                           ~        t                                                                                                                    I r

l i t

                         ~
                          ~

a _ t E R t a P- e f n ~ W a  : n i c ,g Wg '

                                                                                                                              ~

W o z E t i d ~ o Iy -

                       ~
                                 ,n-
                                                    ~

d E ~

                                                                                                                            ~              s                     d e

T s IY T ~ e _u'

                   ~            _                                 l                                                                                               t i

Op O " a I p e ~ o " d r u S

               -                                                                                                            "            l
                                                                                                                             '-                                   t m
             -                     Y.                                                                          -
                                                                                                                               -         O                       S a

O r - - f o - P i n u T -

                                                                                                                          '                             .I n

M T - o -

                                                                                                                       ~

N ~ O ~

                                      ~
                                       ~
                                                                                                                    ~

C i} ~

                                                                                                                                                      ?a
                                                                                                                 ~
                                                                                          .-,-                   ~
                                                                                                                 ~                                    t r

t e d c a e cr i l r W d r i l ey e i t r e y l oeye Wla Zl a i Vla

O

                                                                                                                                                                                 ~ ~

O O WEEPS MODEL Nonuniform, episodic infiltration I 11 t v rt i VY

                                                        . ~....
                                                                 ~...                 y      1 y
                                                                         ~.... ~. ...

t . Fault l

                                        ......... ....-- .~ . _ .......                 .. ~
                           .,Iyl .. , ~ ~ ... ........    .
                                                                              ~.. .. ,

s . . . ~y y l . j

                                                                                        -S                           -

1 WeIded .

                                                                                               ~~..
                                                                                                            . . . ~ ~ ~ . ~~~~.s..-
                                                                                                                                 /
                                                                         ~ . . . . . . . . . . . . ~........'~---~/~~~~~~~.

lafer . .

                                                                                                       ~.............. ...._...... .~~. .... ,.~..... Alluvium f    ~....,~ ~        s...'- Upper i
                          ' '                            POTENTM                                                                               ..

Vitric - REPFTTORy ~~~ Iayers layer s.'. ..~. I.

                    ~ . . - . . . . . . . . .~.....       _ , ,_~' ~ ...'.

f f Welded

                                                  ~'-~~~~ ~ ~ ~'

Zeolitic ~~~I!'~ s . _ _ i / law layer /~""'~~~~~~......-.........

                                                                                                              /

Zeolitic f layer Water table 1

                                                                                                                                             = Accessible Saturated zone                                 f                                              environment t

TSPAFBG15125 ACNWit2-16 32

1 lI!li, ll!! 9% 9 9 O_ _ l e 4^ t _ c8 d 2 T T7 o W N O A _ 5 2 M 1 S T 9% G . B 79 e - w o F A P S T - S1 l

               /                   F             _

_ S s E p S e A 9% e . E 22 . L 1 l

               - 9                  W           .

E 1 R O S U l e O 9% 9 3 d _ E c0 o _ U T 8 M Q w A o l F _ t e i s o p 9% 27 m 1

               - 9                     o    O l

1 C

                   ~

                                                                                                                                                                                 ~

O O O

AQUEOUS RELEASES Composite-porosity model & weeps model

! N 10 U 4

        .O

! cJ ! .O L. a O

        .2                 10-'

l 15 ! ~5

E i n

m 10-2 l E O

!' E I O ser EPA limit k Q Composite-porosity model E ._ weeps modei ' O 10-3 2 10 # 10-6 10-5 10 4 10 4 10-2 10-' 10 10 1 10 EPA sum TSPAFBG17125 ACfWd 1216 92

                                                                             \

0

0 5

i 4 K e i 6 0 i m 3 T e 0 e, 3 - ;0 z, _ 6 0 4 . c. e r i . m yu n i a 0 r t 0 _ t i oa r i i  ; 4 5 3 _ s e op i S pm .. x i0 E e e 0 0 . RT N i3 gp.. I - L - w i0 H ;0 )s

