ML20198D786
| ML20198D786 | |
| Person / Time | |
|---|---|
| Issue date: | 12/16/1998 |
| From: | NRC ADVISORY COMMITTEE ON NUCLEAR WASTE (ACNW) |
| To: | |
| References | |
| NACNUCLE-T-0127, NACNUCLE-T-127, NUDOCS 9812230208 | |
| Download: ML20198D786 (318) | |
Text
{{#Wiki_filter:MMMMMMMWW ' b jy d fW M P H$g$ f,g $ $g_g!BNy _ g g ,w g n[y g a s. c hm'k p a-n
- d.
, w--
;. . ... L[
~~ '
h x I,
- w. .
. m a . .. & .
.wmawqu%n[
- a. s y a 4
., .J.._. . ; ;[ ' _ - [ [ J [;_'l . : a [:. ~J; J/.;;LjL./ g.[.: Qj,p/ g.q_g.Q.)n;-m%
G,.. 4 . . , . g, ,, .- _; . - . . .- . ;
. .,_ . * ,, y
..y. QQQ.yymn.w:.,...
> ;;u.s ., e..
, Q.9 * ' ,
- .e, es ,t x . , , l
.g.;
. y ;g.g n.a'
- m p'u hk S .
;g .. . . .
'.. .: 3 - c
'; , : = s ~ a .. ; ; , :'
x .: 1 ..
~. _ y( .
'^
- .l l ' - . l. ~ L._ _.L... j . ~ [: i .l3: ' c I :' a :' :. :. L*lN';
}j:f'f,j yl?. Q }x Q g..'.q
- 41 7 Q ^ "
. $ .. 'yj '-
- >, i. .~ c
,,.~.<~,p.c... .wa ' . . . ;; : :
. . n y;w.
. .v,.s, ; ,w w7;.n.,9 y y[y p4 . ]. y . ' l]. {3.1;.;;(y .3..3{}p;;y
- f . .
f
.g; ._ 3 3:[x.:gqg y $c4;:jp q.mpp , gg $; y,.g., (
-~
_ _ ., p. 8 .
. ~. .. ;; ~ g . _ .;, , , .. ,
- 9. ( . - 7 . .
g g;. g ( g ig y g g g g g4"'
~ .
- ,mm g ' ' ~
. . E; j .-:/
. s ; j f. [ Jp7ggggw ?:mmmy sww '
a..
.g .
; -n .
. ?
. g,;x . .
TRoa (Acnw)
;y ;
;;%j , [, ; ; f.7,f Q[ ggM'Y.d.M.y.j? RETURN ORIGINALif '.4 S g.g . .. .
.. , : q .<c. D. ;s TO BJWHITE $
j - ;.. ;
.- ' . .; R 'o : ..L : " / ~~ : N. L. . . T. ' .y:y;jf,i
; b
- g . . ' '; ; ; ; ., [
- S
- f. v; -f 7.9};[c_Lg;;w lM/ST-2E26
<1s.,13o p&gf z a c .p g g rurugg; y pi z h,. y... .
, $ ; [ .. .
? "
J .
. . f .' . - , . ' p...
. . . :g g .
p.7,N A s .[i Q,;,..pg ypp;m mm s(
- m. y ... ...c..., u ,
- g. ) . ... ,y gg s.,g g.g
.,M. Q ,g / M i g M ] f.?% [.
- }.
i ,[;h,khkkhhk ;
~
- 4 ..
s n.g.;;n p??.. N T S $5.Md,N. i ~ k ~m(
- - ; . , v: y ;; ':.f.y9 i c 77,3;@. '.,
l.l.[
~ . .-
'?.~
n , . ; .. . ;.. : .. y..1.1
.c ; y . . ..f 1 e n ig , 9 i V-4
,; ( 3 .
.' , ;4 ..r
.. q y:p j j. ,g,g ;;; , , ,py .p.. ,
7, ,, ; i s jk ; ; j :l lq dy-n
. . (- . ' .:,,
Y
,-\. _l, [ ,,.{ [. ,. . . -
^D' 6-^
'W'. . " . ' ' ,
' l,'}::.{::,y 7,'k & g &a. ,d ,Y +
Dl *
. ? :lq y, . .; -
' . { Lj';5 [
.Mr . . * .
,% :. 4 L ; :)
W z :: . Y. :q . N
&, M b .$' ? y g a.
.2. j%,
G.: i;,f :..yp)
, .. . . .f
. [.yQ@.;fsp'f,t;% q;:f: a;fi*i ,.: g
( _. ty g
.>i (
g [ : )7.k' p(; . .. Q,;QQQ
- m.
p .. Q
- 1 . .m.
.[
m m .: ,) g 4., f w <;< g[ l g'Q; c r .. . , K 49 a .
;pf.f gp; u ; y ' ./ :1 %C. &g:.F,XV n n . : x w;@.ppc; g y Q:
.. . - J .
.:yu f,q'q_6 Q;
_' *l L- .jgy [h }l?l ' .
. ,: r _ . ,;e g^
D a y, [ef, ., ..
%_ , f} QP%g. R:i::[g. sy;{, m..%%g \8
^ ' '
a
^l (m. ' .l..Y..} fl:,k $N. )f. ffh. jy '
h
- e. a .xl?v u.enn.n
,p m ,lx ; mmyum.. .#
a - y a; , M y y.,
.m- . - y. 2 x m ,r ; .
%.) l '_ ,.e ';3 ] l' . . j ,. e,; - , _ , . ' / .::'. fd.Q f;_ q- (.:h -}@ 'V ' fs _. '. , .f _ _ m...
?
.a.y
;y.
. . . q: 3g ; :g ..g,:,.4
..c . . . .
. . q. , _-..-,p, .f . m,,. , .
j . Q{ %p.4gp 8.. .
} . ,: ' .g :{ + {. -{ . -
,.].
n @y ; , 4
.. . u . . ; - _. ; ( ; 3 2 :q v : ..
. d g'
- u ,:a. y'@ >'f C.:,. :. n,y>
- a. I '
. : . , u. .
h,~ ,y i, , , , i . : . . . . v
- c. } '
..e,
' , ;1 : - '[, g,: ,
;s.{
y ..' .
,/ gg
[g , ' ; f [g 7 ,W .
. . . .. . . . l . : :; y. :
f.l;fMy '
- y
, g? . );
.: ~g,
- 3. m w.%,gy_ . , ,; x,M ,
.. .., . ,.. m. ,;.g f.3gggg
- e. , ' .
i..* . , . - . t'M'WM ) f. :. [ y (# " 7W: W,6M 2 M 1. L L; )VWQ y$N -. @ % [hh.h$h,h[b'm.h..h.._N_h.h.k
. ~ .
;hh d %NM h N.[ .4 (.
[ 3 1 ,
v3 ":a N 3 7&- , , - -
=m* Q ..
< 8 N'.
.J.
g'og,a y .Mg-@rA ', A - l 4 , L 4
'> n ,1' d 'r c Q.
a
*p .C <
t Y't ( ~ - s gp m ;fi & -, _Q:.d r y #:Y P. 1 x hhY s 4 .c.< s ff. ;
; ~
i
~
i <
)
\
3 e ae
.Nj?
'g,z F
~4 YN$
a hY ,dn, n@Kh'_:p 4- 997Nnf f_HQ '3M:[OFFICI ETRANSCRI 6 OE:s i Q ::.4 ' ' ' ,.i kQ. > ?: '
&,h n?:n* *v ?
s f ,\ ' lN:? &Th] h<f[5. "Md: I , lQ n, r
;pm,m s;. x r% m w g' 'lb '
MsJ% t ;p. zwMp: i c.3 s 4!? L 4, - 4 a t
*y.
c,%l;*ygy 2 , - i
^
<, y:A: n t
t p:. n 3, ,:[' : Qy
- d. 'q' Q'%Q t
,c ,y ,
- x ,p., ' N .
"~,
W htrk, b g >
, g. , ~. y .
w w ANUCIlEdEREGULATORWCOMNIISSION: ' m.Mw. y w %. , - m& m # a %c
-m1 %, ~#
w o4 , ~ ,< M F S w.+~ ,eW E_ % x. mrVfSOIMCOMIIT.TEEIO'NIN_CEE wm # . Y Y f? [ RiWXST ~M um%d h WG,6 y ~ P 4
's . ,
.g-.
~b k] %]$: s
% , 2 p; e~u, ~n
,* q
',= ' h%n y
. .w ',', *, #^
< ' -w* -4 .
f: * ' -.: ' s
> m:e s
- m .,j.^
u y3.
. $x y ? s' w '
-" Q &' N %lf* 0v;\,,..*.%QQC": w. ,'g^$$p
' , i
..a ,'
en bW .',t'
.~
'z s
s >fi' r~~.: . . .%: i@a@@N,m&,# mag ATitle:
- v. ? -
S
^ 1 -. . 1 w m ,. 3105TH'ADVISORECOM,MITTEE ON?# 'sn,y Mkiw'l M Cw'+ M N#n~ ' [ . '
~
.J.{ . di t'F -
,M nn N s 7 # 5 ' %~ .
', +
1NUCLEnRN STE(ACNW) ~ d, kh'Mt & i W. . , . lcj
' ' %W(:["}
\
4%4MdME,2 % @/ W A1 - 5
+
s t , nu O < O T Whhh.?f,, m .,. 1 hWs i t.s3J 4
' s
#:' Y i,h E ;ACNW) b M$ iP t d$ TQD s .
RETURN ORIGINAL
+
' 1 WWW@m MDocket:No.J ' #" -e s$y m pn m, sm 2.m w n..
i- y c u m rc uwmre yw
,ed
.~ .
,, .. a - c m, .
, su
! k, h. ; - ' '
41b-7130 1 , .. o ? k.. i y,.t-. <-r[ t N 1 t i r;4 s t ._..t
;- 9 .. i 1 7 .
_ ( , . [ l y , g . . . , m$PWi x $# AVorkLOrdsriNa.MASBi300-595y x? s %m n an- - m $a$w r o 71 wgC A ,' M.J gp ~ E ~ 99u ,
'I
@[ h/ 4' ' ' W 1
. y F.k ,
s
- p~ v ..
"(
.[ > I f oN49Q
,NM,7 p...pfi,$p~ [m,A.
&+
JF M* rON.k4 3
; ,n'.
q / _ > - , y,. y pf.,,p e f wa. . - h
.s A. -
;g
,lW, z , V. '^+;ef 1
-m j;
;_ s, d ' ' .}e:. , ,1 t 4 4 g
g , y N s W ,
) g 4
& )"'
t , u , 4 %y;y. :, ',?.- . Ms. P GJ4'fMf 3;v ( , Qft? * *
' -rs 9% ; '* p \ 4 ?? '
QV yD s
, , f . ,
i N'd f f.;pg wal'y' f _ h3~w A l f).
.'M e g?~g
/ ,
p::yg y 'n, &a:e 0 wi v w',n c . ; ;:7 #p' ; ,e i: 7 .; 9, ::%s ;
+. s 97,, fl\h, ,c .p o% .:f:n >'
u
?r * ,
, j.
J 2
.O S .dm x4g ,i '
LMNa I:Qh _
, a
+ c. .
.e Y-W MMbyg, s;LOCATI'ON' %' .- - ..: -
- . Rock'ville,MD -.
h ' my I e wn , %
&%uQR
~m.
, >p c
s
^
- ) I vy. ,nw1e ++. v :4u: ;
gg ;n n .., v. - Mh y n 34.. h;f,u ' q;.
.DATEL
' Wednesday, December 161998 .
, t PAGES: 180 385 QfMyw$ %w..N6 C M;" 4 I ' '
1 $MFX@n < 9812230208 981216 PDP
. Wh e iO ; ADVCM NACNUCLE Md 1-O127 PDR j
&lW@@MM
&;[; >y P, -
,g ' ANN RILEY & ASSOCIATES, LTD.
1 i d 3%p%mp, , 1025 Connecticut Ave.,NW,Suste 1014 69 i j . m. l G " h' M. b. . 19 Washington, D.C. 20036
%Jl#. # '
"(202) 842-0034 ' j i
J ;
.i.c: 4 4 .
W'JHEOFEOFTW@@ Ny/0FFICCCOPY RETM@N@ l W W @Mhw s@:
.w hy3w,y e .
's$. Q@., N N[c.mM,1 i na 0 L f..M1t ; & n MN C D ' u. e n
~
s IMA:Q.m%f.g:.m:%l.N! t u.xQh@alI wW; Myc uV +/ wb'%s ' x;' v n. a 3'
1 rh L DISCLAIMER UNITED STATES NUCLEAR REGULATORY COMMISSION'S ADVISORY COMMITTEE ON NUCLEAR WASTE i DECEMBER 16, 1998 The contents of this transcript of the proceeding of the United States Nuclear Regulatory Commission Advisory (D (._) Committee on Nuclear Waste, taken on December 16, 1998, as reported herein, is a record of the discussions recorded at the meeting held on the above date. This transcript had not been reviewed, corrected and edited and it may contain inaccuracies. l 6 v i l
180 l 11 UNITED STATES NUCLEAR REGULATORY COMMISSION
.. 2 . ADVISORY COMMITTEE ON NUCLEAR WASTE 3 -***
4 105TH ADVISORY COMMITTEE ON 5 NUCLEAR WASTE (ACNW) > 6 7 U.S. Nuclear Regulatory Commission 8 Two White Flint. North, Room T2B-3 9 11545 Rockville Pike 10 Rockville, Maryland 20852-2738 11 12 Wednesday, December 16, 1998 13 14 The Committee. met pursuant to notice at 8:35 a.m. 15 16 MEMBERS PRESENT: 17 B. JOHN GARRICK, Chairman, ACNW ) 18 GEORGE HORNBERGER, Member, ACNW 19- E. CHARLES FAIRHURST, Member, ACNW 20 RAYMOND G. WYMER, Member, ACNW 21 22 23 24 25 l~ l if 1 ANN RILEY & ASSOCIATES, LTD. j . Court Reporters
- 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 (202) 842-0034
__ _ _ . . - ~ ._ _ 181 1 PROCEEDINGS p_) ( 2 [8:35 a.m.) 3 CHAIRMAN GARRICK: Good morning. The meeting will 4 now comes to order. 5 This is the second day of the 105th meeting of the 6 Advisory Committee on Nuclear Waste. The entire meeting 7 will be open to the public. Today the Committee will first 8 get an overview of DOE's Total System Performance 9 Assessment, and second, continue its discussion and 10 preparation of letter reports to the Commission. 11 Dr. Andrew Campbell is the Designated Federal 12 Official for the initial portion of today's meeting. At 13 usual, this meeting is being conducted in accordance with 14 the provisions of the Federal Advisory Committee Act. ( ,f 15 We have received no written statements or requests 16 to make oral statements from members of the public regarding 17 today's session, and should anyone wish to address the 18 Committee, please make your wishes known to the staff of the 19 Committee. 20 It is requested that each speaker use one of the 21 microphones, identify himself or herself and speak with 22 sufficient clarity and volume so that they can be readily 23 heard, and I think that for our first agenda item we are 24 going to ask Committee member George Hornberger to lead the 2r discussion, and I think given that we covered the current - O ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
182 1 affairs news items yesterday we will proceed right into the
. ,m t ) 2 agenda. George?
v 3 DR. HORNBERGER: Yes. Okay -- so this is the part 4 of the agenda where we are going to talk about TSPA. The 5 Committee has had a chance to read a number of documents but 6 we are going to hear overviews on several issues. The first 7 is going to be on the Repository Safety Strategy, and is 8 Jack here? Jack, you can introduce yourself, okay? 9 MR. BAILEY: Thank you, Dr. Hornberger. 10 Let me get familiar for a moment with the 11 mechanics. 12 Thank you. Thank you, Dr. Hornberger. I am Jack 13 Bailey. I am the Director of Regulatory and Licensing for 14 the M&O operating contractor, working for the Department of r (_,\1 15 Energy on this project. I am going to talk to you today 16 about the Repository Safety Strategy, and I have to swap my 17 slides here. 18 As an overview, what I am going to talk to is the 19 evolution of this Repository Safety Strategy and its 20 relationship to the post-closure safety case. I am going to 21 talk about the post-closure safety case and what the 22 attributes are to that as we see them, and then I am going 23 to talk a little bit about the Repository Safety Strategy as 24 it applies to developing a site recommendation in the l 25 licensing application or selecting those designs. I [ l ANN RILEY & ASSOCIATES, LTD.
' 'N # Court Reporters 1025 Connecticut Avenue, NW, Suite 1014
- Washington, D.C. 20036 l (202) 842-0034
183 1 The Repository Safety Strategy, which is available l) 2 in the room, and"I believe it's been made available to you 3- is in Revision 2 and has just been issued. 4 There are many' things in it. The parts that I 5 want to stress throughout the day and I'll point them out 6 have co do with the attributes of system. performance, and 7 that is going to be carried on as a theme. There is an 8 organizing structure here which we use throughout the 9 development of the performance assessment, and the 10 attributes of the system performance are going to be 11 discussed and inside of those attributes of performance for 12 the design that vas selected to be evaluated here there are
'13 a series of what we cal 1~ principal factors that explain the 14 relationship through the system. It is actually a physical 15- relationship which I will describe in a few minutes.
16 That is how we look at the performance assessment. 17 The post-closure safety case actually describes l 18 those aspects of w7rk that must be put together to make the 19~ argument in our mind to take this forward into licensing. 20 Those three big pieces are what we are going to talk about. 21 There are some other pieces that we will talk about in the 22 course of the talk but those are the major issues that we 23 want to cover. L
.24 The safety strategy provides this framework, this 25 organizing principle so that we all approach this the same i-gb ANN RILEY & ASSOCIATES, LTD.
Court Reporters
- 1025 Connecticut Avenue, NW, Suite 1014
[ Washington, D.C. 20036 l (202) 842-0034 l i..
184 l 1 way, that integrates site information, the repository design ! -[U) l 2 information, and the results of the performance assessment 3 of that site information and that design so that we can 4 focus on the information that is necessary to complete the 5 post-closure case. This is our organizing principle, to put 6 it into a certain form and then to understand and do an 7 evaluation of what it is that we still need to know. 8 Now it is organized by the key attributes of the 9 repository system, and I am going to digress here for a 10 moment. I have addressed this group several times in the 11 past as the Director or the Assistant Manager for 12 Engineering, and I normally talk in terms of an engineering 13 discussion, mostly inside of the engineered barrier system. 14 As the Director of Licensing, in my discussion
/}
( ,, 15 today I am going to talk about two things -- one is the 16 system, and I mean the system from when the rain forms in 17 the clouds until the water is consumed by the receptor 18 downstream of the facility, and I am going to talk about 19 design in a broad sense, and that broad sense is as the 20 entire design of the system. 21 I consider the overlying rock units to be a part 22 of the design, the repository system. Although we can't 23 really change them per se, we can learn about them and use 24 their attributes and their performance characteristics as 25 part of the system and so when I talk design I am not (m
\--
ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 ! Washington, D.C. 20036 l (202) 842-0034 i
. - .. .~. .. - -- .. -- - . . . -. . - . . - . - - . . - - .
185 1 talking-just'about that engineered barrier system, that 5 () 4 - 2 meter drift. I am talking about the entire system. 3 Now the key attributes of this repository system 4 4 are to limit the water contact in the waste packages, to 5 'have a long waste package lifetime, retain the radionuclides 6 inside of a controlled area as look as possible. Once that 7 is breached, if you will, then have a-low rate of 8 radionuclides from that breached package such that if we can 9 keep it inside or we can make it difficult for it to leave 10 the package we still have the radioactive material l 11 contained, and then as it leaves the package and flows ' 12 through the system out to where our potential receptor is 13 .then see what can be done with regard to concentration ! 14 reduction, and concentration reduction could be through j
) 15 several-forms. It could be through retention of material.
-16 It could be through'a dilution in the saturated zone or it 17 could in fact be a reduction because of the time associated 18 with decay, and many of the things we rely on here are 19 associated with the time of decay -- isolating it for a long 20 period, allowing it to decay, and then slowing the release.
1 21 Now the Repository Safety Strategy focuses on what 22 information is needed to complete this post-closure safety 23 case that I talked to, and we show on the chart here the big 24 arrow, which is the evolving Repository Safety Strategy, and 25 that is important that you note that it is evolving. 1
- ; ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1025 -Connecticut Aver.ue, NW, Suite 1014
~
Washington, D.C. 20036 , (202) 842-0034
186 l 1 As we learn more and more about the system, we ' fn i N_ -) 2 change our views about the system, we may change our views 3 about what attributes are. The principal factors which we 4 haven't shown you yet, but the principal factors may in fact ' 5 be different from system to system, depending upon what the 6 system design, what the actual parts of the system are, or 7 how we choose or how we choose to make them important, which l 8 ones we choose to be most important. l 9 The post-closure safety case, however, which is i 10 located inside the blue, which is virtually unreadable from 11 here -- I guess you will have to look at your page -- may I 12 have to do with five pieces. That is, when you assess the i 13 expected post-closure performance, and that is what the 14 performance assessment tool is for, it looks and says when gw/ (_, 15 we wicker this whole system together, all the natural and 16 all the engineered pieces, how do we believe the system is l 17 going to behave, and understand that so we have an 18 assessment. 19 The second thing is we look at how do we 20 incorporate margin and defense-in-depth? The performance 21 assessment does a nice job in giving us the general I i 22 performance and uncertainty associated with it. The idea of 23 margin and defense-in-depth is to provide some method 24 associated with the uncertainties and what those 25 uncertainties might mean to us, and I am going to have a l l [ '- ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l
i 187 l 1 talk later on in the day with regard to defense-in-depth. i 1 ( ) 2 We have to have a consideration of disruptive 3 events and processes, those things which despite our best 4 effort at understanding how the system works, is there some 5 disruptive event which completely throws everything out the 6 window, such that all the work on how the system normally l 7 behaves creates an untenable situation, so we have to 8 evaluate that. 9 We want to make use of the natural and man-made 10 analogs. What do we know from nature that has been around 1 11 for a very long time, or what has been made by man that has ! 12 a fairly long life that we can use to bolster our case that I 13 we believe we can reduce uncertainties and understand how- l 14 various parameters are going to perform and then finally we j ! r~s ! I,) 15 have to put with that the Performance Confirmation Plan,
)
i 16 which suggests how do we look at this on the back end to see ) l 17 if our predictions or our processes as represented in the PA i 18 are actually coming to pass. l 19 Now the Performance Confirmation Plan can be I 20 far-ranging. It can be a continuation of initial studies or ! 21 it in fact could be a confirmation plan of we believe this 22 is how that is going to happen and in this timeframe we can 23 gain enough information to have a higher confidence that 24 what we have chosen is actually going to work. l 25 CHAIRMAN GARRICK: Jack? L ; [ ANN RILEY & ASSOCIATES, LTD.
\- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 i
?
l-l 188 ) 1 MR. BAILEY: Yes, sir? l r"N
.2 CHAIRMAN GARRICK: You are emphasizing the
- !N J' 3 evolving aspect of-it, and I think that is appropriate. To j 4 help the Committee some as we go along, and maybe you are l ;
5 goi.g to do that, can you highlight or emphasize how things I 6 are changi>g and especially what has changed since we last i 7 heard from you? 8 MR. BAILEY: Yes. Yes, I can, Dr. Garrick. i 9 CHAIRMAN GARRICK: .Thank you.s i 10 MR. BAILEY: In fact, the next chart, which is 11 unfortunately also unreadable, talks about the evolution of j 12 the Repository Safety Strategy and I won't highlight it for l 13 too long. I'll point out that in the site characterization ! 14 plan we had what we call principle elements that were relied t' lA 15 on -- unsaturated rock units, the saturated rock below the 16 repository, potential repository, the engineered barrier 17 system, and we are going to characterize the disruptive 18 events and processes. j 19 We have new site information. We changed to a 20 dose-based. standard which changed our approach because it 21 was no longer a release standard, and we looked at a 22 modified design in its evolving nature. l 23 Rev. O suggested that the principal attributes of i 24 the system were seepage, containment, radionuclide L ! 25 mobilization, radionuclide transport and dilution, and we l l 'f\\- ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 I Washington, D.C. 20036 (202) 842-0034
. . . . =- . _. .. -. .