                                                                ,5 r T                                                                       2 et e

A i m P i0(

                                                                ;0                   e 0X W                                               ~                ,
                                                                 ,2 O

L F-g5 i i 0 0 5 1 S , A G xm. s y

                                                                  ,0 i

0 0 1 7 i i

                         <e               ;

i 0 e j0 i5

                     .g_

i i i 0 0 0 6 0 0 4 g g 0 0 6 e 1 ' ,

                ?;-                                  m f     ll              jl                            l!ll                    1[    l<

                                                                                                                                                 ~   ~

O O O C-14 TRAVEL-TIME DISTRIBUTIONS 1 10 . E 0 ( 10 - -

         *5                                i                                  --             -
                                                                                                            /

rj m . l  !: \ D i I 2 ' l l [

l

o. .~ .-. /

o 10_3 .

                                           .                            .. c-         ,v- - -                   :-
                                                                 ;-            :. t                                                                    ,
         .9 3                               .
                                                 '         .i
                                                           -                 -n.

p .

                                                                         )f :~
                                           ~

5 10~ - ------ - l-o u i f. - _ _2 ., p-360K

                                           ~

3330K 315K 300K

                 -3 4                     5 2                                  3                       10                          10

! 10' 10 10 Time (yr) TSFAFOGt9125 ACNw 1216 92 L .- _

CONDITIONAL CCDF FOR GASEOUS RELEASES I 10 0 -__________

                                      .O
                                                                ^

j ' __ ,,,  : - es

                                                                                                       .,.}
                                      .O                        :          :                             :         .              :                      ,              .             .

O  :  : -  : . Q.  !  ! i ',

                                                                                                                                   -                   I
                                                                                                                                                                    / '/ -

g  : -

                                      .?

10-1 -

                                                            ----i---- -- -i - ----
                                                                                                                                               -- d    .                   -

3  : i i . 3 . l'.

                                                                                                         ~

E i i .

                                                                                                                                                                \

s  :  : O [ f L 10-p m  :  : C  :  : O - g E-  : (D EZ) EPA limit I, Q . Composite-porosity model  :* , E _ _ _ _ Heem mod _e_i_ _ _ _ _ _ _ _ _ _ _ _.

                                                                          ~
                                                                                                                                                      .i                 :

0 10-3 ' ' 0 2 10-7 10-0 10-5 10-4 10-3 10-2 10-1 10 10 1 10 i EPA sum TSPAFBG20125 ACNW/1216 92 O O O . .

                                                                                                                                   ~ ~

O O O TREE FOR HUMAN INTRUSION Human Intrusion a n I l Mining Drilling

                              ;                                                  I

I 1

                                                          % I rato g                               Production Drilling i                                   i l

Hydrocarbon Scientific and Mineral Exploration h Exp1 oration

                                                              ,                              l I

D.illhole intercepts Drillhole does not Intercept Waste

                         '         1                                                      Waste l

I I Direct Removal OI Direct I Contaminants Drilling Fluids Removal of Orifhng Fluid Enhance Contaminants Forms locally l I 2 I I Unsaturated Saturated Flow Flow Field p; ego Surface Deposition Surface Deposition in Saturated -

                                                                           -           Deposition                   i I       i                      Zone h

Formadon of Surface

                                         '      n Tu!f i

Carbonate

                                                                                     .h            h Pan:cu!ates        Weathenng              Aquifer      Aquifer                         Contact Exposure Contact Exposure TseAreazi.i25 acswn2 is92

i SURFACE-RELEASE DRILLING SCENARIO Input Drilling Fluid Drill Hole Drill Stem

                           ,Aeturn Circulation Waste-Package       wi:h Radionuclide Fragments Emplacement Hole Spent-Fuel Rods in Damaged Wast l

Package Return Circulation of Drilling Fluid TSPAFBG22125 ACP(W!121692 O O O . .

l lll!l !llll l 1l{lilIf! illlI! ~ ~ l-1 O ) 2 9 5 m 2 n u - 0 wu G i m n" b c a x N I a m i I R 1 m a z a e

                                                                                                 ~fl L                      (

s f

                                                                                                                                 -         r a

p L 2 m i t s r I H R k

                          /

s t c\

                                                                                                               ^

2-D e e l r - o i D M h e r 3-O b o m R 3 4 S u F ' P A S E f o O SE - iI ' I 5- g o L A s 6 E e s

                                                                                                   }

L s 4 E i M s e

                                                                                                                      , 7-R                                   r a

s a e e E N l e :8 C R - o A F r e Z  : 9-R U S0 0 0 0

                                            -         ~

y c 0 0 0-1 0 4 0 n 4 1 1 e u q O e r F lfllfllllllt !lI l l!l , lil l lll lL

CONDITIONAL CCDF FOR SURFACE RELEASES FROM DRILLING

       >      0 10

.o

m  :

n .o  : c) . . . . u . . j o- 10 -' - - - - -i- - - - - ! - - - - - i- - - - - - -:- - - - - - - - - - - -:- --i --- - - , e  :  :  : , e  : 3 4 . . E 10-2 ....i.......;....  :. . ...f........i. .. . .i. ... ..i.. .. l... ... . , a . . o  :  ;  ; - L,.

                                                                 ~
                                                                                                           ~

{ g . . i ** . c - 10-3 e ----- - --- - . g- . c) - , a_  :  : E  : i .  : .  : . 2 4 2 l 8 10 i- 10-8 10-7 10 ~6 10-s10 4 10-3 10-2 10-1 10 0 10 1 10 2 EPA sum s TSPAFE 124125 ACNWl12-16 92 1 O O O .

                                                                                                                                                                                                         ~    ~

O O O SENSITIVITY STUDY: DRILLING i N d

   .~

10 0 y m ~

                                                                                                                                                                                    .,,/

7 h'. O i  :

                                                                                                                                            \ 
   .D                              -

O -

                                                                                                                                               \'   .

L 5  :( .

                                                                        ~

10 ~l  :--

                                                     -? -

o  : .

                                                                                                                                                    \. \.:
                                                                                                                                                          \:,
 . g                                                :                                                                                  .
                                                                                                                                                            }i,,i g                                                                                                                                     :

s  !  ! 17 holes in 10,000 years h 10 ~"  : ..\\ 'i.- s o I.' , 170 holes in 10,000 years b e 1: 1 7 a C 10 - -- 340 holes in 10,000 years. -

                                                                                                                                                                   - +i   -

D ^ g  : .

   ~                                :                 .                 .
                                     .                                                                                                   ~
                                                                                                         , , , , , , , ,i , , , , , , , ,i     ,,,,,,,a              . { !. !.. ..a h
                     , , , , , , , ,i     , , , , , , , ,i , , , , , , ,,i , , , , , , , J    ,,,,,,,a y 9_4 10 -4       10 ~3           10 -2              10 -              10                   10'                10*

O 10 ~8 10 ~7 10 -U 10 ~5 EPA sum TSPAFBG25125 ACNW/12-1692

                                                                           "                                                                                                                             ,, l ,8

l OVERALL CCDF FOR RELEASES (WEEPS MODEL}

            >       0
           -    10                       --

bm  ! [*-hw.glgg- . .

           .o                          :                          .

N: 0 10-1 -- - i- -----i------ - i - - ;:-Ng. - - - :. O-  :  :  :  :  :  : . O (Z3 EI'Alimit  !  ! i Total

                                                                                                                         ?. .h, i

[ y Gaseo gu .. . . .. . :. . . ..

               } Q -2 9                .....A9".'".8............                           :

j.I / h Drliting ,, 3 g

                             . .. ... Volcanism..(method 1),
                                                                                 ~

Q'\ ']. 4 -1 -- 10-3 I b -

: A  :

           .m                                        .
                                                                                 .                       .                                  ;\'
                                                                                                                 , , , . . .*                       i       .

g g

                                                                                                                                          +-
                                       ~
                                                                --i                                                           k-10-4 E
           .e

a.  : .