189 1 came.up with 18 hypotheses, which is what is the work we
/3
() 2 need to do to demonstrate this, 3 We then received new information, site data, We 4 improved our TSPA, which you are going to hear a good deal 5 about today, and the design matured and we came up with 6 another safety strategy, Rev. 1, where we_ combined the five 7 into four and came up into the four that I have just i 8 listed -- limited water contacting the package -- water is t 9 what tends to both move and corrode, if you will, the 10 engineered barriers, and move the radionuclides, make the 11 waste package last a long time, keep it in the package, and 12 then learn what we can about concentration reduction. 13 We revised the hypotheses to address these 14 attributes. We wrote it in a more precise manner of what is r (_)s 15 it that we needed to know, and there is going to be an 16 example of that on the next page, and then we thought 17 through a strategy that helps develop the post-closure 18 safety case. 19 What do we need to know just besides how the 20 unsaturated zone works and how the saturated zone works? 21 And then finally in this repository safaty 22 strategy, we went to the 19 principal factors, which is new 23 from the last time we talked about the repository safety 24 strategy, and we talked to a certain extent -- we 25 prioritized those and we provide some of the information A ( ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
l 190 . 1- .needs, which you're going'to hear more about today. 2 That came about because, again, because of new. 3 information. We knew more about the TSPA, some in situ test 4 results and material testing data. We knew more about the 5 system. 1 6 And so these are not what I would say massive L-7 steps and massive changes, but they are evolving. We lea _n 8 more.about it and we focus more clearly, which is what we do 9 here. 1
- - 10 Now, to make that very simple -- &nd you won't see l 11 a one-for-one line-up across the page here with regard to l 12 the evolution of the repository safety strategy, Instead 13- what you see is that the hypothesis for the seepage 14 attribute talked about the percolation flux at depth, the
[
- 15 fracture flow at depth, capillary retention reduces seepage. i I
! 16 As we move to Rev 1, we looked at the -- can we l 17 -bound the flux. The seep J' flux -- how much is the seepage l r
-18 flux? That suddenly becan,e more important to us as we got 19 our models down, and it was in this time frame that we felt 20 that there was probably more moisture in the underground l 21 traveling through the underground than we had thought at a 22 previous stage.
23 We looked at what happens because of the thermally t 24 induced seepage, and you'll notice we're focusing very much , 25 on seepage, and that is the moisture that's in the [ t ! l t I I g ANN RILEY & ASSOCIATES, LTD. d Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202)'842-0034 l h, , ,, -, -
191 1~ underground, ma have a pretty good handle on, but how does () 2 3 it really get into the drift and what does it do to the package? 4 And then-finally we looked at, for example, the 5' precipitation -- and these are the principal factors for 6- limiting water'-- the precipitation -- how much comes down 7 actually from_the rain; the percolation to depth -- what do 8 the~ overlying rock units do in order to limit the amount of 9 water that gets down? When it gets to the drift, what 10 ' really seeps into the drift? That's a key issue for us. 11 When it does seep into the drift, how and where 1 12 does it drip onto the waste packages? And then finally, 13 what's the humidity and temperature at the waste package? 14 What's the environment that this waste package has to sit l 15 in? 16 -So you can see, I think, from those three that we 17 looked at what were the things that we needed to know, until-18 you get to the last one, and you're more down to, frankly, 19 an engineering approach of what are the specific parameters 20 we need to know in order to track and calculate what's going 21 to happen at each one of these steps. So we've gone from a 22 general approach and slowly worked it down into something 23 that's more direct.
- 24. Now, our basis for the repository safety strategy 25 is that the majority of the radionuclides in the repository i
I' ANN RILEY & ASSOCIATES, LTD. l (' Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
l l l 192 1 are not particularly mobile in the Yucca Mountain n 2 environment. The bulk of the radionuclides will never move.
-(v) 3 The site'is such that it will contain them. They are either j 4 insoluble or they sorb strongly to the type of rock that i 5 exists down there.
6 The remainder, however, could be transported by 7 water movement, and that's what we have to face with regard 8 to the system as we put it together. 9 However, the natural features are favorable for 10 limiting the amount of water movement, how much water 11 actually comes in, and the site, we believe, provides a ; l 12 predictable and stable environment for engineered features 13 to further limit the water movement. 14 So we understand the environment we're working in (. ( ,) 15 and we can take advantage of our understanding and create a i 16 system that works in concert with the natural system. 17 Now, in revision 2, the performance assessment 18 that had been performed was then exercised, if you will, to 19 provide sensitivity studies that provided us a basis for 20 testing our hypotheses and finding what was most important 1
-21 to the system. There are parts of the system that are '
22 extremely important and parts of the system that are not as 23 important to the system. And the hypotheses in Rev 2 have 24 been replaced by these principal factors, and I'll talk to 25 those in a page or so, what they specifically are. l l I \ ANN RILEY & ASSOCIATES, LTD.
\-) Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036
- (202) 842-0034 l
I l
- _ ~ - . - . . .- - . _ _ -- - -
193 1 The potential to reduce uncertainties for those l principal factors important to performance was used to (s~') 2 l 3 prioritize the technical work for the license application. 4 Those items which really make a difference are what we want 5 to study, and those items which we probably have in some 6 'means bounded or an understanding of are probably not where 7 we want to place our resources, at least not as much as 8 those areas that are as important or are the swing, if you 9 will, of what the end result can be. 1 10 In the RSS, we have summarized the elements of the 11 post-closure case, which establishes basis for showing the 12 repository system will protect the health and safety of the I 13 public, and we're back to the five portions of the 14 post-closure safety case. j (Oj 15 Now, as an example of the principal factors, 16 hypothesis four, seepage into drifts, will be a fraction of 17 percolation flux. What we have found moving into Rev 2 is 18 that a seepage model we put together fits that preliminary 19 data from a liquid release test to niches in ESF, and that 20 will probably be talked about a little bit more this 21 afternoon. But where we were thinking of seepage into 22 drifts will be a fra eion, we are now approaching the point 23 where we have a model to where we feel we probably can l 24 bound, limit and discuss what's going to happen with seepage l 25 as opposed to identify the fact that we know what's going to l l [} \> ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 i
194 1 happen, what.will happen. ( 2 Our preliminary estimates appear consistent with 3 ~the hypothesis that,-in fact, it is only a portion of the 4 flux, and the analysis continue to indicate that seepage is 5 a principal factor; that is, when we do the sensitivity 6 analysis, we find that the seepage is a very important piece 7 to the overall effort associated with water contracting and 8 waste packages. 9 Now, our approach is to understand the required 10 performance of the. system. . Performance assessment, which is
'll the first attribute or the first element of the post-closure 12 safety case, understand from the performance assessment 13 what's the required performance of the system, what does 14 each part of the system have to do, then identify principal
() 15 factors that affect the performance of the system and break 16 it down'in a manner such that we understand that 17 performance. And our method of breaking this'down was, in 18 fact, to take a drop of water and walk it through the 19 mountain, and a drop of water in a rain cloud until it hits 20 a receptor -- what happens to it, physically and in a 21 process manner as it walks through the mountain? 22 You're going to see in another talk how we relate 23 the models of the PA to how the principal factors work, and 24 this was done in an effort to try and gain transparency and l 25 to try and identify where we really need to do the specific l i f' . V[~ ) ANN RILEY & ASSOCIATES, LTD. Court Reporters
- 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 (202) 842-0034 l
._, .-. .m -.-
n i 195 o 1 li pieces of work to move forward. () 2 Then we did sensitivity studies and uncertainty l
'3 analysis to find the'importance of those principal factors 4 and what their effect was on overall system performance,
.5 which onesLwere most important, which ones were lease l '6 important.
l
'7 Now, what that says -- and this chart is a little
- '8 hard to read, so I'll point for a minute -- the top line'in 9 each is in fact the attribute of performance and the hash 10 marks show what the principal factors are. And I already 11 ran through the first one.
12 The first, as 1 said, is the attribute of limited 13 water contacting the. waste packages, and you can see we 14 tried to walk through that drop of water and how it moves. 15 The long waste package lifetime -- you go to the chemistry J 16 of the water on the waste package, the integrity of'the. 17- outer carbon steel waste package barrier. And this is, 18 'again, design dependent. This was our design of an outer 19- package of carbon steel, an inner package of corrosion 20 resistant, and then the integrity of the inner corrosion 21 resistant package.. 22 Then you go to the low rated radionuclide release, 23 .which says once you breach the package, how does it get into l 24 the. package, how well does the clad protect, and we chose in l 25 'this model to use the clad because it was present, and 1-
-[ ANN RILEY & ASSOCIATES, LTD.
Court Reporters , 1025 Connecticut Avenue, NW, Suite 1014 L
. Washington, D.C. 20036 (202) 842-0034
196 1 rather than ignore it and make a very conservative ! [5_'/l 2 assumption it doesn't exist, we chose to include it. And 3 then we looked at how does the spent fuel in the waste form 4 actually dissolve, how does it get into the system, what's 5 the neptunium solubility, because neptunium in the longer 6 time frames is very important, so it becomas a very 7 important parameter, the formation of the radionuclide 8 bearing colloids, ¬her transport mechanism where not 9 solution, but into colloids, andthenatranspbrt through 10 and out of the waste package. 11 So a lot of time spent getting the water there, 12 how do we break the waste package, how do we get into it, 13 what happens inside the waste package and what comes out, 14 and then finally, the radionuclide concentration reduction
) 15 during transport -- it has to move through the unsaturated -
16 zone to the saturated zone. It is, in fact, diluted as its 17 brought up from the aquifer during pumping. And then 18 finally, how is it actually transported within the biosphere 19 and uptake by the receptors when you get there. 20 So we look at all of these pieces and broke the 21 performance assessment. Since the performance assessment 22 starts at the precipitation, ends at the receptor, we broke 23 it down into these pieces to understand which ones are the 24 most important in where we should place our efforts. 25 Inside of the RSS, you find when you read through I ['\s )/ ANN RILEY & ASSOCIATES, LTD. Court Reporters i 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
197 1 -- and, in fact, inside the RSS, we go through it on a ]'~)h N_ 2 principal factor basis grouped by attribute, and you will 3 read in there that it talks about what's the significance of 4 this principal factor with regard to its uncertainty. How 5 important is it to know about this? 6 And we took an approach, fairly mechanical, but we 7 basically said that if we were to take the expected range of 8 uncertainty for this parameter, for this principal factor, 9 and were to expand it from the bottom of its range that we l 10 believe it's in to the top of its range, if you will, does 11 it create more than a change of a factor of five in the 12 overall end result, then we considered it to have a low 13 impact. 14 If it was a factor of five to 50, then we p) (, 15 considered it to have a medium impact. And if over that 16 full range of uncertainty, it made a difference to the 17 overall answer by a factor of greater than 50, then we 18 considered it to have -- those uncertainties to have a high 19 impact. And we needed to go back and do more work to look 20 at what the uncertainties were, can we in fact reduce the 21 uncertainties or understand, because you can't always reduce 22 uncertainties, unde stand those uncertainties better. 23 DR. GARRICK: Did you do the same thing for 24 different time intervals? 25 MR. BAILEY: We did it for different time [)
\~/
ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
198 1 intervals, yes. We did it -- in fact, I believe Holly rx 1 2 Dockery will talk about that. And we looked at that, we _ (d 3 created the principal factor, and it isn't in mine because 4 I'm talking to the RSS. Holly will be able, I think, in her 5 presentation show you some charts that actually brings up 6 the importance of each factor in three time frames, in the 7 first 10,000 years, in 10,000 to 100,000, and 100,000 to a 8 million years. And we talk about what happens in each 9 arena. 10 MS. DOCKERY. Mostly Ernie. 11 MR. BAILEY: I'm sorry? 12 MS. DOCKERY: Mostly Ernie. 13 MR. BAILEY: Mostly Ernie will. Okay. I'm sorry. 14 And in fact, we'll go through and look at exactly l gy ' () 15 what you asked, Dr. Garrick, which is what happens in each i 16 time frame, because it's entirely possible that the 17 signific?.nce of each item -- it changes-from time frame to 18 time frame. In fact, that's why Neptunium shows up. It's 19 not an important factor in the first 10,000 years, but it l 20 becomes an important factor in the 100,000 year time frame. 21 And so we did it by time frames, we did it by different 22 factors so we could understand where to focus our attention 23 and on what time frames. 24 Now, the next thing in the repository safety 25 strategy that we do is we look at various design options, I ('k _/
') ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 L Washington, D.C. 20036 (202) 842-0034 l l l
l I 199 l l l 1 l and we chose three that we had looked at and the three had l [~)V 2 to do with the use of -- actually, we used two. The third 3 was backfill. We looked at a drip shield and we looked at a 4 ceramic coating. And I say a third because in order to use j 5 those two, we thought we had to put a backfill in to provide j 6 mechanical protection for both the ceramic coating and the I ! 7 drip shield. So it looks like three, but we didn't take ' l 8 much performance, I don't believe, for the backfill. l But we basically took a couple of options and ran 9 l l 10 those with limited data because they were not fully l 1 11 developed, with limited data. We ran those through to look l 1 12 at what happened to the system as we added options and, in l l l 13 fact, affected the importance of the various principal I i 14 factors or we affected the process that the system was going ( ) 15 to see. v 16 And as it says, could they improve the system 17 performance and reduce remaining uncertainties. In other l 18 words, could we move the uncertainty associated with 19 seepage, which is where is it going to seep and how much is 20 it going to seep? Could we put in a drip shield, for 21 example, that makes it irrelevant of knowing exactly where 22 it's going to drip or exactly how much it's going to drip, 23 and in fact create that sacrificial barrier so it could drop 24 anywhere, and now the uncertainty associated with seepage l 25 might not be as important because of a design feature that (~D ANN RILEY & ASSOCIATES, LTD.
\_s/ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 I
Washington, D.C. 20036 t (202) 842-0034 l
l 200 l1 .we put in'. [q/'Y _2 It's important ~ to the drip shield, but if we treat 3 the. drip shield as a sacrificial, then perhaps we don't need 4' to know as much about the seepage as we would if we depended 5- on knowing when and where the seep was going to hit a 6- package in particular. And so'that was the strategy that we 7 used'here was evaluating a couple of different options. 8 And the options that we chose also we used the 1 9 same metal, being a reactor historical type. We used the 10 same metal in our approach for the' drip shield, so now you 11 had, if you would, a common cause or a common mode in that 11 2 ' the same material that was a. corrosion-resistant is now the 13 same material._that is the drip shield, and so you basically j 14 have a bigger package to a certain' extent. () _15 On the other hand, we looked at the ceramics, 16 which is a different material and fails in a different 17 manner and has a different failure mechanism. And so we
- 18. looked at that again to add to our thought process i 19- associated with defense in depth, which I'll talk about a l 20 little' bit later.
21 So we showed that they had the potential for 22 improving system performance, but they of course change the 23 system. And we have a chart here which suggests from what
.24 sua call a base case is the top line, as a result from a 25 reference design, followed by placing the drip shield on,
[] 1 \r ANN RILEY & ASSOCIATES, LTD. Court Reporters l 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
> _ . . _ . _ _ _ . . _ _ _ _ _ . . . _ _ _ _ . _ . _ _ . . - . ~ . . _ . . _ _ . ._ - . _ _ _ _ . _ .
t 201 4 l l 1 which you can see delays a great deal what.happens.to the ; 1 1 2- waste package, which has both to do with the kinetics j 3 associated with the corrosion of the waste package and a } 4 delay again, because you decayed off more radionuclides, and I 5 then the ceramic coating, which pushes the system very far 4 6 to the'right in time, as it has a different failure { /7 mechanism. So you can see that the options can very easily 8 make a difference if we can justify the uncertainties in the 2
- - 9 options.
F 10 Now the options were not as well developed as the 11 main case,'and they probably have a great. deal of 12' uncertainty that we have to work on, but when we looked at 4 13 what work do we want to do to get ourselves ready for the i t 14 license, we saw something like a drip shield or ceramic
) 15 coating, and in fact there are many, many more options that s
16- we can have that we needed to consider those as we worked 17 .our way forward. 18 I'm going to talk to the postclosure safety case t' 19 for a couple of minutes, and this is the information that. l I 20 . DOE intends to use in the regulatory setting to provide a 21 . reasonable assurance that the repository at Yucca Mountain I 22- would adequately protect health and safety of the public. ). , 23 And again, as I said earlier, we focus on the 1 24 elements that we believe are necessary to put together an 25 argument, and that's what this is, is to put together an , O ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014
-Washington, D.C. 20036
-(202) 842-0034 I
202 1 argument that says we understand how the system is going to n 2 work, the assessment of the postclosure performance, and 3 then an incorporation of the safety margin and defense in ) 4 depth to mitigate the uncertainties. 5 Clearly there's a good deal of uncertainty I l 6 associated with this entire system since these natural j 7 processes, nor the natural processes associated with the 1 8 natural setting, or the natural processes associated with l l 9 the behavior of the engineered system is not precise. 10 There's uncertainties associated with all of them. 11 And so what we wanted to do was make sure that we 12 had the ability to either have margin or to have a means of 13 examining from a defense-in-depth point of view, which is a I 1 14 reactor term, if you will, and again I'll talk about it this 1
/~
( ,N) 15 afternoon, but how can we make sure that we don't have any 16 major vulnerabilities? How can we be sure that this 17 performance assessment will work? And then we look at the 1 1 18 disruptive processes and events, is there something that can l 19 undo the basic performance of the system, look at the 20 natural and manmade analogs, make the argument, and then a 21 performance confirmation plan, as I discussed earlier. 22 Our major focus as we move to the license l l 23 application is postclosure safety. This is not to minimize 1 24 the preclosure safety. The handling of fuel, the packaging l 25 of fuel on this basis hasn't been done on this scale before, ; [~}
\-
ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l
I l 203 1 but the general tech'niques that are going to be in use have
-, 2 inLfact been used, and we'believe we understand how to'do l 3 .the'preclosure. aspect fairly straightforwardly.
.4 The question is to focus on the postclosure 5 safety. As I said before, a large fraction of the radionuclide inventory decays rapidly or is essentially
~
- 6 7 insoluble, and they're not of great issue, so once again we 8- don't want to protect against things and spend energy and 9 resources protecting against items which are not going to 10 contribute to the final result, and so we look at that small l 11 fraction of the radionuclides that's mobile to see how we l
12 can deal with that particular set. I
- < 13 As I said,.certainly uncertainties remain 14- regarding the-characteristics of the repository system.
'( ) 15 Large uncertainties still remain, and as they influence dose 16 rate. And you'11'see in'some of the other charts there are
^ 17- several items which in fact.have a high or greater than~50 18 effect over their range of uncertainty. 19 When they're addressed, and notice I didn't say 20 reduced, when they are addressed, when we understand them, 21 and we can put the five aspects of the postclosure safety 22 case together, then the postclosure safety case will be 23- complete. And so there is a tradeoff here with regard to 24 knowing everything, choosing a design to mitigate those
.25 uncertainties, and considering the evaluations associated 1
ANN RILEY & ASSOCIATES, LTD.
- Court Reporters 1025 Connecticut Avenue,'164, Suite 1014
- Washington, D.C. 20036 (202) 842-0034
- r -., -
204 1 with defense in depth.
/} 2 We are going to do testing and analysis to address 'n 3 the uncertainties in both natural and engineered barriers, 4 and we're going to look at design options and alternatives 5 as necessary to select a design for this mitigation of 6 uncertainties to make sure that all the elements of the 7 postclosure safety case can come together.
8 The remaining technical work emphasizes the 9 highest priority principal factors, those which make the 10 biggest diff ~nnce. We need to update our performance 11 assessment models using improved process models as 12 appropriate, and the results of the design alternatives 13 study. And what we are doing right now is we have a process 14 in place where we are looking at what we call both design (O,) 15 alternatives and design features. 16 Design alternatives is a fundamentally different 17 design than we've presented in the past, one which may have 18 a low thermal load, may in fact locate the repository at a 19 different elevation, take greater evaluations with regard te 20 ventilation to control heat, for example, and then a series 21 of features where we have different engineered additions 22 which could be added to basically any alternative to improve 23 it, for example, the drip shield with ceramic coating. And 24 we're walking through all those alternatives and all of 25 those features to try and come up with a design that will
/"'T ANN RILEY & ASSOCIATES, LTD.
C/ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 l (202) 842-0034 l
205 1 satisfy our needs with regard to performance assessment, the
/N 2 defense-in-depth margin questions, and the reduction of d
3 uncertainties in those tradeoffs. And these criteria, if 4 you will, the approaches that we're using here are factored 5 into the criteria of that evaluation. So as we go through 6 the design alternatives and we look at the models, that's 7 how we control the work that we need to be doing to move
.8 forward.
9 I talked about margin briefly, and that's to 10 ensure that we have sufficient design margin to account for 11 the variability in properties, and uncertainty conditions to 12 which the barriers will be subjected. We have not developed 13 this fully at this point in time. Does that mean that you 14 make the waste package an extra half a centimeter thicker? () 15 My view of that is in the performance assessment 16 you account for that. Maybe we have to because of 17 uncertainties. So we make it-a little thicker. Defense in 18 depth, on the other hand, suggests that we need to look and 19 find what the worth of each one of the barriers is, and are 20 we solely dependent on significant barriers such that if'we 21 are " wrong" or it is in fact at one end of the range of 22 uncertainties. Could it have a very large effect, and does 23 that large effect in fact prevent us from satisfying a 24 standard? And we need to do that evaluation so that we 25 don't have any sole vulnerabilities. So we look at it from f\ ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
l 206 1 a defense-in-depth point of view.
~'\ 2 The disruptive processes and events, we are going
(& 3 through and updating our current status on those with regard 4 to vulcanism, seismic hazard, and human intrusion, and we 5 note in Part 63 draft that's available on the web that there 6 is in fact a new stylized human intrusion scenario which we 7 are reviewing. We believe that in the vulcanism and seismic 8 hazard arenas that we're coming close to closure. 9 Further evaluate the probability and consequences 10 of postclosure nuclear criticality. That's an ongoing piece 11 of work that we are still doing with regard to disruptive 12 processes. 13 Natural and manmade analogs, as I said, the areas 14 in particular that we're looking at, and Ernie will talk () 15 more to this, solubility and colloid facilitated transport 16 of radionuclides, retardation characteristics of the 17 alteration materials and the fractures, infiltration in the 18 shallow unsaturated zone, and then behavior of the manmade 19 materials. And if you have specific questions, as I say, 20 Ernie will be able to go through some of the specific pieces 21 of work. But those are the areas where we believe that 22 there is information from the natural or the manmade setting 23 that we can help ourselves with reduction and understanding 24 of uncertainties. 25 And then finally again the performance ['} A_- ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C.. 20036 (202) 842-0034
207 1 confirmation plan to revise the existing plan taking into
) 2 account the design of the license application. Our current 3- performance-confirmation plan does-not necessarily recommend 4 what we should do for performance confirmation. What it l 5- does do is go through and identify what seems to be 6 everything that_we can reasonably measure, and is it 7 something that'we would need for the performance assessment.
8 So we are in fact comfortable that we can gather information 9 for the major attributes or major portions of the 10 performance assessment analysis. 11 We haven't selected what the right plan is in 12 terms of how to measure or what to measure. And we will be 13 doing that once we choose the license application design, i 14 understand the uncertainties, and a performance confirmation () 15 plan may in fact be necessary. For example, I guess an easy 16 example would be the waste package material. 17 We've done a few years of testing, which you can 18 do in an accelerated manner for materials to get an idea 19 what's going to happen to the material. On.the other hand, 20 it may be appropriate to spend a 10- or 20- or 30-year test 21
- of material in not quite such an aggressive environment to 22 . find out what happens with regard to that material. In that 23 regard, the performance confirmation plan may in fact 24' . include new' testing.
12 5 - On the other hand, we may be interested in just g ANN RILEY & ASSOCIATES, LTD.
% Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
l
)
208 1 continuing the rain'and finding out how the rair. is going to ()L 2 go and not do any more work with regard to gathering 3 information with regard to precipitation around the world. 4- We may conclude that all we want to do-is, you know, do a 5- spot check to see that something untowards doesn't undo our 6 data,- And that's a decision that we can make with regard to 1 7 the performance confirmation' based upon the level of 8L uncertainty and'our confidence in what the work is that we 4 l 9 have done. Juld so 'those that 's the work - that - we will be I 10 doing in the future. 11 Our system evaluation for the site recommendation, 12 the LA, again is to go back to the repository safety i
- 13 strategy and look at what are the system attributes, what L
14 reference design 1:s in use, and what principal factors does () 15 that reference design portend. l 16' .I've talked about 19 principal factors today with l 17 regard to a reference case. Were we to put a drip shield 18 in, then there would be first of all another principal 19 factor associated with keeping water away from the package, 20- and if you went back with that performance assessment system 21 and graded the importance of each of those principal factors 22 with now one more attribute in the system, you would find 23' that the uncertainty importance associated with the [ ~24 attribute may very well change. For example, the example I l- 25 gave with regard to seepage. Other ones could very easily h
/ \ ANN RILEY & ASSOCIATES, LTD.