1: . E

NI 0 10-5 1 2 10 10 10 8 10-7 10-6 10-5 10-4 10-3 10 2 10-1 10 EPA sum TSPAFBG26.125 ACN#'t?.1692 1

h . . s - . _ _ _

o o 9 OVERALL CCDF FOR RELEASES (' COMPOSITE-POROSITY MODEL} e

               >            ^0
                         'I U       ------........                                           .                      .                               .
                                                                                             .e.....             .
             .O

s. .. . .

(t3 . . .

             .o                                    :                   :                     .

O - - -

                                                                                                                                                                                                               .s
                        ]Q~1        ........... .........                             .................

:  :  :  ; :s  : -

G

               ~,

EPA limit - e dIU^ -2 Total . . (t.1

                                    ..          Gaseous                  .                        .                                                 .           .     .                 .

3  :

                                         .....A.9"..*.*.'.s.............                                            .                  .            .            .

E <*ii' ins

                                                                                                                                                                    ~~~ .

3 Volcanism (method 1)-  :  : 0

                                                                             -^

y/' C 10-3 ------?----i .

                                                                                                                                                    ?
                                                                                                                                                             - - - -..i. - -                     --. .

i _as .. c .

                                                                                                                                                                       .. ..u.
                                                                                                                                                                                         . 1       .

g . . E 10-4 -:- -- - - - -- -- --- ----

                                                                                                                                                                      - &*         c     '-

on . . .

.  : : 1 .

O-  : . . :e i. .. 8

                                                                                                                                                                                                     .-                         i E

O 10-5 0 1 2 10 4 10-7 10-8 10-s 10-4 10-3 10-2 10-1 10 10 10 EPA sum TSPAFBG27125 ACNWr1216 92 l

TECHNICAL CONCLUSIONS In this study, gaseous releases are the most important

        - Aqueous releases are the next most important The composite-porosity model predicts higher releases than the weeps model
      - Human intrusion is not ignorable but is less important than " nominal conditions" Direct releases due to volcanism are inconsequential TSPAFBG28125 ACNW/12-16 92 O                         .O.                                         O    .       .

                                                                                             ~             ~

O. O O TECHNICAL CONCLUSIONS (Continued) k i Our results are most sensitive to the choice of ' conceptual model for flow - We.need to reduce uncertainty and conservatism in assumptions to achieve greater accuracy We need to model more sequences to increase confidence. l l We need to include thermal effects in calculations l .of aqueous releases i TSPAFBG29125 ACNW/12-16 92

MEETING OUR PURPOSES

   - Development of the abstraction process
     - We need more detailed modeling
   - Assembly of data combinations
     - It is possible and perhaps not a major problem
   - Guidance for site characterization We.made specific recommendations Guidance for future analyses i
     - We have planned .t he next cycle l.

TSPAFBG30.125 ACPfW/12-16 92 l

                               .   .................__...,..____.m__.._

O'

                                                                                   ~

O O l PROPOSED FEATURES OF TSPA-93 l Gas flow and transport: new C-14 travel-time a distributions

      - Transient calculations coupled with thermal pulse

. - More-realistic permeabilities and stratigraphy 4 1 . Saturated-zone flow and transport: more-elaborate j modeling 1 Detailed permeabilities from new data base

      - Improved treatment of matrix-fracture coupling                               ]
                                                                                     .i
      . Source term: more: detail, less ' conservatism
      - Hot repository with drift emplacement
      - Medium-hot repository.with borehole' emplacement
      - Mechanistic model for container failure
    -  Sensitivity studies Formal identification of important parameters l
 ~                -                    -      --        . . _ _ _ . - -   _ --

PROPOSED FEATURES OF TSPA-93 (Continued) I Percolation distribution: more elaborate, ' less conservative

        - Use of additional data to derive distribution Validity of 1-D modeling
        - Variation of fluxes to determine range of validity Thermal effects
        - Use of insights from lab studies
        - Multiphase, nonisothermal. detailed modeling TSPAFBG32.125 ACNWit? 16 92 8                             9                                            9  .

                                                                               ~ ~

O O O l ( i PROPOSED FEATURES OF TSPA-93 (Continued) Sensitivity studies of colloid phenomena l l Weeps model: improvements

            - Variation of weep sizes and temporal patterns
            - Incorporation of thermal effects l

Composite-porosity model: improvements

            - Properties of columns
            - Incorporation of thermal effects
            - Effects of water-table changes TSPAFBG33125 ACNW/12-1692

                                                                                                                \

PROPOSED FEATURES OF TSPA-93 (Continued)

                      -        Parameter correlations: cross-correlations, spatial
                              . correlations

• Tectonics

                               - Effects on containers, source term
                               - Water-table excursions 1

Volcanism: use of latest data and insights, if any-Dose module: addition to modeling tools l

                                                                                                              .[

t TSPAFOG34.125 ACNW/12-1692 O O O . _ i

O O OL I I 1

       . I                                      L  .

9

SUMMARY

OF TOTAL-SYSTEM PERFORMANCE ASSESSMENT PERFORMED BY SNL  : I i I PRESENTED TO THE !- ACNW WORKING GROUP ON PA .. l k PRESENTED BY

DR. FELTON BINGHAM SANDIA NATIONAL LABORATORIES DECEMBER 16,1992 "

3 BETHESDA, MARYLAND f

i i

O 6 05 ~ . I I L

                        >   I                                                                                               L    .

SUMMARY

OF TOTAL-SYSTEM PERFORMANCE ASSESSMENT ~ PERFORMED BY PNL PRESENTED TO THE ACNW WORKING GROUP ON PA l PRESENTED BY PAUL EISLINGER l PACIFIC NATIONAL LABORATORIES DECEMBER 16, 1992-l

                          ]                                                                   BETHESDA, MARYLAND f

l I I

                                                   ~ '

O O O I I l l

 . J                                          L  .

l

SUMMARY

OF TOTAL-SYSTEM PERFORMANCE ASSESSMENT PERFORMED BY PNL 1 PRESENTED TO THE . ACNW WORKING GROUP ON PA PRESENTED BY PAUL EISLINGER PACIFIC NATIONAL LABORATORIES DECEMBER 16,1992 BETHESDA, MARYLAND f I I

                                                                             ~

O

                                                                         ~

O. O i STEPS IN PERFORMANCE ASSESSMENT Choose a. regulatory performance measure h Develop mathematical models for performance predictions l Collect site-specific data Validate models L Make predictions of performance l TSPNLPE1P 125 ACNW/12-1&92 i.

                                                                               "i

    !      , llj                 !   !                      !

j! il{!! ,!: 9 G N I e s N a e n RT l e e o EN r e s d VE v ) a e e s OM i t 2 l e a a 9 r b GS S l u m 9 1 e v b e SE u

                                          ,     i t         e         go c           3           aso NS                   s           b
                                          . l u                nt y

OS i d a e m n sl ne F TA I w e , uo ci eik AE ) r u 4 ea t cl i t u 9 LC l s s t rl n f a aup rb o o eio a n UN i se r dee s o f n G dw E A rr mta r rp eu l guu eo R M vtacc e.u nt sd iaa n sn uk R 5 n n a l. kne a n oinmes b ues e r FO OF 8n a c 9im wc eo 1 cr e (i t o s. (i cd o mo rd YR 1 sfrdo t 1 s nl 9l aa t nf a ol RE i i 9 lie ocu OP 1 b pd 1bmu i rd i i t fd T Rareboi t Rabfoiv a ei cv i i Fojma Fori l S CrPMi L Cr ed u g pd _ I H 0 0 PPin eSin 4 - - - 4 - - R-9_