\sbl Court Reporters
- 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 l (202) 842-0034
._ y _ . - . _ . _, ~o - _ , . . . . . , , - , . .-- . _ ,
.~ - , . ..- .- _ _ _ .
209 1 be affected as well. And so what you have to do is take it
-w
) 2 on a system attribute, a reference design, and a 3 principal-factor basis and work through that.
4 Now, going to the right on the chart, you then 5 would evaluate your features, evaluate your alternatives, 6 which is what we're doing right now, and integrate those ) 1 7 design featrees and alternatives into the natural system, 8 and you cre. exactly what I just said, which is you create j 9 your performance assessment and your principal factors for l 10 each set of alternatives and designs. 4 11 The factors really don't change. The factors 12 merely come in or out of the system depending on what the 13 components of the system are, but their importance can 14 easily change, ia ( ,) 15 Then you identify the set of principal factors 16 approach to each combination of engineered and natural 17 features and you have to do this for each system, and then 18 from your Repository Safety Strategy, looking at what you 19 are trying to accomplish, you come and you evaluate those 20 principal factors with regard to their importance to 21 performance, their importance of their uncertainties, and 22 then from that you select a design, and then of course that 23 becomes the site recommendation in the L.A. system design. 24 So this is a systematic process of select design, l 25 select what you are trying to accomplish, and we are working i (,,~) ANN RILEY & ASSOCIATES, LTD.
\~/ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 (202) 842-0034 l
m <-A an. 8-8h 210 1 inside the four attributes of performance that I talked , .2 about.early-on, and then you select the design. 3 The natural system remains fairly constant with 4 what it does, if I can use that as a verb -- I guess.it 5; is -- what the system does,'but then you look at what the different engineered features-are that you can put cn it. 6 7 You look at how that drop of water travels and 8- what it encounters, and you look at what the models are that
'9' you put together and you look at the importance of each one 10 of those models and factors now, and as you do that, ybu do 11 that for each one of the systems, you can look at what is 12 important to the system, which should be of interest because 13 you find out whether or not you have that vulnerability --
14 something that really makes the whole system work or not b g ,/ 15' work -- and that is not a good design in our mind, because 16 if you are wrong or you extend to the -- and then you look 17 at the importance of the uncertainties, if you extend to the 18 range of the uncertainties and that particular piece of the 19 system fails, then you have got a problem, so you go through 20 this in a very systematic method and that is our approach, 21 using the framework of the Repository Safety Strategy and 22 coming up with the design. 23 In summary, which I probably just did, we use it 24 to report the evolving approach -- that's what the blue book 25 is for, so people can understand where we are today and what ANN RILEY & ASSOCIATES, LTD. O( / Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
. . - . . . . . . - - . - ~ - - - . . . . . - - . - . . . . - . - - . . - . . ~ _ . . . - . - - . - . . . . - -
f 211
- 1. we are thinking and how we are going to move forward, and to
() 2 3' update the licensing case and decide what work we have to do. 4 We use this very much in determination of our work 5 of how to move forward with regard to the design, and that 6 is,the system design that we select. 7 I'll be happy to answer questions. 8 DR. HORNBERGER: Thanks very much, Jack. I am 9 sure that th'ere are going to be lots of questions today. My 10 guess is that we are going to want to jump into detail. To 11 the extent that you want to defer detailed questions until 12 after presentations, feel free to do so. 13 Questions? John? 14 CHAIRMAN GARRICK: Well, kind of taking a top-down () 15 view, and looking at your four key attributes,'are you able 16 to extract from the performance assessment, work some sense 17 of the role of each of these attributes in terms of specific 18 parameters and uncertainties in those parameters? 19 I guess the question is are you able to assign to 20 each of the four key attributes for example, looking at it 21 from the top down perspective for a moment, a L 22 quantification, some sort of quantification based on your 23 current state.of knowledge, and that may be something that l :24 Holly or some other presenter is going to cover later.
-25 MR. BAILEY: -Holly is nodding at me that she will i
i' i ANN RILEY & ASSOCIATES, LTD. 1.O Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l: is {
1 212 I 1 cover it specifically. () 2 I canialso talk about it and I will this afternoon 3 briefly in terms of defense-in-depth and we are looking at ! the defense-in-depth approach of can we show l-
- l5 defense-in-depth in each of the attributes as-opposed to i
l :6. 'each of the factors? 7 So I think Holly has the answer with regard to 8 quantification and I think I.can tell you that our approach L
'9 is in fact.to look at each one of those as a building block.
t
-10 ' CHAIRMAN GARRICK: I think in order to give us a l l'1 ' sense of whether these factors can really relate to each 12 other or be compared to each other, we need to have some l 13 measures that give us -- it could turn out for example that
- 14 one of the four is a hundred times more important than the 15 other three --
16 MR. BAILEY: Yes.
- 17. CHAIRMAN GARRICK: And that is the kind'of thing i
18- that I ac. really asking. l 19 MR. BAILEY: That in fact is particularly being l [ 20 . looked at with regard to the defense-in-depth, and I think 21- that Holly will also talk to that as well. 22 CHAIRMAN GARRICK: And another question that I l E 23 have is also going to come later, but with the increasing 24 dependence on the waste package design and the design in the 25- near field for reposicory safety, has that translated into a J ANN RILEY & ASSOCIATES, LTD.
's. Court Reporters
! 1025 Connecticut Avenue, NW, Suite 1014 I Washington, D.C. 20036 (202) 842-0034
213 1 deemphasis, shall we say, in site-specific date or data on r'N i 2 the natural system and incorporating that in the performance' l t G i 3 assessment and almost a very biased emphasis, if you wish, 4 on the near field and the engineered aspects of the 5 repository? ; i 6 One of the concerns I think that exists is that it I 7 is not as clear that the site-specific data is manifesting 8 itself in the performance assessment, particularly over the 9 time period that is anticipated to be the time of compliance 10 as in the past because of the increased safety case, if you 11 wish, being dependent upon the near field and the engineered l l 12 systems. 13 My question has to do with site-specific data and l 14 its role, especially in regard to the natural setting. 3 l
'( _,) 15 MR. BAILEY: I guess I have two or three pieces to 16 the response. The first is that, as I opened my 17 presentation, we talk about an entire repository system and 18 that system includes the natural setting as well as the 19 engineered setting.
20 The natural setting has been studied for many 21 years, as we know, in order to gain that understanding of 22 the environment where the engineered setting, if you will, 23 we'll believe, and so not to say that the scientific study 24 is ahead but I would say that the engineering study is 25 probably behind to a certain extent because until you i fs ANN RILEY & ASSOCIATES, LTD. (\--} Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 i
c - 1 214 1 understand the environment that you are in, the specific / ,, () 2 engineering aspect of it cannot be as fully developed, and 3 so we are of course placing an emphasis on the engineering 4 aspect at this point in time in order to make use of our 5 understanding of the scientific setting or the natural 1 6 setting. 7 Now we are continuing to gather information with 8 regard to the natural setting and incorporate it into our 9 model, so that we in fact have this balanced case thet I 10 have tried to discuss, and that balanced case includes what 11 is happening in the natural setting, what is happening in 12 the engineering setting, and what is the importance of each 13 one of those factors or the attributes, if you wish, what is 14 of the most importance with regard to effect on performance p) (_ 15 or the uncertainties associates with that performance. 16 As I said, we can never completely reduce the l 17 uncertainties associated with what occurs. We are talking 18 about a several square mile area of several hundred meters 19 of rock to which we will never be able to have a 20 deterministic value and so there is uncertainty associated 21 with that. 22 The cost of characterizing that to a preciseness 23 of deterministic case, you know, you go back to the old case 24 of let's mine it out and put in an engineered one, because 25 that is the only way you can know everything that is out l [, ANN RILEY & ASSOCIATES, LTD. 1
\~ Court Reporters l 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
- . _ _ ~ ~ , . . _ . . _ ~ _ . . _ . _ _ _ - . . _ . _ _ ..._.__.. ... - - _ . _ _ . _ _... -.._.-_. . . _ _ . _ .
t 215 1 the,. but'we can work within those uncertainties and there L() 2 is a level of uncertainty which turns out to be acceptable 3 and that level of. uncertainty is on a system basis, it's at l l 4 the attribute basis and then we look at it with regard to 5' how important it is on a performance factor basis, and what l -
- 6' the uncertainty of that performance factor is, and then we L
L -7 use the defense-in-depth approach, which'we will talk about i 8 a little bit this afternoon, to make sure we don't have a- ' l 9' vulnerability by not having, if you will, full and complete' I 110 . knowledge.in this specific arena. l
.11 So the answer is I don't believe that there is a )
12 deemphasis or an emphasis in.either area. What we are l 13 .really trying to do is come up with this even approach, as I 14 describe it, gorerned by the performance,- the importance of k 15 uncertainties in a defense-in-depth approach to mitigate, if I 16 you will, the importance of those uncertainties. 17 CHAIRMAN GARRICK: What do you think has been the l 18 -primary change in the safety strategy in the last five l L 19 years? t. 20 We used to call it the Waste Isolation Strategy, L 21 but it has been a subject of continuing interest and very ; 22 ~important. How would you characterize the most significant 23 changes in the strategy? 24 MR. BAILEY: Well, I personally can speak from L 25- about four years, not quite five -- four years, but I hav-ANN RILEY & ASSOCIATES, LTD. Court Reporters i .
.1025 Connecticut Avenue, NW, Suite 1014-Washington, D.C. 20036 (202) 842-0034
~ ._. .. . m.-. _ . . - _ . _ _ . . _ _ . . _ _ _ _ _ . _ _ _ _ . _ . _ . . _ . . . _ . _ . -
l-l 216 t I read the material. () -2 3 I think-that four years ago the Waste Containment and Isolation Strategy, as it was called, in fact was more 4 focused cn1 what is.it that'we need to do, and that the 15 strategy now has focused on an approach to accomplish that. 6 I think that is where the major differences come, and itHis 7 evolving, as we said before. What do-we need to do in order 8 to be successful? I think now we have charted a path that 9 if we do this, we believe we will be successful, so we 10 focused the work and focused the activities and focused the 11: approach much more closely. I 12 I think I will let my colleagues -- Ray? 13 DR. WYMER: I have one observation, maybe. ; 14 Certainly the things that you have discussed and emphasized
- i
() 15 are vital to the performance of the repository, but now that ' 16 you have gone, maybe, to as much as a 300 year preclosure 17 period, it seems.to me that the operations and the operators
. 18 during that period become exceptionally important in a 19 licensing application.
)
20 I don't see much discussion of that aspect of this 21 whole thing. 22 MR. BAILEY: We do do that work with regard to the I 23 preclosure. We refer to that as the preclosure period, ! L 4 24' .which as you say may stretch as long as 300 years at this 25' point.in time. l ANN RILEY & ASSOCIATES, LTD. ' -04<- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
217 1 We in fact are developing a preclosure safety
- G f
4 2 argument, if you will, which is based on the classical O l 3 approach of design basis events, determining the design 4 basis events, dealing with worker radiological safety and 5 then health and safety of the public with regard to the 6 design basis events and allocation of performance to 7 systems, structures, and components in the preclosure vith a 8 great deal of dependence on, as you point out, procedures, 9 quality assurance programs, those types of things to 10 maintain it, and that in fact is being developed as well. 11 The Repository Safety Strategy per se pertains to 12 the postclosure. 13 DR. WYMER: Okay. 14 DR. HORNBERGER: Charles? fm () 15 DR. FAIRHURST: I'll try to get near the 16 microphone. 17 One usually thinks of sort of strategy with regard 18 to three main barriers -- the engineered barrier, the 19 geological barrier, and then, if you like, the biosphere 20 model. I 21 Here I am maybe betraying my own ignorance -- and i I 22 you have concentrated pretty much on the first two. Is 23 there any room for you to analyze alternative biosphere 24 strategies, biosphere models or implications, you know, once 25 this is released? ! l l ANN RILEY & ASSOCIATES, LTD. ' \_- []- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D..C. 20036 (202) 842-0034
. .. . .. - . . . ~ . ~ - . . . .
l 218 1 Do you take that as a given and thus -- i -s iI ' i u) 2 MR. BAILEY: No -- 3 DR. FAIRHURST: -- thus may be regulated by 4- somebody else?
'5 MR. BAILEY: They are in fact regulated by someone 6 else, the Environmental Protection Agency is charged with 7 proposing a standard that we have to meet, and there are L 8 'obviously several ways to skin the cat, if you will, inside 9 the biosphere.
10' We have for our modelling chosen a certain 11 methodology. We don't know that that methodology is in fact 12 the methodology that will be dictated or specified by the 13 regulator, but we expect that the specifics of the biosphere 14 modelling will be pretty closely held by the regulator, f'h N. ,) 15 either the EPA or the NRC. 16 DR. FAIRHURST: So you just go as far as saying 17 we'll see minimizing what is released and then what happens 18 to it will be less than if it's more. 19 MR. BAILEY: We look at -- you are correct. We do 20 our very best to minimize what is released and-then we look 21 at how it gets handled inside the biosphere. 22 If we feel that we have, if you will, an unfair 23 biosphere model, for lack of a bettar word, one which is not 24 representative, not appropriate, not up-to-date, then 25 clearly we would argue that, or DOE would argue that. l ANN RILEY & ASSOCIATES, LTD. [ O-l
\~ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
. . - ~ . _ - . . . _ . -. . ~ . --- .._ __-- - . . . - . . . -
l l' l 219 i L 1 DR. EAIRHURST: There are several different models 1 ( 2 that were suggested by the TYMS committee -- 3 MR_ BAILEY: Yes, yes. t 4 DR. HORNBERGER: Marty? 5 DR. STEINDLER: The general notion of shifting 6 your system design in an iterative fashion based on the t
'7 results of your performance assessment and prioritization of 1
8 'importance factors-sounds like a reasonably decent logical 4 l 9 system, but I wonder if the iteration, considering the 10 uncertainties in models, the quality of the models, the lack
- 11 of parametric precision, et cetera, allows you to in fact 12 evaluate gross design changes, particularly in the EBS.
13 In other words,.is your tool sensitive enough to 14 give you any kind of ability to truly make optimization in 15 design, which is what I think you are doing? 16 MR. BAILEY: We are trying to make -- I understand 17 your question. You are correct. If you don't have good 18 tools, then you are not going to have good answers. 19 Our tools are fairly good and I think the secret
.20 to the use of tools is to recognize how good they are-and to 21 place the dependence on the tools based on how good they 22- are, and we are looking at our results, I believe, with that 23 jaundiced eye as to whether or not we really know how well
- 24 we know this, and to make a jump to a fundamentally 25 different design with a potentially flawed tool, I think 4
I j .- g- [ ,,] ANN RILEY & ASSOCIATES, LTD. , Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l
220 1 that would be a hard thing to do per se. ( ) 2 I think what you do if you find something l 3 fundamental -- and the systems reveal something to us every 4 day about what is important -- and what the parameters are 5 or the processes are that are really important, and with 6 that, if it were to reveal to us that there appears to be ! 7 something happening here that is worth studying and if it 8 needed to be, we would go back and look at it before we made 9 that decision, so I think we have to recognize that our 10 tools may in fact not be perfect. 11 DR. HORNBERGER: Andy, do you have anything? No? 12 Go ahead. 13 CHAIRMAN GARRICK: One of the things that I wanted 14 to ask Jack is what was the real basis of the 19 factors? (O (m ,/ 15 How did you come up with the 19 factors? 16 MR. BAILEY: The 19 factors were arrived at 17 through a very tough session that lasted, oh, about two 18 weeks on an iterative basis, back and forth, and what we did 19 in that is we did two things. 20 The first was very simple. We walked through the 21 mountain and said where does the drop of water go? And when 22 we walked through the mountain and looked at where the water 23 went we picked what is the process that is happening at each 24 point, and then we asked Performance Assessment, and we did 25 this very early on, and we asked Performance Assessment how [\- ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
. _ _ . . _ . ..- _. .. _.______.__._._.._..._.__.___..__.m _._ m.
-221 1- important is this?- Can you break this down? Can you
!'(fs) 2 understand this particular segment of the performance t ! 3 factor, because the models aren't arranged directly to b 4 . correlate to the performance factors, and you will see that, 5 I beliuve,'in Holly's talk. l l 6 So we wanted to make sure that we really could get 7' an understanding of each one of the principal factors 8 'through the tools-that:we had available but we didn't want [ 9 to get ourselves caught up in trying to find our way through 10 'the models -- let me' finish. 11 Then what we did, as the design moved forward, we
- 12. actually ended up changing those principal factors as we
~13 developed the newer performance assessment and went back and 14 said what is really important to us and what isn't important
' O(s ,/ 15 _to us -- and neptunium solubility was not a part of the 16 original principal' factors from a year ago basically because 17 we didn't see neptunium solubility as being much different 18 than the dissolution of the fuel, but when we actually broke 19- down the performance factors for their importance over 20 timeframes, we found that neptunium itself was something we 21 had to consider, and so it was an iterative process trying
'22 to get.a certain transparency in how the system worked, as 23 well'as understanding what was really important to the
-24 overall performance of the system.
25- CHAIRMAN GARRICK: Have you made any attempt to cI \ ANN RILEY & ASSOCIATES, LTD. ! \-- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034'
... . . . _ . - - . . ._.__..._.._.__._.____._....._.____.._.__..._._.m___m____.
222 1' . map these 19 factors into the NRC's key technical issues. () !2~ MR. BAILEY:- Yes, we have. We have looked at the 3 -- we have done just exactl'y that, and what the key
'4. technical issue and how they relate. And we, in fact, have 5 'some~of those maps available. I. don't have them with me, 1
6- but we have those maps.
'7 -CHAIRMAN'GARRICK: Have you found'any holes between the two?
9 MR. BAILEY: We find some differences, if you. 10 will, in importance . Things which our analyses are u 11l revealing may be more less important than the NRC's analyses 1 12 or evaluation of older analyses reveal'to be more or less 13 important, which is the purpose of ongoing discussions. 14 CHAIRMAN GARRICK: Have the Issue Resolution and 15' Status Reports been a source of material for evolving to the 16 19 factors and the method of dealing with them? 17 MR. BAILEY: .I believe that the Issue Resolution L 18 Status Reports are a fine piece of work by the NRC. They 19 provide us with a very good communication tool of the NRC's l-20- understanding and what is important, and how to go about
- 21. resolving issues and reducing uncertainties. And they point 22 out what is-important with regard to the NRC and a 23 regulator.
24- -Accordingly, we pay a great deal of attention to 25 them since we have to apply to the regulator for that L ANN RILEY & ASSOCIATES, LTD. 4 F -(s/ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 L Washington, D.C. 20036 l (202) 842-0034 Y
223 1 license to construct and operate and, ultimately, emplace .r~N () 2 waste is the site is found suitable. So we use those in a l 3 great deal. 4 I wouldn't say that we necessarily scour the IRSRs I 5 in order to find the principal factors, however. 6 CHAIRMAN GARRICK: But, again, there is the 7 ability to relate the two? 8 MR. BAILEY: Oh, yes, it is clearly the ability to 9 relate the two and look at what the importance is back and 10 forth between them. But that was not the driving basis for 11 me, the factors. 12 CHAIRMAN GARRICK: And this is a question we will 13 also ask the NRC staff. One other thing that I just want to 14 comment on, and I think that this is something that will () ( ,/ 15 come in the course of the other presentations, but there is 16 a tendency for us sometimes to think over-simplistically 17 about this repository. We see a sectional of a canister and 18 a tunnel, and inverts, and drip shields, and backfill, and 19 different -- corrosion resistant materials. 20 MR. BAILEY: I know that drawing, sir. 21 CHAIRMAN GARRICK: Et cetera, et cetera, et 22 cetera. Your famous drawing. And I think sometimes we 23 forget about the extensive space that the repository 24 occupies. I have forgotten the number, but, as I recall, it 25 is something like 800 acres, when it is -- [\s ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
_.- . _ . - . . . _ _ . . _ . - _ _ _ . _ . _ . _ _ . - . . . . _ . _ _ _ . _ _ ._ _,_...m _ _ 224
1 MR, BAILEY: Seven-hundred-and-forty, I believe is 2 .tne number, sir.
-3 CHAIRMAN'GARRICK: Yeah, when it is full. So, you l
4 are-into-a space. situation here where not all conditions are 5 the same. 6 MR. BAILEY: Yes. 7 CHAIRMAN GARRICK: And so you have got different 8- fluxes, different geological conditions. Are we going.to 9- hear about what I might call.a degradation model as a-l 10 function of the geometry of the actual repository, and how 11' this enters into the calculations of initial failures? L 12 MR-. BAILEY: We -- I am getting lots of nods from 13 the back of the room, yes. We, in fact, consider the 14 spatial size, and you will hear several pieces about how the 15 repository is divided'up into different regimes with regard i i 16 to the natural processes and what occurs, and that, in fact, i 17- .the results in different regions are, in fact, different
]
18 based upon the natural characteristics of each. And, in ; 19 fact, the unsaturated zone is -- it is basically in a very 20 large section -- it is considered in the same manner. R21 CHAIRMAN GARRICK: Yeah, and I think of the things 22 ~ we look for is whether or not the consideration of this 23 problem is mechanistic as opposed to just a statistical 24 .model. Obviously, to be as realistic as we would like to 25 be, the preference is for some sort of mechanistic basis for I' I
~
ANN RILEY & ASSOCIATES, LTD. p* Court Reporters L 1025 Connecticut Avenue, NW, Suite 1014 ) Washington, D.C. 20036 j (202) 842-0034 l l
225 1 the degradation rate of the repository. l ,f% () 2 DR. HORNBERGER: Ray? 3 DR. WYMER: Let me see if I can rile you a little 4 bit. What is your view or position on some of the claims 5 that have been made that the water table has, in fact, 6 gotten up as high as the repository will be? 7 MR. BAILEY: That is not my field of expertise, 8 and I don't think it is appropriate for me to try and make a ! 9 technical comment without regard -- 10 DR. FAIRHURST: Are you getting any nods from the 11 back of the room? 12 MR. BAILEY: No. Actually, no. Actually, they 13 are looking on eagerly. 14 [ Laughter.]
/"N k s) s 15 MR. BAILEY: That is not area of technical 16 expertise.
17 DR. HORNBERGER: To rephrase the question, you 18 have mentioned that you, in terms of disruptive events, you 19 were about to close out, or you thought that you were near 20 closure on the two main issues, volcanology and seismology, 21 and the question is, in light of the phoenix rising from the 22 ashes again, about the possibility of hydrothermal activity, 23 are you doing any work or planning any work? l 24 MR. BAILEY: Yes we are, in fact, doing work, i 25 planning work, and going to do work. I believe Ernie Hardin T
/'N
( 4 ANN RILEY & ASSOCIATES, LTD.
\'
Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 i
226 l l 1 will talk about that, or can talk about it when he gets I ( 2 here!'but, yes,othat is an issue which we continue to 3 consider. l 4 DR. HORNBERGER: Charles'. l l 5 DR.'FAIRHURST: Yes. 'And this,.if it is something L 6 'that is going to be addressed later, by all means, just ! 7 defer'it. In your Rev. 2 in this, it is actually Section
'8 2.1.6 and.it is talking about humidity and temperature, the ^
l 9 waste. package. You say the analysis shows that the variable 10 temperature and relative humidity of the waste package did b 11 not significantly affect long-tert? waste package 12 performance. 13 I was at a workshop recently at which we were 14 looking at ventilated drifts and so on, and talking about 15 att'empts to keep the humidity below 60 percent because that ; l 16- had.some significance. And, secondly, there was an ! 17 engineered barriers workshop here. And I think, in fact, I
.18 have a hunch maybe you were there. Maybe you weren't.
19 MR. BAILEY: No, I actually didn't attend it. 20 DR. FAIRHURST: Professor Payer and others were L21 discussing the fact that the temperature of the canisters 22 did have a significant influence on'its lifetime. And so I 23 am just interested, you said it was a two weeks, no holds 24 barred session but that came out of it, and I am just trying 25 -to see how we reconcile that with what other people have
'r l ANN RILEY & ASSOCIATES, LTD.
f-f ('s i Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
227 1 been saying about C-22 and corrosion resistance, [n) V 2 MR. BAILEY: I am not familiar with what happened 3 in the last couple of weeks specifically. The way that -- 4 and maybe Holly wants to address it, the way that I would 5 think about it right now is that when we did this work, we 6 looked at varying the humidity over its likely range of 7 uncertainties. .And the variation over that likely range of 8 uncertainties didn't make a lot of difference to us. 9 DR. FAIRHURST: Okay. 10 MR. BAILEY: Now, going to a new design where you 11 maintain an entirely different range may, in fact, have an 12 impact that could be helpful. That's why you have to think 13 of it both in terms of design or in terms of uncertainties. 14 I see Holly wants to help me. (._,1 15 MS. DOCKERY: If you want help. 16 MR. BAILEY: Go ahead. 17 MS. DOCKERY: I'm Holly Dockery from the M&O PA. 18 I think that the humidity issue is mostly a corrosion 19 allowant material issue, and it didn't show up as much of an 20 issue at all for the alloy-22 degradation rate. And we 21 actually did -- I will show you a little bit about some 22 sensitivity studies. But it is true that C-22 degradation 23 rate is one of our top two uncertainty issues, and there are 24 studies that are ongoing to address some of the specific 25 issues, and some very detailed sensitivity analyses trying [#'~ I ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 (202) 842-0034 i l
. . _ ~ . . . . . _ . . _ _ . _ . . . _ . _ _ . _ . _ _ _ . . _ _ . _ . . _ _ . . _ . _ .
228 L 1 to look at specific parameters in which -- areas in which we
) 2- might sort of focus our work to try to narrow that 3 uncertainty range ac.d loc.. at how it'affects the TSPA.
4 DR. FAIRHURST: .The temperature controls leads 5 into repository design for thermal loading density and a lot 6 of other major _ considerations. 7 MS. DOCKERY: As Jack said, I am not familiar with 8 what has gone intensively in the last two weeks with the 9 'TSPA peer review team. I think they are going to be giving 10 us a report within the next month or so, and so that 111 specific issue -- but I am aware of several of the others 12 like the specific chemistry at the bottom of pits, and 13 things like that, that have come up as a result and overall 14 dealing with the alloy-22 degradation rate, and the 15' variability of that degradation rate is a big aspect of what 16 the program is looking at. 17 DR. HORNBERGER: What I suggest, since we are I 18 getting into details, is that unless we have one question on 19 the more general area, we will move into the other 20 presentations. Andy.
'21 DR. CAMPBELL: Okay. This is kind of a generic 22 question. We often say in PA one of the key areas of L 23' uncertainty is, in fact, conceptual model uncertainty, as 24 opposed to uncertainty in particular parameters or
- j. 25' -particular process level models. How does the safety 4
( . ANN RILEY & ASSOCIATES, LTD.
\~ Court Reporters
-1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l
l
229 1 strategy, and maybe later we will hear about in TSPA, how do
; 2 you deal with alternative conceptual models? And what got l
3 me thinking about this is this whole issue of hydrothermal 4 activity rising up into the mountain. And is that 5 incorporated as a fundamental component of the model or is 6 that kind of a side analysis, an ancillary analysis that is 7 done independently? 8 MR. BAILEY: I am probably not the best person to 9 answer that particular question. I think that you have to 10 look at it in a couple of ways. One, with regard to the 11 rising of the water as you decide if it is, in fact, an 12 event or a process that you have'to consider. And if it 13 doesn't make the cut-off of being considered, then it 14 doesn't make the cut-off of being considered.
,7
, ( ,) 15 An alternative analysis can be run to show, from a 16 defense-in-depth point of view, if you want an alternative 17 model consideration, that, in fact, it would be very 18 significant, or it would not be very significant. That is 19 one to make the issue go away, for example, if, in fact, you 20 run it and it has no impact. 21 But our approach at this point in time is, in 22 fact, to determine the features, events and processes that 23 are most important and to work with those, and that the 24 alternative models, I would guess, would be handled on a 25 separate basis of here is something that has to be i (\- ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 l (202) 842-0034 l
230 1 considered as, if you will, an alternative -- an alternative t ( }s 2 approach with an alternative answer, and why it was or 3 wasn't considered, or if it turns out to be important, 4 consider it with regard to defense-in-depth. 5 DR. HORNBERGER: Okay. Thanks very much, Jack. 6 We are next going to have an overview of TSPA by Abe, Abe 7 Van Luik. 8 MR. VAN LUIK: Now that Jack got all the tough 9 questions. 10 DR. HORNBERGER: No, we saved some tough questions 11- for you. You're our favorite tough-question guy, 12 MR. VAN LUIK: The idea is to use this microphone? 13 Is that correct? 14 Okay, this is my topic. This is you; this is me.
- f~'T
(_) 15 Now we really did look at what we presented last 16 time about the TSPA-VA and tried not to duplicate much of 17 that. 18 DR. HORNBERGER: Good. 19 MR. VAN LUIK: So I hope that's okay, because if 20 yo'1 ask a lot of questions that show that I should have 21 brought my previous viewgraphs, I'll feel bad. 22 What I wanted to do is show the correlation of the 23 attributes of the first -- oh, this is the old viewgraph; 24 the new viewgraph doesn't say this -- basically to correlate 25 the safety case with the TSPA model components and the KTIs. .[ ) ANN RILEY & ASSOCIATES, LTD.
\J Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l
i l 231
'l I'll do that in the'next viewgraph. I wanted to briefly
() 2 update you on the conceptual models used and discuss the 3 results in some detail, and then the way that we have L 4 structured.these talks is that when we get to the hard part, l l 5 Polly takes over. Okay? Which is only fair. 6 DR. HORNBERGER: This is helpful. l 7 MR. VAN LUIK: And I think -- oh, my gosh, this is 8 completely unreadable. ( 9 Well, I guess the screen serves as a pointer to 10 . what's important in the package. But Jack has gone through i 11 the attributes of the safety strategy and the principal 12 factors, and'then the. thing is that principal factors are a l 13 nice way to scientifically and logically break down the i
'14 repository. But how do we handle this in our TSPA modeling?
.15 Well, the unsaturated zone flow model takes precipitation L
i 16 infiltration and -- 17' DR. HORNBERGER: Percolation to depth. 18 MR. VAN LUIK: Percolation to depth. We have a 19 separate seepage model which brings the seepage into the 20 drifts and looks at the effects of heat in the excavation on 21 the flow. Thermal hydrology is done at both mountain scales 22 and drift scales. And there we look at dripping onto waste [ '23 packages, humidity and temperatures at waste packages, et L .24 cetera. In other words, a lot of these principal factors l' l *25 are lumped into some of the main model components of TSPA. l
- (W)
- ANN RILEY & ASSOCIATES, LTD.
Court Reporters 4 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 L u . . . ._. - - . - - - - . . - - .
.. _ _ - . _ . _ _ . _ _ _ _ _ _ . ._~ ._. ._ _. . . . . . _ ._ . . .
232 1 Now if,we move this a little bit, the question () '2 3 awhile ago was the correlation to the key technical issues of the NRC, and you can see that different models address 4 different of the key technical issues. And then we have an 5 umbrella issue, which is the TSPA issue, which basically 6 picks up anything that was left undone in the others. 7- The next viewgraph is. kind of a discussion that 8 went along with.the' previous-viewgraph and Just makes a very 9 basic point. 10 Thank you. And I'll try not to do that again. 11 Our TSPA modeling components were developed to 12 represent various process models produced by the site 13 characterization and design organization. Both of them
- 14 produce process-level models. Principal factors represent I:
D), 15 aspects or-parameters used to model the processes, and then 16 in ti?e charc that I showed previously disruptive events were 17 not discussed. They are included in the TSPA and they're 18 correlated to a different element of the postclosure safety 19 case. 20 This is one of the few viewgraphs that you've seen 21 before, and basically it's just a reminder of all of the l 22 different process models that we've discussed. Now the 23 question that was asked a while ago is why do we feel 24 confident that our modeling is good enough to allow us to 25 make judgments on the engineered system. l [ ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 (202) 842-0034 i
233 1 As you probably are nell aware, where we need more (~ ) 2 work is in the near field environments and its effects on
%J 3 the waste package. For example, the expert elicitation on 4 the degradation rate of C-22 -- Alloy 22, more correctly --
5 the reason that they gave is a wide range of possible 6 degradation rates is because we had large uncertainty in the 7 evolution of the near field environment. So an obvious 8 thing that we still need to do is work on that near field 9 environment and try to narrow the uncertainties in it so ~ 10 that we can narrow the uncertainty on the Alloy 22 corrosion 11 rate. 12 These are the processes that I showed on my chart 13 just awhile ago, and across here are the disruptive events. 14 Like I said, in previous briefings we have walked you (h) ( , 15 through the structure of TSPA, which this is. And then the 16 biosphere model on the very end right here, all of the 17 results that I will be showing in a few moments are for a 18 dose receptor at 20 kilometers receiving a dose calculated 19 from all pathways, not just drinking water, and this is what 20 we believe will be close to what's going to be required by 21 the NRC, but of course that remains to be seen, depending 22 somewhat on what the EPA does. 23 How did we show -- when you get your copy of the 24 VA, which is going to roll off the presses and be rolled out 25 on Friday -- I was hoping it was today, but, you know,
.O ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
)
234 1 somehow -- it'll be out on Friday. In the VA you~will see lY g 2 that several different calculations are shown. We did 3 deterministic base-case analyses. We did probabilistic base 4 care analyses. And we did comparative analyses, which 5 Charlie will be talking about quite a bit. 6 Why did we do all that? We did deterministic base 7 case analyses, which we call the expected-value case, to 8 allow us to do one off studies. This deterministic case was 9 a single realization. All the uncertain input parameters 10 were sampled at the mean of their range. And the usefulness 11 of it is to illustrate the relative influence of various 12 components or subcomponents on individual dose values. A 13 deterministic base case would not be used to assess system 14 performance against a regulatory requirement. () 15 Great caveat. 16 So what are the results from the deterministic 17 analysis? And here you get the scoop on what's coming out 18 on Friday. Although if you have gone to our web site, you 19 will realize that for weeks now all of these results have 20 already been on the web in PDF files, and if you want to 21 suffer a lot, you can download those PDF files. 22 [ Laughter.] 23 'DR. CAMPBELL: We've tried. 24 MR. VAN LUIK: Oh, have you? 25' No, we put them in sma]' enough packets. O ANN RILEY & ASSOCIATES, LTD. \- / Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
i 235
-1 DR. CAMPBELL: Right.
2- MR. VAN LUIK: Okay? ( Okay, but it's still a pain. 3 In a few months we will also have HTML files to 4 download, and then you can manipulate things a little bit 5 easier. But there's a certain beauty to PDF. Anybody can
~
6 read it or use it, so the reader is free. 7 If we look at bottom lines from the deterministic 8 analyses, and as it says down here, this little caveat is on 1 9 all of our viewgraphs. We don't pretend that these are 10 ready for licensing. Ten-thousand-year dose rates are quite 11 low; 100,000-year dose rates are still quite low; 12 million-year dose rates we do have some peaks and valleys in
'13 there.
L 14 For the deterministic base case we assigned the !q ( ). 15 occurrence of climate changes, and this explains some of the i L 16. two peaks here. They are 400,000 years apart because these 17 are super pluvials. In the work that's being done by the GS i 18 to help us with the SR and LA, the super pluvials are l 19 disappearing because.they believe now that the long-term 20 average, which means basically a glacial climate up north, 21 that the super pluvial is not indicated by the ice core L l- ) 22 record, the sea bottom sediment record, et cetera, that i 23 there's just more variation that we should take into account [ 24 in the long-term average. So these peaks will disappear, 25 but you will see that the climate changes are going to have hv ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticuc Avenue, NW, Suite 1014 ! Washington, D.C. 20036 (202) 842-0034 4 d y-- -
,e,.
4 236
.1 . a.little bit'more. variability to them.
I
}
2 But I think,'you know, these are of passing I 3 interest. What we need.to do is get down to why we actually l l 4 did a deterministic case. We'll revisit all these results 5 again in-the probabilistic case. l 6 But why we did the deterministic case is to look 7' at the various components and look at their importance. For 8 example, we assigned the first climate change.here, we 9 randomly picked it between 1,000 and 10,000 years for the l
- 10 probabilistic case, which means the~mean value is about 11- 5,000 years, so we impose the 5,000-year climate change from l 12 the current dry climates to the long-term average or a 13 pluvial climate, which doubled the precipitation and l 14 multiplies the infiltration significantly more than double, I) 15 because it's also cooler, and there's more -- or less j 16 evaporation, et cetera.
i 17 When we look at the mean values of the things that l ] 18 determine waste-package corrosion, we see that we have 19 probably about 14 to 17 waste packages failing over 10,000 20 years cut of 10,000 packages, but we assigned one to fail at 21 1,000 years. We just assigned that to look at the effects 22 of basically a flawed waste package being in place. The 23 reason we put it at 1,000 years is because at about that 24 time we see that the thermal period would allow water to l 25 come back in. The thermal period's pretty well over. In 4 $#) ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 l (202) 842-0034
l 237 1 order to explain these others to you, I think I'll have to 2 [w )/ go to my paper copy, because I can't read what's on the 3 screen too well. 4 DR. CAMPBELL: Abe, that single case of juvenile 5 failure is a representation, not an estimate of actual 6 juvenile failures, is it? 1 l 7 MR. VAN LUIK: It's a representation in this 8 analysis. We have done some work, the engineering 9 organization has done some work on the likelihood, and it 10 comes out that perhaps one to ten packages could be failed, 11 but, you know, this is a31 speculative at this point. 12 DR. CAMPBELL: Out of how many total? 13 MR. VAN LUIK: Out of 10,000. And so we just put 14 one in. Actually I think the range was from 0 to 10, with rr ( ,) 15 the likelihood being very close to 1 for the expected value. 16 DR. CAMPBELL: Okay. 17 MR. VAN LUIK: Okay. And I also have a problem 18 reading this, okay? 19 If we look at the seepage flux history on this 20 side right here, then -- 21 MR. SCOTT: We have 11 by 17s. 22 MR. VAN LUIK: Oh, wonderful, wonderful. Some of 23 us are getting older. And so they make allowances for that. 24 [ Laughter.] l 25 Okay. This chart right here is the area available [N; ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 ! Washington, D.C. 20036 (202) 842-0034
l l 238 l l 1 for advection and diffusion releases. In other words, the
/
( x ,1
) 2 red line right here is the juvenile failure patch area. In 1 3 the juvenile failure we assigned it to fail a whole patch, l 4 which is a sizable area. If we look at the corrosion, i
5 however,'and ignore this juvenile failure, the area l 6 available for release on the waste package starts out at 7 this value at about -- well, it's close to 3,000 years, . l 8 2,500 years or so -- and increases in this manner. And then l 9 similarly the pitting corrosion starts at about the same 10 time and comes up in this area. And I think this i 11 dramatizes -- how can that be, we have two copies? 12 DR. CAMPBELL: Yes. 13 MR. VAN LUIK: Oh, okay. Yes, I can use all the 14 help I can get. l () rs 15 But the point is that the pitting failure is not 16 going to be a great contributor to release. It's the 17 larger-scale failure, the patch failure, that's going to 18 contribute to your release. If we now go to the seepage 19 flux history -- this is more difficult than it deserves to 20 be. 21 Here we're looking at seepage into the drift. 22 Here we're looking at the reduced seepage, a couple orders 23 of magnitude Urat gets into the juvenile failed waste 24 package. And then here we're looking at still more 25 reduction in the volume of water coming into the corrosion l
; ANN RILEY & ASSOCIATES, LTD. \-/ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036
, (202) 842-0034 l
239 l 1 failed waste packages. l (n) 2 If we look quickly over here at technetium ! 3 release, you can see the juvenile failure right here with 4 -some of its readily available technetium coming out at a 5 .high level, going back down, and then diffusion and l 6 advection picking up, and then the climate change comes.in 7 here, and we get a huge jump in the -- well, an
- 8. order-of-magnitude jump -- in the release from the advective 9 case because there's more advective flow. And then the 10 diffusive release actually comes down a little bit, and then 11 continues to increase. .But the point here is the climate l 12 change raakes a break between the advective and diffusive 13 release.
14' Now if we had something like the Europeans are p). (, 15 doing, a bentonite barrier, you would see a bigger 16 difference between the invective and diffusive. But we 17 don't. 18 If we look -- and this answers a question from 19 before, only partiall f, I realize -- but if we look at the 20 six areas that the repository was broken into, it was broken 21 into those areas because of considerations such as largely 22 the infiltration map that was given us by Flint, et al. We 23 broke it into areas where the infiltration and percolation 12 4 .were comparatively similar, and I showed this map the last 25- time we talked,-broke the repository into six areas, and you ANN RILEY & ASSOCIATES, LTD.
-( )-
Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
240 1 can see that the illustrative case that'I show is always for [} 2 the NW, the northwest area, because it has the highest v 3 release. The other areas because there is less infiltration 4 mostly have lower releases. And so we do do sensitivity 1 5 studies, and I'll show a few more. i 6 The reason that we pick on technetium is because i l 7 no sorption, high solubility, means fast EBS release and 8 fast transport. So -- and that's for the 10,000-year case. 9 MR. VAN LUIK: Have you got two copies of that, i l 10 too? l 11 DR. CAMPBELL: Yes. 12 MR. VAN LUIK: Okay. Here we see the mapping of 13 the repository area into its six areas, and then we have a 14 similar mapping of six areas for the saturated zone, which
/~.
(v ) 15 is also based on the characteristics of the saturated zone, 16 and then we have a translation of radionuclide travel using 17 the UZ transport model of radionuclides from these areas to 18 these areas. But you see that there is no correlation, 19 really, in the mapping of these areas and these areat. It 20 is not, you know, just taking this same area and sticking it 21 down here. This is based on the characteristics of the 22 saturated zone as well as we know it. 23 Again, we look at technetium. And if go now -- 24 can we get both of these on the same page? Yeah, that's j 25 pretty close. If we look now at the six regions, and here 1 [9 f ANN RILEY & ASSOCIATES, LTD. ~ \_) Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l Washing _on, D.C. 20036 (202) 842-0034 I
-. - . -- ~ , .- -.-.. = . ~ . - . . - - - . - - - . . _ _ .- -.- -
241 1 they are-region 1 through 6, and here they are given the () 2 3 names right here. zone regions. The region-11through 6 is the. saturated-These are the unsaturated zone regions which 4 are given locator names and these.are just numbered. 5 But you can see that there is variation in the 6 receipt of-radionuclides for each of these six areas, and 7 some of them are extremely low and some of them definitely 8 lead the pack. But if you look at the flow lines from the
~9 UZ to the SZ, you will see that region 5 in the SZ, right 10- here, most of the flow lines kind of go in this direction, 11 and so region 5 receives most of.the flux -- I mean -- yeah, 12 most of the flux and most of the radionuclides. And then it 13 is, you know, the total is less than an order of magnitude 14 Lhigher than region 5. So the others contribute very little.
() 15 If we go-to the right, and this is a question that 16: may-be of some importance to some of us, if we look at the 17 dry climate that we are in right now, we can see that the 18 breakthrough from the UZ to the SZ is right around, oh, a 19 few thousand years for the mean value, right in here. Then R2 0 - if we go to the long-term average climate, we are talking 21 about several hundred years for the median breakthrough, i 22 And if go to the super-pluvial and, basically what is going l l23 to happen in the calculation that comes after this, is that l 24 the characteristics of these two -- I think, you know, this 25 will be movcd to the left a little bit, and this will
/~'- ANN RILEY & ASSOCIATES, LTD.
f k -} u Court Reporters ! 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 ! c l e
l 242 1 1 disappear. But the super-pluvial, we are talking about a i (' ) 2 hundred years basically for the 50 percent breakthrough from 1 i 3 the UZ to the SZ. 4 So even though we have rapid -- you know, these ; 1 5 tails come right down to zero almost. Even though we have l 6 some rapid flow paths in the mountain, still the bulk of the 1 7 activity will arrive, you know, at approximately these times 6 for the calculations that you are going to see. 9 Okay. Moving on now to the last of these big 10 pictures, which will be a relief to you as well as me. Now, , 11 we are going to illustrate some of the results going.from 12 the contact between the UZ and the SZ and the 20 kilometers 13 at which we calculate a dose. I guess we can start right l 14 here since that is convenient. fg t
! ) 15 Saturated zone breakthrough curves -- and . tis is 16 a tricky one right here, this is not based on modeling that 17 is shown in these others. But if we haC a constant source 18 at the rTZ , at what point in time would you start seeing 19 arrivals at the 20 kilometer boundary? And you can see 20 that, basically, in a few hundred years you start seeing the 21 first arrivals. About half of it comes through at about a 22 thousand years, this is just for the saturated zone, and in 23 less than 2,000 years, you are a constant state which 24 matches your input value. So --
25 DR. HORNBERGER: This is for a constant source at i ANN RILEY & ASSOCIATES, LTD. f>) x- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 j Washington, D.C. 20036 (202) 842-0034 1 l
l l 243I t 1 the repository horizon? ! /~N 2 MR. VAN LUIK: Constant source at the repository 3 at time zeta. And so this is not to be compared with these 4 other flux values which are based on the actual TSPA. But 5 we just wanted to see what the effect of the saturated zone 6 was. And then that is for technetium. Of course, other i ! 7 things that are somewhat retarded, or greatly retarded, 8 would have very different arrival times, with some of the 9 plutonium isotopes coming in at 100,000 years and reaching a 10 steady state that never is equal to their input value. 11 DR. HORNBERGER: Abe, just to make sure I 12 understand this, though, this is for a constant source at 13 the repository horizon, so it incudes UZ? 14 MR. VAN LUIK: No, no. It is a constant source --
,7
() 15 I'm sorry -- at the UZ-SZ interface. 16 DR. HORNBERGER: Oh, okay. 17 MR. VAN LUIK: Yeah. It is strictly looking at 18 the sensitivity of the parameters that govern the saturated 19 zone. 20 DR. HORNBERGER: Okay. 21 MR. VAN LUIK: And I must also add a caveat that 22 this is for the model used in the VA. And you know that we 23 are developing a new model for the saturated zone that 24 should be out this year. So this will probably be somewhat l 25 different the next time around. [~j
\-
ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 (202) 842-0034
u 244 H 4 l 11 Another -- if we can go right to the dilution ! [
\ 2 factor here. Another' contentious point was the dilution s-) -
~3 . factor for the saturated zone.
4 CHAIRMAN GARRICK: Excuse me a minute, Abe. Is i !
~
5 there transport' data.to support that at all from the l 6 saturated zone? f-7 MR. VAN LUIK: The modeling that is being done at 8 this point is taking into account what we have learned from 9 the Seawell complex, and the modeling here is based, 10 basically, on the expert elicitation for the saturated zone. l !. 11 So the answer is,.yes, there is some data, but it was not 12 used specifically in this model, except to inform the people 13 that gave us the ranges'of things to be. expected for the I 14' saturated zone,
/~%
l 15 CHAIRMAN GARRICK: Thank you. 16 MR. VAN LUIK: So, as'I think you realize, because 17 we explained this to you before, that the saturated zone 18 model that we went into the VA with was dropped and we came 19 up with this stream tube model as an intermediate fix until 20 we can get a new model this coming year. So for the SR and 21 LA, you will see a much more sophisticated, much more based 22 on ground information model, but for the VA, we thought that ! 23 this was a very conservative approach. You know, in fact, 24 we are being criticized for being so conservative, which I l 25. like. , = [~') ANN RILEY & ASSOCIATES, LTD. i 3d Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
..n... m .a l
'245
- -1 CHAIRMAN GARRICKi -Is this sort of what you l 2 expected or is this. a surprise?