  .      ;           :       ;                                :l1               l L l;

                                                                         ~ ~

O PROCESSES MODELED IN THE 1991 TSPA Undisturbed repository performance Radionuclide source term Gas-phase transport Liquid-phase transport Human intrusion from exploratory drilling Disruptions from basaltic dike intrusion 1 Disruptions from tectonic activity Dose to an individual I TSPNLPE3P.125 ACPf#t12-16 92

a CCDF FOR TOTAL SYSTEM PERFORMANCE 10 100 10" 10-'

    ,@  10-2                                                         10-2
    =

10-3 10-3 e-

    $   104                                                          10 4
    ~E
    @   10                                        (               10-5 W  10-6
                                                         )           10 4 10-7                                                        10-7 l

10 4-- 10 4 (- i 10

• 10-7 10-8 1 0 -5 1 0 -4 10-3 10-2 10- 10 10' 102 Ratio with the EPA Limit.

O O O WHAT DID WE REALLY DO? t Modeled a single scenario from several scenario j [ classes I Developed mathematic models for the scenarios- i Limited the analysis to only 10 nuclides 1< Exercised the models (computer codes) using preliminary data Demonstrated that the computer codes would run a TSPNLPESP,125 ACNW/12-1692

DIFFERENCES BETWEEN THE PNL AND . SNL APPROACH TO TSPA

      .PNL used much more detailed process models 2-D transient unsaturated zone hydrology and mass transport Fully transient thermally driven gas-phase movement of'4C PNL ran many less. cases than SNL to get the CCDF of performance
      .PNL performed individual dose calculations for several release scenarios l

rsnar usacuana

  --.L. u .ia t f .h.3, _.w A p. y-        a p- e u w. e m. - .a      w     e - -v-             u O                                                  O.                                       O

SHORTCOMINGS OF THE 1991 TSPA Only a few scenarios were modeled Only a few nuclides were examined LMany site-specific data were not available Computer codes were not QA qualified The modeling approach and data did not

                                     . receive early OCRWM-wide review

( . l l

                                                                                  . TSPNLPE7P.125 ACNW/12-16 92

l: l: LESSONS LEARNED FROM THE 1991 TSPA ANALYSES o Several new models (and codes) are required to l , address all scenarios L Bounding analyses may not be appropriate for modeling processes Models efficient enough for CCDF's may not address processes lof interest

                . Site-specific data' will modify the preliminary. findings The cost of doing " realistic" PA work is high TSPNLPE8P.125 ACNW/12-16 92 9                                        O                                  O -

O^

                                                                       ~

o O WHAT CAN WE DO BETTER? Everything - in both scope and efficiency , Choice of scenarios l EBS models

     . Gas-phase transport Unsaturated zone hydrology and transport Saturated zone modeling Dose.modeling i

TSPNLPE9P,125 ACt#//12-1692 i

                                                                   ~

y

HOW WILL WE GET THE REAL ANSWER? Achieve regulatory convergence Obtain scientific consensus on models and data Maintain a proper philosophic position We are not predicting the " future" precisely

   - We will present a plausible range of performance within which we think the performance of a real repository will be contained i

i ISPNLPEtoP 125 ACf."W/1216

O O- 05 -

                                                                                 ~:

i I

     . I                                                   L  ,

i i

FUTURE DIRECTIONS FOR i

DOE TSPA , = i i PRESENTED TO THE ACNW WORKING GROUP ON PA 1

PRESENTED BY JERRY BOAK YUCCA MOUNTAIN SITE CHARACTERIZATION PROJECT DECEMBER 16,1992 -

       ]                 . BETHESDA,' MARYLAND i                                                  i

O

                                                                                                      ~

L O O l l 1 i

  . I                                                                                      L      .