1: ~3 MR. V101 LUIK: Well, to tell you the honest truth, ( 4: I expect when the modeling is done that some of these curves 5 will move to the right'a.little bit, but not all that much. I l-6 I really don't think that we are that far off, even though 1 I l 7 we -- I hope we are conservative, but I think that when they 8 do the ground truth modeling that -- and, also, we have some 9- work-going with Nye. County. 10 They are going to drill several holes in the 11 saturated zone, all the way into the alluvium, you know, 12 from the repository. Where we stop having data, they are i 13 going to drill systematically for_their early warning f 14 program, and we are going to analyze those cores for the j () l'5 p material that is in them,-and we hope to get some properties L _16 for'the alluvium. Right now we have no properties for the I l 17 ~ alluvium except assumed properties. 18 So we are looking at, by the time of license 19 application, having a lot more information for the saturated l 20 zone model. But in the next year, for the SR, it will be
- j. 21' based, basically, on what we know now. So I see continual 22 improvement in this model as we go along, but this model 23 here was based on conservative assumptions.
i 24 DR. HORNBERGER: Conservative assumptions being l 25' groundwater average linear velocity of 20 meters per year. a [ l' ANN RILEY & ASSOCIATES, LTD. A /' Court Reporters l 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 L li .. -. - - - - . . - .-. .
! -j l 246 1 I=mean that is-what we infer.from that graph. r-MR. VAN LUIK: 4 2 (.J- Yes, exactly. And then the J 3 dilution factor, as you probably know, the experts came in a ( 4 which an estimate of about 1 to 20, right her, with an 5 expected value of about 2. This is looking at sensitivity 6 studies, looking at different processes in the saturated
- 7- zone and what they would do to you in terms of dilution l 8 factor. And you can see that there is a possibility of the 9 dilution factors going-much higher. But we picked this as 10 -- because it was based on the-expert elicitation, and, l 11 also, we feel that it is quite conservative.
12 I don't know if it is that important to go to the 13 right, but if we go to the left, if we look at the release 14 from the unsaturated zone, here, and then the concentration l () L 15 at the water table, this, again, illustrates Cle regions and
.16 their importance. It is region 5 that has the highest flux 17 coming in, and so it is also region 5 that has the highest 18 concentrations. But, of course, you know, going from flux 19 to concentration is a modeling effort that also requires a 20 lot of expert judgment, and so we hope to, you know --
t 21 again, it is expert judgment based on conservative 22 assumptions -- we hope to refine this part of the modeling 23 considerably for the next t.ime around. 24 T think that iF about it for the large pictures, l 25 and I really appreciate the help. i. ANN RILEY & ASSOCIATES, LTD. 1% . Court Reporters
- 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l
l l l l 247 1 Now, let's get to the more fun part, the f~) x/ 2 probabilistic case. Our probabilistic modeling uses linked 3 deterministic models with their related parameter ! 4 uncertainty, propagated using a Monte Carlo technique, and 5 multiple realizations, of course, are used to define the
.6 range and dose rates. So, when it comes to regulatory l 7 compliance, we believe that this is the type of calculation l 8 that will be required.
9 Now, what do you get out of running a hundred 10 realizations? You get a horse tail and I think no horse 11 would be proud of this tail -- with a hundred mile an hour 12 wind maybe. But what do you make of something like this? 13 This is the 100,000 year case. What we make of it, and I 14 will need my cheat sheet to read it, because -- or can that ,(m) 15 be focused? I don't know how to focus it. You think it can 16 be? So even I can read it. Oh, yeah. Yeah, yeah, yeah. 17 You do it good. Okay. 18 The thing that we have to extract from this horse 19 tail is information. And if we are looking at -- let me 20 cheat a little bit here. If we are looking at the different 21 time periods of interest, 10,000 years, 100,000 years and a 22 million years, we are looking at, basically, the mean values 23 and the absolute highest values. 24 But I think the important thing to illustrate here 25 is the huge difference between the me an and the mean. The f~} N_/ ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 r
l I l 248 1 median, which some of us would like to argue is very ( ) 2 representative of the range, is two orders of magnitude l [ %_/ \ 3 different from the mean. One order of magnitude different l l 4 here in this case, and also in that case, it is two orders 5 of magnitude. 6 The regulators, however, insist that because means 7 are more conservative and more readily understood, I must 8 say, by the general public, that mean values are to be used 9 in regulatory compliance analyses. When you do sensitivity 10 studies, such as we have done, on, you know, whether it 11 makes any difference to these curves to do a hundred 12 realizations or 300, or 400, or a thoucand, the means change 13 because you bring in, you know, more points on the tails, l 14 but the medians are very stable, and I think that is just t (q) 15 interesting background. 16 But, as you can see, the mean value for 10,000 17 years is about .1 millirem at 20 kilometers. The mean value 18 here is about 30, I believe. Yes. And then the mean value 19 for the peak dose over a million years is about 200 20 millirem. So that is the kind of information that we get 21 out of these runs. 22 Let me show you the curves for 10,000 years and a 23 million years. By a million years every one of our hundred 24 realizations contributed something, but some didn't even 25 start to contribute until about 800,000 years. Obviously,
/~N ANN RILEY & ASSOCIATES, LTD.
N.- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l
249
~
1 -that is an outlier just like the top.one is an outlier. For )
'f~Q 2- l i,
m/
). 10,000. years, many realizations did not contribute at all. I l
3 I think there is like 72 realizations plotted here. And, I 4~ you-know, it is a challenge to come up with meaningful ' 5 interpretations.of these types of things, but we meet that i 6 challenge. 7 CHAIRMAN GARRICK: Just to think out loud a little 8 .with you, Abe, here. 9 MR. VAN LUIK: Yes. l 10 CHAIRMAN GARRICK: Isn't the issue here really the 11 contamination of the-groundwater or the. biosphere rather ) 12 than the dose, on the basis-that at least at the dose level
)l 13 you have the option of interdiction and it is more difficult i
14 to interdict at the biosphere and groundwater level? I mean ! (,,r 15 I am not talking at it from the point of view of the NRC and i 1 16 what they are using. i. 17 MR. VAN LUIK: Okay. l 18 CHAIRMAN GARRICK: I am just saying that isn't the ; 1 l: 19 real -- if you back off and looK at this and ask, what are l l 20 we doing, because there is no way the public is going to get 21 these kind of doses unless we have gone to complete 22 primitive land, because of interdiction. So it is l
'23 meaningless -- it is meaningless from a dose standpoint. It i 24. is only a potential issue.
25 On the other hand, the contamination of the O ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025-Connecticut Avenue, NW, Suite 1014 L Washington, D.C. 20036 (202) 842-0034 t t
'250 .
1 L li ~ biosphere, the groundwater is not' potential. So, in fact, ! ,t 2. it strikes me that that.is where the real threatLis, it.is i l1 - 1 L 3_ not dose to people . I l 4 MR. VAN LUIK: In fact, when you look at the Draft 5' Environmental Impact. Statements, it does address those types j , 6 of issues. l i '
- 7 CHAIRMAN GARF CK: Yes.
[ 8 MR. VAN LUIK: The are looking also at contaminant i 1 I 9 levels other'than radionuclides because of exactly'that' type I 10 of concern. And the idea behind the.EIS is to show that the
. 11 contaminant levels are within expected ranges from a L 12 societal prescribed' point of view. But I believe you are
-13 right, that the real. issue is what are we doing to the
'14 ' environment with this type of system. On the other hand,
?( f; 15L the environment is interpreted, usually, as impacts on human 16 health and safety. And, so, from the NRC's point of view 17: and the EPA's point of view, that is the bottom line. The 18 EPA also wants to throw in groundwater protection, but -- l 1
19' you know, which sounds very noble, but I have some real i l 20 difficulties with how you would implement that. l I-21 But in the EIS, in NEPA space, we do address 1 22 exactly the types of concerns that you are raising. 23 CHAIRMAN GARRICK: Yes. l 24 MR. VAN LUIK: Which is, what is your actual 25 contaminant levels? And do they fall below the allowable 4 !;["' ANN RILEY & ASSOCIATES, LTD. I Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 e r.., . - - - ,. ,, ,. , . . - _ . _ _ _ . . _ . . _ _ . , _. _ _ _ _ _ _ _ . _ _ _
. _ . . . _ . . _ _ . _ . _ . . . . _ _ _ . . . _ _ m _ . . . _ .__
251
. ~
.1 . contaminate. levels for each specific poison, so to speak?
- (A) 2 Like chromium is an issue, lead may be an issue, so we are 3 analyzing-all of those from that perspective in the EIS 4 . space. That is being done through performance assessments 5 like this, but they are being done, with our assistance, by=
6 the EIS. contractor, and I think they are doing a good, 7- honest job of it. 8 CHAIRMAN GARRICK: Yes. Okay. 9 MR. VAN LUIK: The reason that I threw this one up 10 here.is to, basically, give you a statistical interpretation 11 'of the horse pair diagrams. If we can go here first. What 12 we have plotted is for each of the time periods, like this 13 is 10,000 years, the 95th percentile of the -- in other 14 words, leaving off the top five and then coming down, that () 15 -dose rate, the mean value dose rate,.the median value dose 16' . rate, just a self-serving thing to throw in, and then.the . 01 7 fifth' percentile never even makes it onto this chart, or 18 this chart, for that matter. i 19 But I think this is very helpful in interpreting 20 what the horse tails actually mean. If you are looking at a
'21 .mean value compliance calculation, this is .ind of the line i
22 that you would be looking at, and these values, if you look 23 in;the VA for exactly this chart, you will find others that 24 have' rates that are slightly higher. What we have done for 25 these values.-- and I should highlight that -- is that we f 3 ANN RILEY & ASSOCIATES, LTD. 3 Court Reporters i
'1025. Connecticut-Avenue, NW, Suite 1014 Washington, D.C. 20036
'(202) 842-0034
7__ 252 1- ' averaged,-took the mean median and fifth and 95th
-( ) _2- percentiles at each time step;-these happened to be
- 3. hundred-year time steps; and this.is consistent with the
-4 . methodology in the draft CFR Part 63, which is available.on 5 the web.
6 Now for 10,000.and'100,000 years, the fifth 7 percentile dose rate is.zero; for.a million years, of 8 course, it does come into play, as I said, about 800,000 9 years down the road. 10 DR. WYMER: These are all technetium; is that 11 right? 12 'MR. VAN LUIK: No, this is actually all 13 radionuclide doses. 14 DR. WYMER: All of them? ,. ~ 15 MR. VAN LUIK: Yes. But technetium is the 16 dominant one in the early period. Neptunium comes in at 17 about this stage right here. And then we actually have some 18 contribution from plutonium in the colloid state down in the 19 hundreds of thousands of years. So, yeah, it is a total for l 20 all radionuclides shown in these charts. Oh, is that my l 21 next viewgraph? Okay. Thank you, Holly. 22 But I think this kind of analysis really helps, 23' 'and we appreciate the guidance that comes from the NRC in 24' this case, because before what we were doing is doing the 25 means and the fifth and 95th percentiles, going by the peak l 4 ANN RILEY & ASSOCIATES, LTD. (/ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 i
253 1 dose value regardless of time. Doing this in discrete time "( ) 2- steps, I think, actually gives you a more representative 3 ' understanding of how the system is working over time. 4 In fact, the other way, now that we see this way, 5 doesn't make much sense, l 6 Thank you, Holly, for pointing this out. 7 For the 10,000-year period, we have technetium 8 dominating, iodine.next, carbon-14 being a very small , 9 amount, others being negligible, and then zero doses in 27 l
-10 or 28 of the realizations. It's because this is less than 11 one -- you know, the round-up, sometimes it's 28, sometimes 12 :it's 27. l i
13' If we look now at the 100,000-year period, we 14 start to see some contribution for plutonium. Iodine is
- 15. still important, and technetium, but neptunium comes in as a i i 16 very important one.
17' And then if we go to a million years, plutonium 18 rises still; iodine is still there; technetium is there, a 19 very negligible amount; neptunium dominates and accounts for
- 20 70 percent of the peak dose.
21 So as Jack said, the vast majority of 22 radionuclides don't go anywhere. It is just these four that 23 contributes out of the large number that are there. 24 DR. CAMPBELL: Abe? i 25 MR. VAN LUIK: Yes. i b ANN RILEY & ASSOCIATES, LTD. ' \_) Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
254 1 DR. CAMPBEL'L': Quick question about that. In the 2- .NRC's analysis'of these previous analyses, there's a 3 somewhat different mix of radionuclides; forLexample, 4 . americium and even curium show up. Do you have a ready 5- explanation for that? Is that in terms of sorption l
'6 parameters or in terms of radionuclides considered?
7 MR. VAN LUIK: I think it's -- yeah, the 8-discussions I've. heard of that is that there is a difference 9 both.in sorption and solubility ranges over which they are ! 10 sampled and, you know, it's -- you know, they have a 11 different basis for coming up with.those numbers and, you 12 know, I don't think that the difference is that significant,
.13 actually, because our dominant ones are also their dominant 14' ones. They just'have contributions from some things that we . () . 15 don't see. So --
16 -DR. CAMPBELL: So the overall impact of whether 17 it's plutonium over long times frames versus americium isn't 18 that large? 19 MR. VAN LUIK: I don't believe so, yeah. l 20 Okay. Is it the zoom button you were using? 21 Yeah. If I do this long enough, I can get my engineering 22 degree and get a real job. 23 The disruptive events are the chird element of the 24 post-closure safety case. What we looked at in the TSPA 25 that is going to come out on Friday is igneous activity, I j ANN RILEY & ASSOCIATES, LTD. l (- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 L
- . . - . . - . . - . . . - . - . - . _ . - . ~ - ~ . - . . - - . . . - - _ . . - - --
l 255 1 looking at' direct release, enhanced source term, and f 1 q( f 2- . indirect effects; looking at seismic activity; inducing' rock l 3 fall; and also.looking at indirect seismic effects; looking 4- at nuclear criticality, in package and out of package; and human' intrusion.
'5 This is the stylized release'to the-6 saturated zone that's basically described in the National i
7 Academy of Sciences' recommendation. I 8 I'm not going to do justice to these, I'm just I i 9 going to basically show some. bottom lines'. 10 In the analyses that were done, only about 5 11 percent.of all the igneous events resulted in direct 12 releases. In other words, 95 percent did not. Doses were 13: assessed,.again, at 20 kilometers; that's where their 14' critical group'is located. About 10 percent of all l
) 15 realizations with releases had a wind direction from north 16 to south and resulted in a dose effect; 90 percent of the i
17 cases based.on the. wind rows for the location. The wind was 18' blowing from, I think, southwest to northeast, and there was I 19' no population to. receive a dose. So these are some of the i 1 20 caveats on these results that you will see. t 1
- 21 If you look at the base case release -- I mean 2
.2 dose rates, and then you look at the volcanic direct release
- 23 dose rates, of course, conditioned by the probability of l l 24 that event occurring, you can see that it's a rather l
25 negligible contributor to the total. i-
. () ANN RILEY & ASSOCIATES, LTD.
Court Reporters ; 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 l (202)- 842-0034'
.. .-. .~. - . . - - . _
...<_..m- .. .. . _ ._.. ..._._. . _ _ _. _ .. _ _ _ _ - . _ _ . - . _ _ . . . _ _
L 256 l' However, if you do a conditional plot and say 2 ~kay, o what are the effects"of an enhanced source term, given i 3 that we will have a volcanic event, then you see something i-This'is the' effects of a volcanic intrusion into 4 like-this. l 5' the' repository here and here; and:in this case it is L.
'6 increasing waste package degradation rates, altering the 7 chemistry of the waste form, and this means that
.8 radionuclides are more accessible to the groundwater and
,. 9 more easily dissolved. 10 And then the altered source term resulting from i-L 11- 'that is then put back into the base case groundwater flow, 4 1 l 12 and you can see that there is a two order of magnitude 1 j L13- increase here at this exact time, and about a one order of 14- magnitude increase here at this point in time. But, again, 15 these are conditional calculations. This is assuming that 16 -they happened,.and if you multiply by the about 10 to the 17 n .us'8 probability, you know,.these points are rather ! l l .18 - . vanishing, and I think that is what is shown on the next ! l 19 .viewgraph. l 20 DR. WYMER: Do those take into account the fact l l 21 .that it's going to be hot and it will dry the water out of I l 22 the system? l 23 MR. VAN LUIK: I don't think so. I don't think
-24 that was -- that we were that sophisticated. We just 25 basically -- Rollie did the calculations, and he wants to
[ \ ANN RILEY & ASSOCIATES, LTD.
\"s Court Reporters
. 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 l (202) 842-0034 i f i- -- , , . , - - . . ,. -
L 257
' 11 standLup and something.
2 MR. BARNARD: There's a. quick answer to that'. 'I'm 3 Rolson Barnard from Sandia Labs M&O. There's two main time l L, 4 periods involved. One is the time.at which the igneous 5- intrusion occurs,1and then a long time later, after it is-6 cooled and water is able to return, is when this enhanced
'7 source term situation.would arise, p 8 MR. VAN LUIK: Thank you, Rollie.
i 9-. This. puts those two peaks back into perspective 10 because you can see that when you take away the 11 conditioning, assuming that they occur, and put it in l '12 perspective of the likelihood of occurrence, that on a CCDF ! l 13 . basis therefis not much of a contribution. i , 14' When.we look at seismic activity, there's a caveat i A
- k_,) 15 on the analysis that waste packages need to fail to a l 16- certain extent before a rockfall can do.any damage to them.
i 17 sand so if we'look at this figure here, this is'the number of 18 waste packages that are failed, and you know, by rockfalls
- 19. 'over time.
l l 20 Now if you compare this with the failures -- I may 21 need your help on this, Rollie -- do these spikes here l 22 indicate where the base case is, or is the whole graph ,- 23 everything? I L 24 MR. BARNARD: No , that's only rockfalls. I 25 MR. VAN LUIK: Only rockfall. Okay. P ' ANN RILEY & ASSOCIATES, LTD. ! Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 !. Washington, D.C. 20036 (202) 842-0034
- -- v r
i 258 1- And if we look at that compared to the corrosion
$[) 2 ' failure rates, this puts it in perspective okay, here is the 3- combined waste' package failure. Here is the failure from 4 corrosion as it was used^in the base case, and here is the l 5 contribution from the rockfall, so what you saw on the page 6 before only contributes this much to the total. And the 1
7 ' assumption is that because of the integrity of the waste 8 packages that no rockfall damage occurs from the size of
~
l 9 rocks that we postulate falling into the repository until j l 10 after 30,000 years. So this is why I think the comment was 11 made earlier, we think that both rock seismic effects are 12 not that important. I 13 If we look at criticality now -- and this is a ' 14 conditional analysis again -- assuming that a criticality l l (O s ,) 15 occurs, starting at 15,000 years and continues for 10,000 l 16 years, what is the contribution. Well, here is the base 17 case dose rate, up to a million years, and here is a 18 contribution from that conditional event right here, and you 19 can see that that contribution basically is just a creation 20 of a little bit more inventory, and the inventory that is 21 created, of course, includes more iodine, more technetium, j._ 22 neptunium and uranium, and those happen to be contributors j 23 to the dose, so you basically get some minor influence, but i 24 nothing even remotely approaching the base case. 25 DR. WYMER: That 's everything in the reposif y 1
'[\/ } ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 i
- - , ~
l i l iL 259 1 going.. critical?
'2 MR. VAN LUIK: No, no , this is a. waste package 3 going critical.
l l 4- DR. WYMER: Just a single waste package. l t 5 MR, VAN LUIK: A single waste package, yeah. 1 i And the reason that we think that this is a 6
)
' i 7 no-never-mind is if you look at the in-waste package K l l 8' effective, it goes up over' time and then goes down over I
-9 time, and then if you look at the time of waste package l
10 degradation, you can see that at the time that you start i 11 having significant failures of waste packages, the results l 12 of the -- basically the likelihood of your having a 13 criticality is already close to zero again. So the basic 14 point is that we don't expect in-package criticalities to !( 15 contribute because when they are most likely to occur is not 16 when you have waste packages failing, and basically the 17 'in-growth effects from the criticality, as we showed l -18 previously, are followed by waste package degradation. 19 DR. GARRICK: So the argument here is that as the 20 waste package degradates, you move into a less favorable 21 geometry; is that what you are saying, to support 22 criticality? 23 MR. VAN LUIK: Yeah. Why don't we have the person 24 that did the analysis explain that one, yeah. 25- MR. BARNARD: John, it isn't so much the geometry { l'- iD ANN RILEY & ASSOCIATES, LTD. lV Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
. . _ . _ . ~ . - . - . _ ~ . . . . . - . . - _ . _ . - . _ _ . - . .. .-.-.
260
-I as it'is theLfact.that you-have: decay.which changes,the-12 . inventory.of fissile nuclides that are essentially what.give
-3 you-the' reactivity. So-after roughly;100,'000 years, your f4 .freactivityLis down by about 30 percent, which in most cases-5 is-enough to make it_so that it's essentially. impossible to 6 'have a thermal-criticality.
.7 DR. GARRICK: Yes. I was glad that was the-answer 4 8 because if it was a geometry answer, that's a pretty.
9 difficult one to: guarantee. 10 MR. BARNARD: Yes, but we just have Mother Nature 11 ' working for.us -- l l 12 DR. GARRICK: Right. 13- MR. BARNARD: -- on that decay time. I 14' MR. VAN LUIK: And, finally, if we look at human I 15 intrusion-- .and this again is that_ stylized calculation ! 16 --assuming the waste.is moved directly into the saturated 171 zone at 10,000 years, we did a sensitivity study saying, L 18 well, how much mass could have been moved to the saturated i
- 19. zone,-and took two cases that kind of spanned the range of-i 20 what we thought might be likely, 550 kilograms, and 2700 i
L 21 kilograms, and then to exacerbate the effects, we decided on h 22' the high range of the dissolution rates, and then the low p '23 range of the dissolution rate for those two. We basically L 24 spread ~them apart as much as we can. 25 Like I said, this is at 10,000 years. So you see l' i ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut. Avenue, NW, Suite 1014 [. Washington, D.C. 20036 f- (202) 842-0034 L i
. . , . . ,, , _s -_.4 _ . , , , _ ,
261 ! 1 this little spike on the million-year diagram. Blowing up 2 ( that spike, you can see that is about a half order of 3 magnitude, going from the base case to the low dissolution , l 4 rate, low release into the saturated zone case, and then a 5 couple of orders magnitude increase right here. But 6 basically if you look at the absolute values of the dose 7 rates, we go from, you know, a negligible value of .01 to 1 8 millirera because of that intrusion. 9 Now this is what is shown in the VA. We are 10 looking forward with eager anticipation what the regulators 11 are going to tell us we need to do for this calculation, 12 because we see some problems, you know, depending on when 13 you select the case, that if you are -- you know, this could 14 easily become the target, in other words, for the entire ('h
\_) 15 licensing case if it has to meet the same exact target as 16 the base case. So -- and depending on when you select the 17 implementation of this, you know, it makes a difference.
18 But the way we see it is that the repository is quite robust 19 against human intrusion effects, given this type of 20 analysis. But there is, of course, other effects, as you 21 know. 22 Okay. This is where I was going to stop, and the 23 tough part now comes in. This is the base case results, 24 both probabilistic and deterministic. Now what we did learn 25 from all the hundreds, maybe even thousands, of sensitivity / \ \ ') ( ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 i
262' 1- and uncertainty analyses that were done? And I was going to () 2 turn things over to Ho'lly at"this point, unless.you want to 3 give me some questions. 4 I might just answer one thing. What we looked at t 5 in the' SPA is that whenever we had a pluvial, which is 80 6 percent of the time'for the probabilistic case and the 7 deterministic case, 80' percent of the time we are in a 8- -pluvial condition is'the water table was raised'120 meters. 9 We have not addressed the-Szymanski hypothesis in our TSPAs 10 whatsoever, because we basically believe that they are 11 incredible. 12 DR. HORNBERGER: Thanks, Abe. 13 What we are going to do is ask Abe questions, if 14 we have them, and then we are going to take a 15-minute 15 break and reconvene, okay? 16 MR. VAN LUIK: Oh, wonderful. 17 DR.'HORNBERGER: And we'll go counterclockwise 18 this time, so let's start with Marty. 19 DR. STEINDLER: The draft version of 63 tells you 20 when-the human intrusion is going to take place, a hundred
. 21 - years after closure. If you believe that, and you then go 22 through the analysis, do you find that you have to incur any 23 other EBS design changes in order to meet the apparent 124 requirement?
25_ MR. VAN LUIK: It is my impression that we have ANN RILEY & ASSOCIATES, LTD. O. - Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
......_-.._,_.__.,-._._....._____._.__.m. _m._._ . . _ . _ _ . . _ . _ _ . _ . _ _ . _ _ _ _ _ . _ _ .
263 1 not evaluated that at this point. The advent of 63 on the Mj i 2 web is relatively recent, and I don't think we-have really > 3 looked at that. What we have done is quickly looked at what 4 would be the effect of a -- you know, of an intrusion that
.5 was a worst case analysis like this,.and the absolute value 6 is not too scary, but:the implications for licensing in 7 terms of intervenors coming in with worst-case scenarios 8 worse than we have, you know, postulated here, there's a lot 9 of potential for mischief. I don't think that's an 10 ,
intrusion that hits one waste package, no matter what the 11- design, even the current design, it's really unlikely that a 12 drill bit used in that kind of rock would penetrate that at 13- a hundred years because it is still perfectly intact. So -- 14 but I think, you know, the basic answer is we have not
\ 15 really evaluated it that early. We have -- in fact, even at l
16 the thousand years or the 10,000 years that we have done the 17 waste packages, basically we have to presume that they are 18 using some drilling technology that will penetrate those 19 metal. waste packages, and-that they will not stop drilling 20 when they see strange material coming up in their drill bit.
..21 But obviously we are doing those analyses. l 22 DR. HORNBERGER: Charles?
23 DR. FAIRHURST: No, I'll-hold my questions for 12 4 later.
' 25 MR. VAN LUIK: For holly.
1 () ANN RILEY & ASSOCIATES, LTD. Court Reporters l l 1025 Connecticut Avenue, NW, Suite 1014 ' Washington, D.C. 20036-(202) 842-0034
- ..- _. .- . . _ . - - - - . - . . . - . . . . = . - . - . . - _ - . . - . . . .
I 264 11 DR. WYMER: I'll' follow up on something that sort ' () 2 of caught my interest on this criticality situation. What-3' . sort.of concentration process was assumed for uranium-235 4' and.perhaps plutonium ~in this analysis?
-S- MR. VAN LUIK: I think Rollie-will'have to address
- 6. that.
7 _MR. BARNARD: Let's talk aboutLit in-package. 8 MR. VAN LUIK: This is Rollie Barnard, Sandia.
;9L MR. BARNARD: Yes. Sorry. I'm Rolston Barnard, 10- still.
11' [ Laughter.] 112 MR. BARNARD: In-package, what was looked at was I 13- design basis fuel, which specifies the inittal enrichment
)
14 land the specified burn-up which results in an amount of a () 15- U-235 remaining, plus some in-growth of plutonium.
. And at 16; the time that Abe showed of 15,000 years, that is the point 17' at which the decay and so forth gives you the greatest 18 reactivity. l
\
19 Essentially design basis fuel is the one that they l i
- 20 have looked at -- no, excuse me, they have also looked at j 21 that particular fuel'which is the combination of enrichment E2 2 - and burn-up which would give you the maximum reactivity l 23- also. So that has also been looked at.
24-' Does that answer the question, or -- 25 DR. WYMER: Basically you didn't look at anything i ~ 1-ANN RILEY & ASSOCIATES, LTD. l '[- Court Reporters 1025 Connecticut Avenue, NW,-Suite 1014 Washington, D.C. 20036 (202) 842-0034 L
, . .- .._.-. - ~ . - - . . - . . . . - - . - . _ . - . . - -
1 l 1 265 1 outside of the package? You didn't look at the far field? y 2' Yes,'that has been looked at also, f MR. BARNARD: 3 and there are several scenarios for methods for
'4 : concentration in the far field and in the near field. They
'5 have been. looked at, not in great detail in this most recent !
6 TSPA, but only sufficiently to show that they look like they 7- I are of even lower probability of occurrence than the ' 8 in-package case. 9; MR. VAN LUIK: Particularly when you'have an I 10 average burn-up of maybe 40,000 -- ! 1
'll- MR. BARNARD: Yes, 35.
12 MR. VAN LUIK: Thank you. l l 13 DR. GARRICK: I want to alert everyone to the fact l 14 that we have a hard deadline of 12:15, all right, so we will 15 -- we can run with questions until 12:15, but we will have 16 to break at 12:15. 17 MS. DOCKERY: Thanks. 18 DR. HORNBERGER: Andy? No, I have one that 19 follows up on a question that somebody asked Jack earlier.
.20 From what you have given us, both the base case and the 21 probabilistic analyses, it certainly appears that your 22 treatment of the engineered barrier system has a very 23 significant influence on the calculated doses. Given that 24 various groups, including your TSPA peer review panel has 25 pointed out that your knowledge base and your models for h'j ANN RILEY & ASSOCIATES, LTD. %/ - Court Reporters J. 1025-Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 I: (202) 842-0034 V
( [ 266 l l [ 1 base' degradation and the geochemistry of the.near field are l r X r U 2 imperfect, shall we say, to be charitable, how do you see 3 the analysis that you have presented in terms of being L 4 realistic? 5 MR. VAN LUIK: That's an excellent question. 6 Jhat's one that we wrestled with. t L 7 MR. BARNARD: That.and many others. i 8 MR. VAN LUIK: Yes. The expert elicitation on 9- degradation rates for the waste package.I think pointed out 10 very fairly that it's because of our uncertainties in the 11 .near field that they had great uncertainties in the 12' performance of LR-22 in particular. They thought that 13 basically the corrosion allowance material we were in pretty 14 good shape on; a lot of experience with that material; that 15 we know the de radation rates; and it goes away, you know, 16 relatively quick compared to the alloy 22. And basically 17 their challenge to us was to define that near field 18 environment more closely'so that we can model the 19 degradation of alloy 22 more closely. 20 At the same time they recognized the deficiency in 21 the, you know, the data base for a new alloy, basically. 22 It's been around a few decades is all. We hope that through 23 our testing program, which is ongoing and from which we have 24- now I think about one and a half year's worth of data, the 25 ongoing program will continue to confirm basically the range i i j ' ANN RILEY & ASSOCIATES, LTD. Court Reporters l l 1025 Connecticut Avenue, NW, Suita 1014 Washington, D.C. 20036 (202) 842-0034
267 1 that was used in these analyses and even-bring that range () 2 down a'little bit. 3 However, you know, we see as one of the challenges 4 for obtaining.not only the construction license but the 5' amendment to the license to emplace waste is to have a much 6 ' firmer story on.the design and its anticipated behavior. So 7 we'see that, you'know, at the time offthe initial license. 8 application to construct, we will:still have an incomplete 9 ' data base to make a convincing -- I think we will make a 10 convincing case, but the uncertainty will still be great, l 11 but'that uncertainty should come down, you know, before we 12 receive the license to emplace-and, of course, by the time 13 of repository closure,.if we go 100, 200 or.300 years, I l 1 1 14 think that the NRC will insist that we continue work to 1 ! 15 narrow the uncertainties before we finally close, whenever
- 16. that might be. I 17 So you got me. It's an area of great uncertainty
- 18. at this point. I think we have captured the range of 19 uncertainty,_and I-think that range is shown in the horse 20 tail diagrams. Those diagrams would collapse quite a bit if 21 the-alloy 22 corrosion rate could be pinpointed more 22 exactly. I l
L 23 DR. HORNBERGER: Okay. Thanks very much. I 24 We are going to take a 15-minute break, which, as 25 my colleague has pointed out, would be when the big hand is L:- s a-f- () ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 1; i (202) 842 0034 l m ., w w iwv,ry
^ <-- pr w- -
5-a
I t 268 i 1 two minutes past the 12.
/"'N
( ) 2 [ Recess.] L/ , i 3 DR. HORNBERGER: In case you haven't heard the 1 4 gavel, we are reconvening, and the next presenter will be 5 Holly Dockery. We are going to hear some more on TSPA. I 6 want to alert everyone to the fact that we have a hard ; 7 deadline of 12:15, all right? So we can run with questions 8 until 12:15, but we will have to break at 12:15. 9 Are you wired, Holly? 10 MS. DOCKERY: I am usually loud enough. Nobody l 11 ever complains. I'm wired now. Can you hear me better? 12 DR. HORNBERGER: Okay, please. Go ahead. Start, 13 Holly. 14 MS. DOCKERY: This is really just a continuation p) (, 15 of what Abe was talx!ng about, and what I am going to be 16 talking about is some of the comparative analyses, the way 17 in which TSPA analysts have tried to take apart some of the ) l 18 detailed calculations and tried to determine where the 19 uncertainty lies and how various parameters, various factors 20 affect the overall results. 21 So the primary goal of the comparative analyses, 22 the sensitivity studies, were to try to find out how the 23 uncertainty in the parameters and in the models affected the 24 TSPA results, and of course to some extent look at how the i 25 parameter factor -- each individual parameter is sensitive.
- I I ANN RILEY & ASSOCIATES, LTD. I
! \# Court Reporters l 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
.- . = _ . - . = _
269 1 We have kind of broken up the two types of studies ! ew f 2 2 that we did into the regression-based analyses which were G 3 performed on the probabilistic base case that Ata was 4 talking about and then the run-off sensitivity analyses 5 which'are done on the deterministic, and I will tell you the 6 whys and the wherefores on each individual one. 7 This table is, for your information and when you 8 are reviewing the TSPA information, just so you'll know ! 9 whore uncertainty was handled in the various instances, for 10 the principal factors in various cases there was 11 heterogeneity or variability directly in the base case 12 analysis. We are sampling from distributions, and so you ! l 13 sampled the uncertainty in that way. l (m 14 Sometimes we had additional uncertainty like ( ,) 15 conceptual model uncertainty that was also handled in the i 16 base case. An example of that would be fracture versus 17 matrix flow. 18 Then in some cases we simply looked especially in 19 the one-off analyses at specific aspects of the system that 20 we found to be interesting after we got results and we 21 wanted to explore a little bit further some of the specific 22 aspects of those analyses. 23 In terms of the regression-based sensitivity 24 analyses, the main reason we did this was to try to find out 25 which parameters contribute the most uncertainty to our f l g. ( ANN RILEY & ASSOCIATES, LTD. A Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
r l i L 270 1 'results, so again this is on the probabilistic base case () 2- analyses. ! 3 The results of the analyses are all sampled 4 simultaneously, so you have the full range of uncertainty l 5 that is maintained and the interactions among the various L 6 parameters are also maintained because you are basically 7~ running it in a full probabilistic mode. 8 The results have been examined using various
'9 methods -- the scatter plots, regression analyses,
-10 contributors to variance. Each one of them have their own 11- benefit and ways of looking at the data and scatter plots 1 l
)
12 are nice visual ways to identify nonlinear behaviors and you 13 can certainly see things visually, but they are not quite i i 14- as -- they are a little bit more qualitative than I
/~ l i
15 quantitative. 16 Regression analyses are a more quantitative way of 17 handling the results and running the results in such a 18 manner to pull out the specific aspects of the uncertainty 19 and what I am going to show you in a minute are the
-20 contributors to variance. We found that to be a nice visual 21 way to look at how much variance is decreased every time you !
22 take out a specific parameter or a specific factor -- 23 besides, I don't want to have to get into the details of 24 statistics, since I am just a geologist -- all those T tests 25 &nd things make me nervous. i !ID ANN RILEY & ASSOCIATES, LTD. ,- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
i , i 271 1 Be sure this looks good. It's over there, Andy. () 2 It looks horrible to me. 3 This is the 10,000 year dose rate history plot l 4. that you have seen before, and so you can see, the 5 uncertainty range in the 10,000 year plot basically is this l 6 entire sweep because we have shown every' single realization 7 and you can~see based on all of the parameters and all of 8 the uncertainty that is incorporated in those parameters in 9 the conceptual models we have a fairly wide range of 10 uncertainty. 11 Then as you.go back and you do a step-wise 12 regression where you start adding variables and each time 13 you add the variable you look back and you see how much did p 14 that change the variance and the dose rate, then you ( 15 eventually can come up with these r-squared loss plots, 16 .which again is just how much does each one contribute to 17 variance. As you'take a parameter out, how much does your i l 18 uncertainty or your variance decrease as you take these. 19 parameters out. 20' You can see that in the 10,000 year history, when 21 we do the regression analysis, we find that the seepage L L '22 fraction, the amount of water that hits the waste package 23 .and the number of waste packages that are hit is the most
- 24 important parameter in terms of uncertainty in loss of 25- variants. Then the alloy 22 degradation rate -- and then I/ '
ANN RILEY & ASSOCIATES, LTD.
\
, Court Reporters 1025 Connecticut Avenue,.NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
. m_. _ _ _ - _ . _ _ _ . - . _ . _ _ _ _ . _ _ _ . . ._... _ . _. . . _ . . , 272 )
1 the number of juvenile failures and the saturated zone
.(-s) 2 dilution factor show up as somewhat lesser in the 3 uncertainty analysis.
4 When you go on to the next timed step, and this ) 1 5 will answer one of the questions about how do you know what 6 is important at various times versus the 10,000 year and 7 100,00 year,-million year, are there differences that you 8 would look at, this certainly shows that in some cases yes 9 and some cases no. This is theLhorsetail or hairball 10 diagram for the 100,000 year dose rate history for all the 11 pathways, and then you look at the r-squared loss plot for 12 this. 13 You will see again in this time seepage fraction I 14 and alloy 22, the corrosion resistent alloy, are still the I i ) 15 primary factors contributing, but then you also start to 16 show some of the corrosion rate variability showing up and 17 the number of juvenile failures is a little bit less 18 important obviously as you get out.to longer timeframes and 19 the corrosion is starting to have more releases than the 20 juvenile failures. 21 Then as you get into the 100,000 year, we can 22- still see that seepage fraction is very important. Now you L 23 also have saturated zone dilution factor, the dose i 24 conversion factor for neptunium, and the alloy 22 corrosion 25 rate is dropping way off as you get to these later times. ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l
I i 273 1 One of the things that Ernie will be showing p) s, N J 2 whenever he goes through his talk on prioritization is that l 3 for each one of the individual factors, principal factors, 4 from the postclosure repository safety case, we have looked 5 at the importance in each timeframe as well as sort of a 6 what is the overall importance, what is the highest rating 7 that we gave it. 8 We also looked at how things vary over time. 9 DR. HORNBERGER: Holly, a technical question. Are 10 these regressions done on the parameter values themselves or 11 on the logarithms of the parameter values or what? 12 MS. DOCKERY: The logs. 13 DR. HORNBERGER: Okay. 14 MS. DOCKERY: One thing that I had wanted to show
/'-'N
( ,) 15 is one of the ways to start looking at values that we 16 thought should have been important, I mean neptunium 17 solubility, cladding failure -- we thought that those should 18 have started to show up but because there was such a large 19 uncertainty range associated with seepage and also with C-22 20 corrosion rate, those kind of overshadowed some of the 21 things that were more important. 22 So the next step we took was taking out both of 23 those highly uncertain parameters and see what else starts 24 to pop up, because I am sure that you would be the first 25 ones to point out that if you have a really large range of l [) A/ ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
274 1 uncertainty or really small range on something that is (3) 2 important, it is not going to show up in this type of 3 analysis and that is why we have to do the other type as 4 well, but what this particular one is showing is you can see 5 compared to the previous horsetail diagram, which was about 6 this wide in a million years, you can see how much 7 compressed the overall uncertainty in these dose rates is 8 whenever you takn out those two highly uncertain parameters 9 of C-22 and seepage fraction, and so in this case you start 10 to see the other parameters popping out and that is the 11 saturated zone dilution factor, the cladding failures, the 12 biosphere dose corrosion factor -- in this case for 13 neptunium again -- there's dispersivity, cladding failures 14 due to mechanical failures, and then the amount of alluvium (D i ,) 15 in the saturated zone, so we start to get a very different 16 suite of parameters showing up as we take those other two 17 out and we would continue to -- we didn't do as many of 18 these analyses as we would like. 19 I am sure that we will be doing many more on these 20 particular results as well as we change our models, but the 21 benefits of this method is that we can look at which 22 parameters contribute the most uncertainty to the dose rate. 23 They don't tell you an absolute value of which one is 24 most -- which parameter is most important, which factor is 25 most important to safety, but they tell you which one is l i
) ANN RILEY & ASSOCIATES, LTD.
l [\- ' l Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
i l L 275 " , xi. driving your. dose rate. () 2 You have.your synergistic effects maintained. 3 Other methods where you sample individually you may have o 4 coupled effects that aren't captured if you don't use this
- 5. kind of method. And it tells you where you need to look.at-
;6 reducing your uncertainty, because as you reduce the' 7 l uncertainty;in your factor it may no longer be as important 8- in the overall system.
9 . CHAIRMAN GARRICK: 'Ho lly, I'm familiar with 10' regression analysis, but I want to be clear -- 11 MS. DOCKERY: You're probably much more familiar 12 than I am. l 13 CHAIRMAN GARRICK: Well, I want to be clear on t 14 . understanding, when'you say remove the variable, just what ; i 15 you're saying. l l
- 16. MS.'DOCKERY: My understanding is-that in the
- 17. stepwise regression. analysis you sinply take out the L 18 parameter associated with first I believe the seepage and I
l 19 then with.C-22, and then your analysis results are shown 20 without'those two variables added into your overall I 21 analysis. 22 MR. VAN LUIK: This is une Van Luik. I think what i L 23 -they did was they took out the uncertainty in the parameters
- 24. and fixed them at their main value.
25 MS. DOCKERY: Right. Exactly. I'm sorry. i: I If ANN RILEY & ASSOCIATES, LTD. l: Court Reporters t' 1 1025-Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
- - - - M e -: v p-pq g--,v..ymvg. -v -,-w+v e----
l 276 l' CHAIRMAN GARRICK: So the. input was -- they had an l [~T 2 input for seepage, but it was just'the mean value.
%I 3 MS. DOCKERY: Exactly. I'm sorry. !
l 4. DR. HORNBERGER: Now wait. Now hold on. 5 [ Laughter.] )
. l
- 6 Are you talking about the modified parameter case, !
j 7 or are you talking about the variance reduction? Because , 8- .the variance reduction you're just removing the. variable q L 9 from the regression. ! l 10 MS. DOCKERY: Right. But in the modified i 11- . parameter case. l 12 DR. HORNBERGER: In the modified parameter, you're
.13 fixing -- j 14 MS. DOCKERY: Exactly.
m) 15 .DR. HORNBERGER: The values of the base case. 16 MS. DOCKERY: Yes. 17 CHAIRMAN GARRICK: Okay. So, one, you're removing 18 the variability. i 19 DR. HORNBERGER: Yes, so the modified parameter I 20 case you've taken the variability out of the seepage and the 21 C-22 corrosion. 22 MS. DOCKERY: And the C-22 corrosion.
'23 DR. HORNBERGER: Okay So that's why the horse 24 tail.gets narrower --
25 MS. DOCKERY: Right.
~\ ANN RILEY & ASSOCIATES, LTD.
f((ss
\ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014
[ Washington, D.C. 20036 (202) 842-0034
'l 1
277 L :1: DR. HORNBERGER: Because you've~taken out those l.
/T 2 ~uncertaintie'. But when you look at the plots on the right, i
V 3 when it's the reduction in r, that's just taking the 4 variable out of the. regression. 5 MS. DOCKERY: Exactly. 6 DR. FAIRHURST: In the first case,.do you just 7 assign the mean value and go'with it? 8 IMt. HORNBERGE'. -That's what's meant by the l~ 19 -modified. parameter case. 10- MS. DOCKERY: Obviously you can't run an analysis 11 without having seepage or without having degradation rate. l 12 So all you do is you take all of the uncertainty that's ) i 13 associated with it and give it the mean value. And so every l 14 time you sample, you use the mean value for that particular () 15 factor or parameter for the modified parameter case. And in 16 that-way you start to look at rather than how that
- 17. uncertainty band, you're sampling over an extremely large i
18 range for.each of your realizations in the. base ~ case. 19 In the modified case you take those and only those L
'20 two'and set them at single values, which is their mean ;
21' ~ value, and sample everything else. And that way you can j 22 look at the other uncertainty -- the uncertainty associated
.23' with other factors and other parameters instead of those
.24 two.
25 Does that make sense? l.. ANN RILEY & ASSOCIATES,'LTD. (( . i Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington,.D.C. 20036 (202) 842-0034 i,
, , . . . _ . - - . . - . _ . . _ - , , = _ , , . , - ,
. _ - - - . ~ .. . - . . - . - - - . - . _ _ - . - ~ . - . - _ ~ . . - _ - _ .
278 l 1 DR. FAIRHURST: Yes. I guess I'm still wondering,
/ \
2 because the mean value, even that is based on a set of y \~sl ! l L 3: _ assumptions. 4 MS. DOCKERY: Absolutely. 5 DR. FAIRHURST: And one could investigate that l '6 ' parameter by itself, and the_ justification for that mean l I i-
- 7 value.
;8 MS. DOCKERY: That's what -- when I.get to the one i
9 on sensitivity analyses, that's a little bit more looking at
- 10. the. individual parameters. This is sort of a way of looking
.11 at all of-the uncertainty in the entire calculation and L 12 where does the most uncertainty -- which parameters does i-13- their uncertainty range drive your spread in dose rates.
14 And so where do you want to put your efforts on
) 15 trying to decrease that uncertainty range so_you'll have a 16 better handle on what your -- you're basically taking your dose range and trying to decrease it. And that one also 18- will show you more -- this-is what this particular parameter l 19 did in terms of changing the overall dose rate.
L 20e Okay. Just for a minute back to the 12 1 regression-based analysis. There are limitations on this 22 method. One is that it is extremely intensive, and so you l 23 can't.really do all the sensitivity studies you might like 24 to do. 25' For instance, we run nine flow fields for the UZ I i t
. ANN RILEY & ASSOCIATES, LTD. ;
Court Reporters
- 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
~ , . - __ _ - -
279 ' 1 flow and then assign probabilities for each one of those (n) 2 flow fields. We can't run an infinite number of unsaturated 3 zone flow fields in three dimensions to go back and do the ; l 4 sensitivity study, so we have to find another way to handle 5 and look at some of that. 6 You can't look at "certain parameters." 7 Parameters for which you have a very well defined range, a 8 very small range, or perhaps a given value are not going to 9 show up in this kind of analysis either, because this is 10 showing you how much the uncertainty and the larger ranges 11 of uncertainty affect your dose. 12 And you have to be extremely careful that you have 13 an appropriate range of uncertainty. If you assign a huge 14 range of uncertainty, then it's going to show up as (~ ty ,/ '15 important or if you show more likely an inappropriately 16 small range of uncertainty, it also won't show up. And so 17 that's something that is really important to be very careful 18 with, and that's part of why we're doing some of the one-off 19 sensitivity analyses, to take another type of a look at the 20 parameters. 21 So going to what we called one-off sensitivity 22 analyses, the reason we did that is to look at the 23 uncertainty that you don't necessarily have in a PDF. 24 Regression analysis looked nicely whenever you have a nice 25 range of uncertainty and probability distribution function I t
,.~
!./ ANN RILEY & ASSOCIATES, LTD.
\ss) Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
-. . . - -. . .-. - . - . _ . .- - -. . - - . ~ ~.- --.
280 1; .for parameter, you can look at it. But there are other
] ) 2 types of uncertainty, and cladding /no cladding. That's a 3 major' uncertainty range.
4 So this is a way to look at some of those factors. 1 5 And we use it on the deterministic base case results, ! l 6 because we're simply trying to look at not the precise value 7 of dose you might expect to get, but rather the effect on 8 -the system. And so you want to look at the change that 9 you're going to get in the system rather than the absolute 10 value for dose. 11 .And we've varied these factors differently 12 depending on what kind of assumptions we used to 13 conceptualize the factor or the parameter. 14 I'm only going to show you a few of the (O ,j 15 sensitivity analyses, and in most cases the ones that Ernie 16 -next is going to tell you turned out to be relatively 17 important in the prioritization scheme. We did a number of 18 -sensitivity analyses, anywhere from maybe two to five for 19 each one of the factors we did different types of analyses.
-20 So I'm going to show you a few of them and just kind of give
-21 you a flavor of how we've deconstructed things and try to 22 look at them.
~23 The first one I'm going to show you is 24 infiltration rate, because, as I said before, with the L 25 mountain scale flow models and the way we modeled i
L ANN RILEY & ASSOCIATES, LTD.
- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
_ . _ _ . . _ . _ _ _ . _ _ ~ . . _ . - _ _ _ _ . . _ . _ _ _ _ . - _ _ . . . _ . . - . . ~ . _ . - 281 1 infiltration in the connection,'it's really difficult to () 2- change the flow fields and come up with a lot of new flow 3 fields-to run. So instead we changed the probabilities 4 assigned to the flow fields.that we had. , 5 We started our with a base case, as you probably 6 are very well familiar, that we had a base case where we had 7 an infiltration rate, and then we divided it by 3 and 8 multiplied it by 3 and came up with different flow fields. [ 9 'Well, in the base case, the I divided by 3 case, 10 had a 30-percent chance of being sampled, the I times 3 had J11 a 10-percent chance, and the standard infiltration rate 12 would be 60. Well, we kind of inverted that in one case and 13 looked-at well, what if it was -- the divided by 3 was 10 14 and-the times 3 was 30. And we sort of changed the values () 15 of the probabilities a little bit. ! l l 16 But in this particular example, what I've shown is L 17 that if we took the base case, which is here, I think, and 18 then we ran the infiltration rates from the base case at 19 their 5th percentile and at their 95th percentile, and held l 20 those values, then what kind of change would we get in the 21 dose rates using the 5th percentile values and the 95th 22 percentile versus the mean value? And we looked at the l 23 change in the dose rate at 10,000 years, 100,000 years, and 24 a million years.
'25 And basically what happens is because you get i
O ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
~ -. - - - - . . . .-
l 282 1 infiltration increases you get more percolation, more l [ 2 seepage, and more waste packages contacted, and this 3 basically causes about a factor of ten times greater or less 4 dose over the 100,000-year period, and what we were looking 5 was from peak dose to peak dose, what kind of factor of 6 change occurred as a result of sampling off of those 7 different -- the 5th and 95th percentile. And then we did 8 the additional look at the changes in probabilities, and we 9 came up with the change ~in your probabilities changed it 10 from a mean of 7 to 8 to about 11. So again there is a 11 'small change, a significant change if you sample the 12 infiltration at the 5th and 95th percentiles, but you're not 13 getting many orders of magnitude change as you are in other 14 areas. l
-s g ,/ 15 DR. HORNBERGER: Holly, just again to make sure I 16 have this clear in my mind, a three-times higher 17 infiltration, your long-term average in your base case is 40 L 18 millimeters per year, as I recall, for percolation. Is that l 19 right?
)
20' MS. DOCKERY: I'm sorry? I 21 DR. HORNBERGER: Percolation flux is about 40 22 millimeters per year in your base case. 23' MS. DOCKERY: That's -- you're talking about 24 infiltration or percolation? 25 DR. HORNBERGER: Long-term average. L
/~'\ ANN RILEY & ASSOCIATES, LTD.
f !s-) Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
283 1 MS. DOCKERY: Long-term average, the percolation I~ 2 is something like 480. [ N_)) 1 ! 3 DR. HORNBERGER: Okay. So three times higher 1 4 than -- 5 MS. DOCKERY: This is infiltration, however. 6 DR. HORNBERGER: I understand. But is it a 7 nonlinear relationship between percolation flux and I 8 infiltration rate? j 1 9 MS. DOCKERY: Yes. I 10 DR. HORNBERGER: Okay. So three. times higher 11 infiltration translates into how much higher percolation ! 12 flux then. That's my question. l 13 MS. DOCKERY: Abe, do you remember the numbers 14 right offhand of what the three times on the base case was? (rN) 15 I can get back to you, but I don't recall the . 16 exact numbers. 17 DR. HORNBERGER: Bigger than 40 and less than 120. 18 MS. DOCKERY: Yes. 19 Let's see. But I can show you what happens when 20 it gets down to the repository horizon and becomes seepage. 21 The seepage fraction and the seepage amount are 22 the uncertain aspects of the seepage into the drift. We 23 looked at what would happen if you took the 5th and 95th 24 percentile of seepage into the drift. When you have no 25 seepage, you get extremely slow corrosion of the Alloy-22. l ANN RILEY & ASSOCIATES, LTD. i f")D
\- Court Reporters
! 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 1
-- . .. - . . . . - . - . ~ . . . - . - - - . - _ _ . - . - . - . - - - . ~ . .
L 284 1 Basically,'as we were talking about earlier, the humid-air 22 J) corrosion has very little impact on the C-22. You have to 3 be basically ~under a-drip before you can start to corrode ~ l 4' the C-22. l 5- ~ So as a result, whether or not you see dripping 15- 'and how much dripping occurs is extremely important. In- j 7 this case what you're seeing is the 5th -- I do'wish I could i 8 read these -- you've got the 5th and the 95th percentile and l l 9' lthe expected-value case -- l '10 Thanks, Andy. l l1 11' For-10,000 years, 100,000 years, and a million i 1 12 years, and you can see that -- something happened with that p 13 little curve up there. Now you can see them.
- 14 ' .The fifth percentile basically is dropping off --
15- start over again. 16' The fifth percentile here is basically at zero, 17 and so you can see'that the range -- l 18 DR. HORNBERGER: It's a little bit to the left. 19 MS. DOCKERY: Is several orders of magnitude 20 change in the overall dose rate. ' 21 DR. FAIRHURST: The same is true at a million 22 . years, isn't it? 23 MS. DOCKERY: Yes, the same is true of a million 24 years, until you get -- down over here. 25 Some of the individual parameters that we looked ([" ANN RILEY & ASSOCIATES, LTD. l Court Reporters lt 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 __ ~ _ . - _ _ _ - . _ . _ _ _ . .
l 285 1 atfafter we -- after you look at the incredible effect that - O g j 2 seepage has on the overall dose rate, then you kind of want i j 3 to'deconstruct and find out well, what aspect of seepage was i l 4 it that was most important, and we looked at fracture
- 5. permeability in the fracture area, parameter alpha, and it turned out that when you went through and did the various 7 analyses, that it was the increased fracture that had the l '8 most effect on the seepage.
9 There were varying degrees of impact, but it's one 10 of the things that we would like to continue with l 11' deconstructing these analyses just looking at the various 12 parameters in more detail and find out if you change l 13 individual parameters within these factors, which ones will
-14 give you the most influence on your dose-rate change. 1 15 DR. FAIRHURST: Which way did that change occur?
E16 If you increase the fracture, did you decrease the seepage? 17 MS. DOCKERY: It decreased the seepage; yes, it. 18 decreased the dose. 19 DR. FAIRHURST: Because it went through the -- ( '20 MS. DOCKERY: I'm sorry? i 21 DR. FAIRHURST: Because it was going through the j 22 matrix. 23 MS. DOCKERY: It's going through the matrix, l 24. and -- yes. 25 DR. FAIRHURST: Takes a long time. i /~s ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 [ (202) 842-0034 ( I tr w *y e*wnr:: me-a m- w-s .m
- y- y- y
286 [ l l' MS; DOCKERY: .Um-hum. Exactly. }( l 2 .MR. VAN LUIK: Can I. address George's question? l 3 MS. DOCKERY: Sure. 4' MR. VAN LUIK: This is Abe Van Luik. For the !~
'5 base-case infiltration we had 7.7 millimeters a year, 8 l 6 basically. . When you double the precipitation for the l
t ,.. 7 long-term average, it multiplies-the infiltration by 5 and ' j 8~ goes up to 42; and'when you triple it, it goes up to 110. 9 So it's quite'a nonlinear relationship. ( And there's ranges L 10 on that also, of course. So -- 11 DR. HORNBERGER: But the answer to the question 12 then is because you multiplied by a factor of 3, your i 13 percolation flux for'the -- was about 110 millimeters per 14 year. 15: MR.. VAN LUIK: That's correct. L 16 DR. FAIRHURST: Four it's 110. i L 17 DR. HORNBERGER: -Eleven centimeters a year. Yes. 18 MS. DOCKERY: The runoff sensitivity that was done
' 19 . for Alloy-22, again, as I said before, the only packages !
20 that start to degrade due to corrosion are the ones that see 21 dripping water, and the mean degradation rate and the b 22 variability are uncertain. And as we said before, that's a 23 lot, because the local conditions in the repository were not L l 24 the near-field environment conditions, the actual conditions l l 25 'down at -- the chemical conditions down at the bottom of p i [1 'r . ANN RILEY & ASSOCIATES, LTD.
- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 r .. . _._ _ _ __ ___ _ , . .-. . _ _ - .
r j l l 287 1 i ! 1 ' pits-are somewhat uncertain. So we've got several orders of () 2. magnitude again in dose rate that's introduced by sampling L 3 .at the 5th and 95th percentile. 4 Some of the things that we have looked at L, 5. specifically in the degradation. rate is whether or not we j 6 -saw a drip and that changed'-- if none of the packages saw l 7 drips, we.didn't see any releases for hundreds of thousands t j 8 of years because the packages didn't corrode through given 9 our C-22 models that exist today. { L 10 The percentage of surface that was wetted really ) 11 didn't seem to.make much difference' The chemistry of the . l i 12 water didn't make much difference but it is possible that we 13 were' extremely conservative in that the assumptions that we L 14 were making and that we might see more change if we change l c (~\ 15
\ ,/- some of those assumptions.
16- Interestingly enough, MIC, the Micrr*>ially Induced
-17 Corrosion, actually did increase the degradation rate by 18 'several orders of magnitude and so that was something that 19 was interesting and unexpected to us, but it is completely
.20 driven by our assumptions as everything else is.
21 Juvenile failures -- only important again in the
- 22 early timeframes. It increased by about a factor of eight 23 and then the patch size, how-large were the patches --
U 24' basically very little effect compared to whether or not you 25 got a drip, t
. ANN RILEY & ASSOCIATES, LTD.
Court Reporters 7 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 b -
. .? , . - - - , - . . ., , , . - , -. .
288 i So we are starting to focus in on the things that ( ) 2 we really need to'look at in our various models. 3~ I thought you guys would enjoy that -- so what's 4 new? 5 (Laughter.] 6 MS.-DOCKERY: Cladding degradation is uncertain -- 7 ha ha. This is a surprise to us all? 8 We didn't have a very wide range though. We know 9 it is pretty uncertain but'the information that we had on 10 cladding at the time didn't give us a wide range in 11 degradation rates and we didn't really get much effect at 12 all until after about 250,000 years or so, and so what we 13 'did is we said okay, we'll look at what happens if all the 14 cladding is basically failed when it comes in the ( 15 repository, what happens if it starts to fail so that you 16 have most of your degradation done at 100,000 years, what if 17 it happens at mostly a million years, so you are kind of 18 changing your degradation rates over time. 19_ As you can see, the bounding assumptions basically 20 say that we are going to increase it by a little over a 21 factor of 50 in the dose rate and this is mostly because you 22 have more technetium available as the spent fuel rods are 23 exposed. You have more technetium but then as you get to 24 somewhat later times, you are solubility-limited again, so 25 it becomes a-little bit closer. i l ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 i
l
.289 1 I am sure you can see'very well it starts to
() 2 converge on more of the' base case parameters because you are
'3 being driven more by your neptunium solubility.
,4 DR. HORNBERGER: So this indicates that for the 5 long timeframe cladding is really important in the base 6- case?
7 MS. DOCKERY: Yes. It's also important in the 8 earlier times as well. It's just that the strange thing 9 right at 100,000 years it becomes a little bit less, but I 10 think if 10,000 years is our standard then cladding is going 11 to be an issue -- if a million years is the standard, it is 12 too. I think we are not out of the cladding woods yet. 13 DR. HORNBERGER: No. I was just noticing that it 14- looks like fairly high doses at long timeframes without 15 cladding. 16 MS. DOCKERY: Yes. Okay. Neptunium solubility -- 17 again this was one that just really surprised us when we did 18 our first regression analysis and neptunium was nowhere to 19 be seen. It's always something that has been shown as 20 important in our previous TSPAs. ) 21- As most of you know, our neptunium solubility that 22 we have gained out of all the literature and experimental 23 data is extremely large. It's like at least three orders of 24 magnitude on the solubility, but at times less than 50,000 25 years. ~Obviously the technetium, as Abe has shown, O ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l'
- . , -. . - - .- . ~ . - - _ . _ . - . _ _ _ - . . . . . - . - - - . - . ..
290 L L1 technetium and iodine are really dominating, but after you- [J s_ C 2 get up past 100,000 years'then it is about a ten' times 3 greater or less dose, depending on how you pick your 4 solubility. 5 This kind of goes to one of the questions I think 6 that George had before is that in this case, on the fifth 7 and-95th percentile you are not sampling solubility. You 8 are on a log beta distribution so'you are' sampling the log, 9 ~ of solubility in this case. l- ! 10 In the late times you are looking at our control l 11 by the amountLof inventory that is exposed and in flux and 12 so that is why the cladding starts to take back over at this l l l
'13 later timeframe.
14 Another one of the things that you will hear is !- ,f- g (,) 15 that saturated zone dilution -- when Ernie talks -- is E 16 certainly one of the things that we think is important, and 17 it doesn't really show up well in this analysis because we 18 don't have a lot of confidence in our model. We are not 19 real happy with the stream tube approximation, so we j 20 certainly will be doing a lot on the saturated zone model i
- 21 and it is not clear that doing a lot of variations on the i
p 22 models that exist today is Toing to tell us what we really 23 need to know in terms of ,:ertainty. l\ L' 24- However, in addition to looking at just the 25: dilution rate, which was obtained from the expert ANN RILEY & ASSOCIATES, LTD. h ~ (\]~/ i
?
Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 L
_ _ _ _ _- . _. . - ._ ~,. -- . _ .. _ . ._ . _ _ . . 291-1 elicitation and we varied it from one to 20 with an expected 2- value of two, we looked at how long did the water travel in 3 the alluvium, you know, if you had a different percentage of 4 the pathway in alluvium, and that had a small effect but not 5 a-large effect. 6 The most important effects were later times and 7 what they kind of did was dampen out the climate changes. 8 We had some big wows in the dose rate due to the climate i 9. changes and that kind of dampens things out and it does 10 decrease the dose curve but less than a factor of 10. l 11 Whether or not -- how you use the stream tube l 12 approximation at least we used distinct stream tubes, one l l 13 stream tube, that also didn't make a big difference as we l 14 looked at the variations. It certainly wasn't as important 1 (' \ 15 ! \s,/ as the dilution factor, so we clearly want to have a better 16 saturated zone model and a better understanding of how the 17 water moves through the saturated zone and dilutes, and 18 we'll also be wanting to look at the connection between the 19 unsaturated zone and the saturated zone because the way we 1 20 have treated it now is simply it comes to the saturated zone 21 and it is spread out across that block, but we don't really 22 know whether that is conservative or not and we don't have a 23 good mechanistic basis for that. l r. l 24 So, benefito and limitations of this "one-off" 25 approach that I was showing you, well, it is computationally 4
'( ANN RILEY & ASSOCIATES, LTD. l Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
292 1 a lot simpler. You can do a lot of realizations and do a r~N (v) 2 lot of changes if you are simply changing probabilities or 3 adding or subtracting entire parameters or factors. So you 4 can look at conceptual models that you couldn't look at 5 before, because you wouldn't see them in your probability 6 distribution function, and you can also look at certain 7 parameters. Things that have a very small range of 8 uncertainty, you can test those and see how much impact they 9 have if you sample them at different intervals. 10 The problem with the method, and something that we 11 can't go off and say that we are going to use this in 12 exclusion either to show everything we need to know about 13 the system, is you can't look at the relative importance of 14 the factors one against another. You can't look at what I A. ( ,) 15- showed you and say, oh, well, I can tell the difference 16 between a saturated zone and infiltration. 17 One of the most important things that it doesn't 18 show is how other factors interact, and, so, you lose some 19 of the interactions that might change the way a particular 20 factor behaves in the overall system. And, so, we expect to 21 continue to use both approaches to keep looking at the 22 sensitivity of the uncertainty and how the parameters need 23 to be investigated for the SS, the SR, the LA. 24 I also wanted to just quickly show you some design 25 option sensitivity analyses which were handled in the same t
/ D L
i ANN RILEY & ASSOCIATES, LTD.
\d Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l
Washington, D.C. 20036 (202) 842-0034 1 l
293 1 way as those "one-off." We have been looking at what would rm ( ) 2 happen if you added some design features or changed DBS 3 performance, and the caveat on all this is that you have a 4 good enough model to really know the behavior. 5 So we looked at three design options, backfill, 6 drip shield, and ceramic coating. 7 DR. WYMER: Is there any significance in the order 8 you have listed those? 9 MS. DOCKERY: No. No significance at all. This 10 is the order in which they are in the technical basis 11 document. 12 What I am going to show you are the analyses with, 13 first, the drip shield, which also has backfill in it, so it 14 is included in this analysis. The assumption was that you r x. (_,) 15 added a two centimeter alloy-22 shell, basically, and so it 16 degrades the same way, it has the same uncertainty 17 associated with it as the C-22 model. 18 The model steel is still degrading under humid air 19 corrosion, because you have -- basically, have the humid air 20 in contact with the package. But you have got to have holes 21 in your alloy-22 before you can get seepage into the waste 22 package, or before you can start to significantly degrade 23 the underlying waste package. And it is assumed that your 24 waste package is going to fail at the same place as your 25 drip shield did, because that is where you are going to f%, ANN RILEY & ASSOCIATES, LTD. (' ') Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
1 294 1 - allow r,he dripping water. (Y
.r]
2 And if you make'some of these assumptions that we 3- -clearly understand all the synergistic effect= are not
- 4. included in here, it is basically looking at.what'if you had c5 .twice as much C-22,'then you don't get any waste package 6 failures for 100,000 years.
7 So, this design option, if we did a little bit 8 more work on the details of the model and the C-22 corrosion 9 rate, and also in the specific drift environment with the ~ 10 thermal aspects and everything, this might be something that 11 would be useful to look at. 12 The same. thing with ceramic coating with backfill. 13 You are putting a nice little skin on it. Some of the nice 14 spinel based ceramics are tougher than anything I have.ever () -15 heard of. And, basically, you are protecting it from any 16 kind of rockfall by having the backfill in place, so you are 17 not getting the cracking from rocks falling on top of it. 18 And you are not getting the model steel degrading as fast. 19 The only problem is you may get some blisters at a i 20 couple of hundred thousand years, according to the models 21 that we were given, and that is when you start to have the 22 package available for corrosion. So you can see, again, 23 basically, no dose at all. That's a fairly big impact in 24 the first 10,000, 100,000 years, and a million years, you 25 are starting to lose some of the packages because they were t I ANN RILEY & ASSOCIATES, LTD. Court Reporters i 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 i
- c. -. - - - - , --. a n
i 295 t 1 blistering, the ceramic was blistering and it is beginning [A ,) \ 2 to degrade. Again, as good as our model. s l 3 But it does -- what it does show us is that if 4 there is any engineering way that we can defensibly show l 5 that we can keep dripping water off of the waste package, 6 then we definitely buy a lot in our dose rate curves. 7 So, the summary for both Abe's talk and mine, l 8 since they were supposed to be basically one talk, is that 9 we came up with some expected value dose rates from our 10 deterministic model using the mean values, just to show how ! 1 l 11 the overall system behaved. It is about .1 millirem, 12 maximum to 10,000 years, about 30 to a 100,000 and about 200 13 at a million. 14 The range in the dose rates, and this is the l 15 absolute range, in the probabilistic case is zero to 10, 16 zero to a thousand, and .01 to 3,00C at a millicr fears, and 17 those are the complete range. We have in other ,: aments 18 shown the 5th and 95th percentile as perhaps being a more 19 useful measure of the range for uncertainty. 20 I think you probably are convinced now that the 21 most significant factors affecting performance is the 22 degradation of the waste package and the seepage into the 23 drifts. As we are meeting here today, the Unsaturated Zone l 24 Flow and Transport Workshop is happening in Albuquerque, and L 25 there is a group that is meeting on seepage into the drifts i l l a [} sr ANN RILEY & ASSOCIATES, LTD. Court Reporters
- 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 (202) 842-0034
296 1 and looking at how good is our model as we change the () 2 geometry of the drift. As it starts to degrade, how will 3 that" change our seepage model? Will it change our seepage 4 model? Do we need some big differences in how we are 5 handling it? And getting a much better spatial variability 6 understanding on the fractures so we can get a better handle
.7.. on how we might look at where and when seepage is going to 8 -occur.
9 Degradation of the waste package is an upcoming 10 . event. We do know that there is a lot of work that is being 11 done in the project to obtain data on the degradation rate 12 of the materials of interest, and we will be incorporating 13 that.into our models as that becomes available. 14 We know that there is uncertainty in'our models. q (,,e 15 We do tend to keep focusing on how uncertain we are and 16 where the uncertainty is. But, on the other hand, I don't 17 think we in the PA organization would sit back and we think 18 that our models are just the most horrible, awful things and " 19 they need enormous amounts of work. 20 We think we have come a long way in our overall 21 modeling capability, and that we have finally reached the 22 point that we can focus in on some very specific elements to 23 be investigated. And I think that is one of the real 24 benefits of,-and one of the positive things that came out of 25 the TSPAs we have just run, if that we have a much better i ? j ANN RILEY & ASSOCIATES, LTD.
~~ j Court Reporters l
1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l l-
__ _ _._.._..-._-_____._._._-_.._.._.._._._._.__m ._. . _. 297 . 1 handle on the specifics, specific factors we need to look
- l[) 2 et.
3: We'do think that some of this uncertainty can be 4- ' addressed and performance improved by having perhaps - 5 additional' design enhancements. And we also are conducting,
~6- and think we need to conduct additional site and design
-7 related scientific investigations to address some of these 8 more specific questions that came up as a result of this 9 analysis. And when Ernie gives his talk, he will talk in 10 more detail about.how these have been prioritized and what 11 those specific analyses are.
12 DR. HORNBERGER: Thanks very much, . Holly. We have
-13 time for questions and we can recall Abe as well if we have
< 14- questions for hin.. t O (ss/ 15 MS. DOCKERY: Recall Abe to the stand. 16 DR. HORNBERGER: John, do you have questions? , 17 CHAIRMAN GARRICK: Well, I think part of my 18 questions are going to come in the next presentation. 19: MS. DOCKERY: You~can hope. Keep hoping, right? 20 CHAIRMAN GARRICK: I think that -- 21 DR. FAIRHURST: You can only use that excuse once. 22 [ Laughter.] 23 CHAIRMAN GARRICK.: I think that one of the things l '24 that, of course, the Committee is very interested is what i 25 all this means in terms of design of the repository. You i () ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l . .~ . .. . - . _ _ , .. . . - _ _ , _ . . _ - - - _ . - _ , _ - _
- - .~ . .
i 298 1 note in your summary and conclusions that the uncertainty (v; 2 can be addressed and performance improved by adding l 3 additional design enhancements. And I guess one question 4 is, what are some of those, and how much analysis is 5 supporting that? 6 MS. DOCKERY: In terms of the design enhancements, 7 the two that we are looking at -- step back. I am sure you 8 are very well aware of the design selection work that is 9 going on right now. There is a huge suite of possible 10 design enhancements, options, alternatives being 11 investigated, and models for those are being run through the 12 TSPA analysis that you saw here, the unnamed analysis. 13 But the decisions that we make are only as good as 14 the models that support those enhancements. So I think the f-x l i i,) s 15 next step will be is if we identify enhancements, as we 1 16 looked at here, the drip shield, the ceramic coatings, then 17 we will have to develop better models to capture those and 18 run them in a more rigorous method. 19 What we are going to have from our design 20 selection I believe is some pointers on areas in which we 21 might most productively spend our dollars on design 22 enhancements. So that information should be coming back to 23 the performance assessment team and the project as a whole 24 in the next several months. We are starting to hear some of 25 the results and see some of the results of the design l l ANN RILEY & ASSOCIATES, LTD.
' '(N') Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 i
I. 299 1 selection. Lrx () 2 CHAIRMAN GARRICK: On the business of looking at 3 sensitivities and importance measurements, we had a meeting 4 some time ago in San Antonio where this was a tuajor topic of 5 discussion. And we talked about the basic modeling 6 structures that might enhance the decomposition of the 7 results into importance ranking of specific contributors. 8 I get the sense that the basic structure of the l 9 model sort of came first and then later, when these kind of 10 questions came up, namely, how do you turn up the microscope 11 on the results and pull out the major contributions, that 12 the process of pulling out the major contributors came 13 later. 1 1 A 14 Have you given any consideration to restructuring l !() 15 the assembly of the fundamental model such that it better 16 accommodates decomposition of results into contributors to l 17 risk that don't have some of the limitations that you note 18 in your summary? I i 19 MS, DOCKERY: Well, I guess there's two answers to 20 that. The first one is you are correct that, as we 21 performed the analysis, we did go back and try to find 22 methods in which to look at individual contributors. And so 23 that is -- you know, you are certainly correct on that. And 24 we learned some methods and we stole some methods from the 25 WIP organization to do that. That's why John Helton's name
;[m] ANN RILEY & ASSOCIATES, LTD.