FUTURE DIRECTIONS FOR l L DOE TSPA PRESENTED TO THE ACNW WORKING GROUP ON PA , PRESENTED BY JERRY 'BOAK ,

      -YUCCA MOUNTAIN SITE' CHARACTERIZATION PROJECT                                                     I DECEMBER 16,1992                                                             '

BETHESDA, MARYLAND f ' I i

W %

                                                                                                                                      ~

O

                                                                                                                                    ~

O- O FUTURE DIRECTIONS FOR DOE TSPA Fiscal Year 1993 PA Priorities Scope of TSPA 2 Participants Methods

                        -                   Schedule                                                                                    l l

TSPAFDJB1P/12-16 92

1

                                                                   ~           ~

O O O FISCAL YEAR 1993 PA PRIORITIES

     -     Support to ESF design
     -     Support to Surface-Based Testing
     -     Regulatory evaluations
     -     System studies (thermalloads)
     -    . Definition of TSPA 2 I

TSPAFDJ82P/12-1692 I

O

                                                                   ~ '

O 'O POTENTIAL SCOPE OF TSPA 2 1 . i Thermal effects on fluid flow & transport Incorporation of new site data . Scenario refinement and augmentation Sensitivity studies to evaluate results Evaluation of. sensitivity to correlation Enhanced source term , Improved modeling of saturated zone Modeling of colloidal transport Effects of water table changes TSPAFDJ83P/12-1&92

m ., L O O O H L PARTICIPANTS IN TSPA 2 i i l i

                       . YMPO/ TAB

! - Direction 1 l CRWMS M&O , l - Coordination, problem definition, total system l evaluation, dose modeling SNL L

                        - Problem definition, scenario development, total system                                    .

r evaluation, natural system evaluation i - LLNL

                        - Source term, EBS evaluation, near-field evaluation

PNL i i

- Dose modeling, EBS support LBL,.USGS LANL

~

                        - Support
                                                                                    -_ n -
                                                                           .._--_ _    ._._ .__ ___.___   . __ _ _t

METHODS FOR TSPA 2 i o ) i Comprehensive scope, stochastic input, j most abstracted process simulation,  ! l evaluate system sensitivities i 4 System l Models Broader scope, partial abstraction, and some coupling of processes, Subsystem evaluate crit: cal processes ! Models 3- t Limited scope, comprehensive , Process Models treatment of modeled process,  ! g for understanding phenomena ii k I i 1 - ,,2., - i i

O O O METHODS FOR TSPA 2 i Four Preliminary CCDFs for Yucca Mountain 1

       .6                                                                                                               i E                                                                                                                !
       .O       1                                                                                                       i 10      -
       $u N                             ;

O- NRC l

        =                                                                                                               i
       $.      -2               SNL                                                       .

m 1O - 1 3 -PNL h EPRI h o ~ D -a Y

       .!g 1 O                                                                                                          i C

G l l E y o. s 1o O

               -4
                     -u
                                   -b
                                            -4
                                                                  -z
                                                                        .          i h,

o 10 10 10 10 1 10 " l Normalized Release to the Accessible Environment i TSFAFDJB6 Pit 2-1692 , st- w a , _m - _ - _ _ ___ ___m______._ .- _ . _ _ - -

[ O O O~ ! CONCEPTUAL SCHEDULE < l l-

f. 1993 1994 rob em Definition Process Model Ca.iculations un'u'$u'uuum

. Scenario Definition num'$ummunu

         ..Sensitivit.ies for T_ SPA 1991- umuu9 mum Source Term Definition au Source Term Calculations                         l mau'uuum'
         . ysten Model Calculations Ss                                               l
         . Draft Do_cument Preparation                    l Internal Review                                  '                                             summum External Review                                 l!                                             I       muuuuuuuuuu'r in addition to routine channels of review, DOE /YMPO is considering requesting review by the OECD/NEA TSPAFDJ87P/121492 L '. _.     ,.          __             - , . . . ,-
                                                              -      -.-...-__.~_ _   . ,_.._. -....___     _ .                 _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _}}