Court Reporters . 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
300 1 will figure very heavily in our uncertainty analysis. We ? rx i i %J
) 2 went back and tried to pull some different ways of looking 3 at things.
1 4 But I think another part of the answer to your 5 question is that the defense-in-depth analysis that Jack 1 6 will be talking about, two talks from now, is kind of 7 putting a method in front of this group to say here is 8 another way, we are thinking of looking at the individual 9 pieces and looking at the specific contributors. And so 10 that's -- to some. extent, I will push that off, and if you 11 want to ask that question again, and say, no, that's not 12 that it, I want something different, then we can address it. 13 But I think that is another method of constructing things 14 differently to try to deal with that issue.
' / \
'Q) 15 CHAIRMAN GARRICK: Okay. I will yield to my 16 colleagues. 17 DR. HORNBERGER: Ray. 18 DR. WYMER: I don't have any. No questions. 19 DR. HORNBERGER: Charles. 20 DR. FAIRHURST: No. It's a very simple 21 observation, but, obviously, if you can stop water coming 22 in, you drop many orders of magnitude on any relationship. 23 MS. DOCKERY: It's water, water, water. 24 DR. FAIRHUPST: It's so obvious, it almost 25 shouldn't be said. But it still strikes me that that is an [ T ANN RILEY & ASSOCIATES, LTD.
-/ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
m , 301 l l; 1 interesting observation. k) 2 MS.;DOCKERY: I remember standing up in front of
'3 .the entire M&O at a TSPA 1993, and they said what are the 4 topIthree most important parameters, and we said percolation 5 flux, percolation flux, and percolation flux. And it just I- -6 hasn't changed, so at least we are getting some robust 7 answers here.
8 DR. HORNBERGER: Marty. 9 DR. STEINDLER: Let's see, two things. One, to ) l 10 follow-up on John's point,-the NRC staff has done some quite ' i 11- interesting work with regard to importance measures, and I J 12 sure that you-learn about.these things in technical 13 exchanges and whatnot, but you might keep your eye on that. l 1 14L .The second question I still have after all of ( ; 15 .these presentations, and' reading lots and lots of material, 16 is that my. understanding is that when we got to this TSPA in 17 '98 -- yeah,'we are still going to get there -- that this 18 was going to be an analysis of a realistic case, you know, L 19 that the word " expected values" here was going to be 20 something beyond just the statistical meaning of the word, 21 -that this was going to be our best guess. 22 And I have this feeling that, and everything I l l .- 23. have heard and read, that your TSPA model still has some 24 bounding assumptions that we are not sure about, and there 25 is sort of.a mixture'of realistic modeling and bounding !~ i O ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 (202) 842-0034
. , ,. . . - ~ . - . - . ~ . - . . - - . - - . - . - _ . - - - . - - . . - -
302 1 _ calculations, and, to not put too fine a point on it, gross
'N
'2.
ignorance in some cases. And I am still a little, I don't 3 know,~ skeptical about then what is that we -- how should we 4 read these dose curves and the sensitivity analyses you have 5 given us? Are these going to change massively when we get 6 to TSPA '99 or '00 or '01, whatever it is? 7 MS. DOCKERY: I would guess, and I will 8 editorialize and then Abe can, you know, give me the 9 political correct and true answer. Given that the next TSPA 10 right now is scheduled to be out, delivered to the Secretary 11 in April of-'00, I do not believe we will see massive 12 changes in how we are handling a lot of the real unknowns. 13 And, again, Ernie will talk about -- there are some cases 14 where we have uncertainty, but we don't really think we can () 15- do much to change that uncertainty. Or a bounding 16 calculation like, how much water gets into the waste 17- package? We don't really have a good mechanistic way to go 18 back and say this is a better way to look at it, so we are 19 going to have to continue to handle it conservatively. 20 So I don't expect that you are going to see any 21 massive changes in the doses. The biggest changes could 22 come if there is a major design change that strongly 23 influences how we model the near field. As we grid up the l-24 repository in a much more finer scale and get a better 25 handle on seepage, it may be because that is such a drive, l i ANN RILEY & ASSOCIATES, LTD. 3 [\ -) Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 L i
303 1 that we may see soma -- it looks potentially like reduction
- [m, s ,)
t 2 in that area. 3 And then, of course, the last is the saturated 4 zone. We really feel that we have a lot to gain perhaps by 5 the connection between the UZ and the SZ, that we perhaps 6 are not taking advantage of, and then the dilution factor in 7 the saturated zone and the alluvium. If Nye County wells 8 become available and we have more information between the 5 9 and 20 kilometer range that we don't have now, I think we 10 may also see some changes. 11 I don't know if I can presuppose what that is 12 going to be, but those are kind of the areas where I see i 13 potentially the biggest changes for a 10,000 year standard. 14 Things like secondary alteration of neptunium and things
/,')\
x- 15 like that, that is not going to -- that is going to affect 16 things at a much longer timeframe. 17 DR. HORNBERGER: Okay. l 18 MS. DOCKERY: Abe, did you want to -- 19 MR. VAN LUIK: I've never been accused of 20 political correctness, but I have no disagreement with what 21 you said. 22 I would like to address part of John's question 23 about the models. I 24 One of the reasons that we selected the RIP 25 platform is because it had built-in features for doing (A 3 ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 l l (202) 842-0034 l i
. .._.m. . - . . _ . . _ . . . _ ~ . . _ _ . - , _ _ . _ _ _ _ _ _ _ . _ _ . _ . - . . . _ . . _ . _ . _ . . . . _ . 't l-.
l 304 l li sensitivity and uncertainty analyses. () 2 -What we have found through experience.is thac they , 3 are' insufficient, that when you have' parameters that are ! -4 'very uncertain mixed with param te ers that are not all that l-5 uncertain you kind of miss the boat, and that is why we are j- 6 . doing'a' lot of.the run-off studies. 7- As far as the changes, what the two-letter word 8 that I wasn't supposed to use,.which will come out later 9 ,this week or early next week, what the purpose of.that was , 10 is to give'us a snapshot of where we were and I have claimed !
,11 all along that it was a trial run or a dry run for the L.A.
1:2 and by golly it has taught us'an awful lot about how to do-
. 13' ' these things, what needs to be done, what sensitivities need 14 to be. looked at, and it has also taught us a lot about what .. /
I - 15 we are uncertain about, i 16 .Now we have the effort to come up with a. design 17 for the_S.R. and the L.A. that is going on right.now and, 18 you know, there are many motivations behind that. 19 One of the motivations that isn't behind that is 20 the 10,000 results from the analyses that will be shown in 21 the VA when it comes out, which you saw a preview of right
. 22 here.
+ 23- The driver behind that is basically to have a more 24 robust case to go into licensing with, and I think what j 25 Holly said was exactly correct. We expect some changes in N 1 l, [I ANN RILEY & ASSOCIATES, LTD. [V [ 1025 Connecticut Avenue, NW, Suite 1014 Court Reporters 1 Washington, D.C. 20036 (202) 842-0034 i
,,a-, ,,. . . ~ - - - - . .._ , ., , , . _ - -- - - -
l 305 l' the modelling,-especially in the saturated zone. The expect I () 2 to have the.near field environment defined a bit better, so 3 that some of the uncertainty goes away, and we expect that 4 the long-term testing activities at Livermore will 5' contribute to a better model for.both cladding degradation 6 and the Alloy 22 degradation unless of course we go to a I l 7 design where that becomes_a moot point. j 8 The design options that are being looked at are a
'9 sweeping set of design options and we will come up with an
- 10 optimum design out of that effort and analyze it, and it may g 11- . require some changes in the way that we model things, but
- 12 basically I think if the design that comes out of that
[ 13 activity resembles and the thermal loading resembles what we 14 have done here, then you will see specific items within the 15 -modelling structure we have now improved, and I think.the 16 performance ~will improve with time. Otherwise there is no i 17 sense changing designs. 18 Now another thing I wanted to say when you brought 19 that'up, is that we do have a large faction within our l' 20 project that believes that the way that we are modelling the 21 water flow through the mountain is unrealistic and that it 22 is in fact bordering on the unbelievably conservative side 23 of things. 24 We hope that some of the activities that we have 25 -ongoing in the niche tests and other things will give us a I
'() .e ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1025 Connecticut Avenue, NW, Suite 1014-L Washington, D.C. 20036 (202) 842-0034
4 306
- . i
'l - better handle on that,'and~it may be that the drip shield
-f ) 2 and other things-are really guarding against uncertainty 3- rather than drips,_because it may-be.that'the drips are much 4' :less likely than we think, but at the same time we can't go ,
5 -into licensing with a case-that is completely realistic by a 6 majority of the - .you know, there's other people in the 7L project that believe we are right about on target as far as 8- -realism and dripping so it is a very dynamic project.
-9 We are still working on a lot of' issues, but if we 10 take conservative view in most case and a realistic view
.11 in some other cases, and this is the hodgepodge you were 12 . talking about, the analysis that you have shown is about 13- what we come out with, so the VA exercise revealed to us a 14' lot of the holes in our own analyses and our own knowledge, P.
[ Ts ,) 15 as you hinted, and we are counting on Ernie in his 16: presentation-to tell us how we are going to fix some of 17 those holes. 18 CHAIRMAN GARRICK: One of the things that we have 19 to be careful about is the language of this business, s 20 particularly when it comes to probabilistic analysis. i 21' If you are doing a probabilistic seepage analysis,
- 22 you shouldn't get into this dilemma that Abe just described l
[ 23. of being accused of being overconservative, if you have l l 24 confidence in the evidence that you are using as the basis t i l 25 for your probabilistic analysis of water, water flow, so I p-. i ANN RILEY & ASSOCIATES, LTD. L Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 L h
l 307 1- don't quite understand that -- that debate. (f 2 -That is the reason you do probabilistic' analysis l3 is to embrace any evidence that-might be available into your 4- probabilistic calculations, so it shouldn't be a question of 5 conservative or non-conservative. It is based on the 6 evidence that is available and that is what the evidence 7 translates into in terms of your probabilistic calculations 8 as far as seepage is concerned. That is one point. 9 Another. observation that I think is interesting is ' 10 that.if in fact we were licensing this facility on the basis 11 of peak dose, and we are not, but in a sense the peak dose 12 is the real world, then your four key attributes really. 13 reduce to two because -- namely, the limiting of the water 14 and the integrity of the waste package, so that is one thing j 15 that I think has been made very clear this morning with your 16 results, because the other is just a matter of' time. I 17 If you are operating on a peak dose philosophy,
'18 whatever the time is that the peak dose occurs would be the l
19 point of the regulatory requirement if that is the direction 20 .we are going. 21 So I think that there's only so much you can do, 22 it; seems to me, with respect to that, but I think one of the 23 things we are really anxious to see translated out of all of 24 -this is, okay, what does all this mean in terms of 25 alternatives that are under consideration for dealing with l-a- ANN RILEY & ASSOCIATES, LTD. i w Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington,'D.C. 20036 (202) 842-0034 t
_ . _ _ _ _ _ . _ _ . _ _ _ _ _ _ . . _ _ _ . _...._--__ _.___ _ __.--._ . . m W 308 1- such~ things as the peak dose, not only in terms of its time
/~N
'of occurrence,.which is a major contributor to its large 2 1 3 uncertainty, but in terms of its magnitude, and maybe that .1 4 is setting up the next speaker.
5- MS. DOCKERY: I think1to some extent it is, but 6 one of the things I wanted to address about the first point 7- about the conservatism -- of course, because we are using a 8' . distribution and it is not a uniform distribution, where you 1 9 place the median value or the expected value in that ' 10 distribution is where it arises that you are either being l 3 11 accused to being too conservative or not conservative l I 12 enough -- l 13 CRAIRMAN GARRICK: Yes. My philosophy on ! 14 probability is different from that. l.( ) 11 5 .My philosophy on probability is the mean value is 11 6 not something that drives the distribution, but the 17 distribution determines the mean value, number one, and 18 number two, the probability distribution has to be justified 19 on the basis of the evidence and if the people who are 20 accusing you of being ultra-conservative have some evidence, 21- let them bring it forward and put it into your analysis and 22 see what it does. 23 MS. DOCKERY: That's -- as is hard to do whenever l l1 24 you are doing snapshots in time. Sometimes the data comes ! -25 in after your analyses are complete and I am aware of what i [,'s) ANN RILEY & ASSOCIATES, LTD. Court Reporters ! 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 h l l -.-, - . _ , . . ,.
309
.1 you are referring to, and we are discussing that.
() .2 I think there was perhaps a misperception-in the 3 team on one of the ways.that things were' handled and so that t may just be what -- if we discuss it with them they'll say, 4 5 oh, we didn't realize that, oh, that doesn't apply --'you I 6 are not too conservative. 7 But there is a range of knowledge. We are trying 8 .to.do the best we can to capture the most reasouable model, 9 and in every case we will try to do that and there will i '10 always be outlier opinions, I expect. We will simply try to 11 come up with the best case given the information that we 12 have and we will continue to evolve toward that. 13- As long as they keep doing studies after we have , f r 14 finished our analyses, we are going to be subject to saying i 15 oh, you missed that one, you shouldn't have done it that 16 way. 17 CHAIRMAN GARRICK: One. thing I did want to say, 18 very positive. I think that today you have done a wonderful-19 job of showing how these results can be presented and how 20 you.can cut them across just about any angle and how you can 21 turn up the microscope on them, and the issue that remains 22 is the transparency of the results and the supporting 23 evidence. 24 MS. DOCKERY: One other thing I wanted to mention-25 is in the second point you were making about the two l l t ANN RILEY & ASSOCIATES, LTD. l Court Reporters l 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202).842-0034 i . . . .- _, _ -
- . - -._. . -. .. - - . . . . - _ - . ~ . . .-.. . - , ._ . - . ,
l 310 1 factors. Of course we have to keep'in the back of our mind l - (/) ' that as we. decrease that uncertainty -- right now they are 2 3 showing up as.very important in part because of the 4 uncertainty we have associated with them. 5 As we can model them better, and because our 6- answers are only as good as the models that support them, l 7 there may be other elements that also show to be maybe not 8 equally important but important,.and the saturated zone is 9 certainly one that comes to my mind as an area where we 10 really want to hold judgment until we feel like we have got 11 a better representation. 12 On the other hand, I don't expect that we will 13 find that water contacting the waste package is ever an l l 14 unimportant phenomenon. I think Abe had wanted to aay I 15 something. 16 MR. VAN LUIK: Yes. I wanted to respond, except 17 you stole my first response, but on the second part, in the 18 re-analysis of the design, the peak dose reduction is one of 19 the motivating factors. It is.not that we are jIst totally l 20 focused on 10,000 years, although for licensing car focusing 21 is shifting now to make that a very robust case. l 22 On the other parts, I think no one will accuse us 23- right now that the distributions that we le using for l 24 seepage, for example, don't represent the whole range of f L 25 opinion in the project, but we have very few datapoints and l l() ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 j Washington, D.C. 20036 (202) 842-0034 i
. . _ . , __ . , .,4
_ . . _ _ _ . _ _. . . ~ . - . - - - . . _ _ _ . . ..- __ _ . - _. . _ . . 311 1 a large range of opinion, and the charge is that if you j ) 1 2 would=get some more datapoints you would narrow that i 3 'uncert'ainty,-because, you know, the basis right now is very ) 4 loose, so that is why I am hoping that we will fix it at one 1 5 point or another with.less uncertainty, either at the high 6 level or the low level. I don't know who is right in this 7 debate. l 8 CHAIRMAN GARRICK: Yes, and I think we have to 9 remember there is a difference between opinions and 10 evidence. 11 MR. VAN LUIK: Oh, yes, except that in performance 12 assessment, as you well know, it is a mixture of expert 13 judgment, interpreting data, incorporating it into models, 14 et cetera, et cetera. We shouldn't overplay that, but at
.(_)
15 the same time there is a lot of expert judgment in every 16 piece of an analysis. 17 MS. DOCKERY: An excellent conclusion. I would
'18 have to say Abe is right. We learned more than we ever 19 dreamed we would learn by going through this process, so I y 20 think we are in a much-improved state, both modelling-wise j 21 in terms'of the data, but also in our understanding of how 22- to run TSPAs, and we are stealing liberally from our friends 23 at the WIPP whenever we can to do a better job at presenting t
24 the results. i ! 25 DR. HORNBERGER: Thank you very much, Holly, and [ l ANN RILEY & ASSOCIATES, LTD.
\#
d Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 L ' Washington, D.C. 20036 (202) 842-0034 1
. _ . . . .- ._. . _ . . _ . _ . _ - . .. . -- ._ . _ . - _ . . . _ _ m .
312 il "thank you,-Abe'.
/~y
() 2! We do have to break'now. We have some commitments at 12:15. We will reconvene'at 1:15. 4 -[Whereupon, at 12:15 p.m., the meeting was
, 5 . recessed, tcf reconvene at 1:15 p.m., this same day.]
t 6 7
'8 9' :
10 11
.12 l ' 13 - 1 L s y .
14 ; 15 l 16
'17, 18- 1 1
r- - l l 19 l 20 21 l ( 22 l-23 i 24 25. 1 !
~[)
' ^ '
ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
i l 313
\
l l i l' AFTERNOON SESSION ! ( 2 (1:18 p.m. ] 3 CHAIRMAN GARRICK: The meeting comes to order.
'4 George, f l 5 DR. HORNBERGER: Thank you, John. We are 6 reconvened again. We have two presentations that will wrap 7 up our discussion on TSPA for today.
8 The next one is on prioritization, and Ernie 9 Hardin is going to do the presentation. l 10 MR. HARDIN: Thank you. My job at the M&O is 11~ the -- I'm the lead for the-group that is doing models for 12 the near-field environment, and after this morning I l 13 understand why'they' asked me to come here and give this 14 talk. 15 LLaughter.] 16 I'm going to talk about prioritization of 17 technical work needed in the next 18 months to two years to 18 complete the postclosure safety case. 19 We start with the postclosure safety case elements l 20 that you've already encountered this morning. And I'm going 21 to address only the first, which is expected postclosure l 22 performance and the supporting evidence for it. And this 23 grounds the process of prioritization that I'm about to 24 describe. 6 25 And these are the attributes of the repository I
, ANN RILEY & ASSOCIATES, LTD.
Court Reporters I -' 11025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 L L -
. - . . _ . . _ . . . . _ . _ , _ _ _ _ _ . ~ _ _ . . ._ . _ _ . . _ . - . . _ _ . _ . . .
314 1- safety strategy,.postclosure,. and the 19 principal factors I 12! that we talked about, which have been liberally discussed. i
.3 The approach ist to prioritize the factors, and 4~ thereby.to prioritize the technical work to increase.our i
5 understanding in these areas._ And it's important to point s. 0 , 6 out that-these factors may change, _ depending'on the design, 7 or for that matter depending on the safety strategy. And if
~
8- that and when that happens, we're at liberty to repose the
- 9. prioritization.
10 DR. FAIRHURST: Ernie, could I just ask.you -- 11 MR. HARDIN: Sure. 12' DR.'FAIRHURST: The current listing, 1 through 19, 13 is that -- that 's not :ba the priority listing now, is it, or , '14 in it? It looks as though it could be. Is that your k 15 ' current sort of 1 through 19 priorities? 16 MR. HARDIN: Yes, it is. These are the principal 17 factors, not in order of priority. 11 8 MS. DOCKERY: This is in order of factor, not in 19 order of priority. 20 MR. HARDIN: And I'll present to you a process 21 that we used to apply TSPA and expert judgment to come up 22 with a-system for prioritizing where the emphasis in 23- technical work should be applied across these 19 factors. i l , 24 DR. FAIRHURST: Okay. 25 MR. HARDIN: I'd like to take just a minute and i
- ! b. ANN RILEY & ASSOCIATES, LTD.
- ># ' Court Reporters
'1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 lY l' . -. - - . ,
l 1 315 1 talk about technical work plans. We have planning at two ; rx } 2 l (u/ levels in the project. We have the plans that are in the j 3 VA, and then we have our program plans and schedules. 4 The work that I'm talking about is described at 5 both levels, and is that which will support evaluation of 6 postclosure performance. 7 In addition there is other technical work in the 8 plans, and, for example, Jack this morning talked about 9 defense in depth. There will be subsystem performance 1 10 analyses conducted in that area. Consideration of 11 disruptive events, the events, igneous activities, seismic
.12 activity, human intrusion, and nuclear criticality will be 13 analyzed, but there are no new data-collection activities 14 proposed at this time other than continuing to monitor the
(_j 15 seismic network and geodetic networks. 16 For natural and manmade analogs there is technical 17 work in the current plan. I'm not going to discuss it at 18 length. In addition to the international projects that DOE 19 has supported in the past and the efforts at Laurence 20 Berkeley and Laurence Livermore National Labs in the area of 21 looking at natural analogs, namely the geysers and the 22 Waireke geothermal field in New Zealand, there are other 23 activities currently contemplated, and one of those is Pina 24 Blanca. We are currently discussing how best to approach 25 that site. It's of interest. l ? ANN RILEY & ASSOCIATES, LTD. (m\') Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 i
._._.y i
l-l '316
- l' 'In'the area of performance confirmation, of
() .2 3-
. course, the regulations dictate that the Department will provide a plan =. There is little technical work. There is 4 little technical work involved with constituting that plan,
! 5 but it of. course will. describe technical work to proceed 6 beyond the' time frame that I'm-going to discuss in today's 7 presentation'. l 8 And over and above that, there is additional l 9 technical work planned for.preclosure safety, and also l 10 technical work to support design-decisions. And in fact i 11 those -- in the area of supporting design decisions there is I 12 quite a bitlof technical work, but again I'm not going to.
- 13. . touch on that directly today.
14 What are the objectives for this prioritization
. task?-
15 16- Well, we're going to determine in a systematic
.17 fashion what technical work is needed to support the SR and 18' the LA. We're going to identify specifically how the work 19 will contribute to the PA. And we're going to base this for
;20 the present on the reference design with the possibility 21 that we may revisit the prioritization in the future. The 22 results of the prioritization, by the way, have been 23 implemented in the project planning at all levels.
24 Some other considerations taken into account in 25 prioritizing technical work, operational, policy, and l' i ( I '- ANN RILEY & ASSOCIATES, LTD. ' j Court Reporters I ) 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 (202) 842-0034 L
1 i 317 t 1 environment, safety, and health, some examples, we have to g
) 2 consider budget. In the policy area we have, for example, 3 the Department has elected not to use or apply for 4 permission to use radioactive tracers at the site. And in i
5 the environmental, safety, and health area we control site
)
6 access and site impacts. ! l 7 How was this activity conducted? We convened a 8 multidisciplinary, multiorganizational team in Las Vegas. 9 The principal investigators from the national laboratories j l 10 and the USGS were included. A consultant facilitator was ! 11 also included for control of the methodology and the 12 assessments. And we generated a formal record and have i 13 produced a report. ! 14 The methodology that we used is straightforward. (~ ( 15 A priority figure of merit is based on the algebraic 16 difference between an assessed confidence goal and an 17 assessed measure of current confidence. 18 We started with assessments of the significance of 19 uncertainty to TSPA, and again this is grounding. We 20 received the information that Dr. Dockery alluded to this 21 morning based on exercising the PA model, and I'll talk 22 about that a little bit more. The current confidence and 23 confidence goal were assessed on a numeric scale; 1 is low, 24 7 is high. And what I'm going to do is I'm going to go 25 through each one of these assessment categories briefly and O ANN RILEY & ASSOCIATES, LTD. ( >) Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
l' l 318 l l 1 describe how it was done.
' [ V) l 2 The TSPA, the current TSPA model was exercised 3 over the ranges of uncertainty implicit in that model. So, l 4 for example, this means that if a particular parameter which 5 captures the behavior of a subsystem is known to be 6 distributed over a range or is assumed to be distributed l 7 over a range, it was that range that was used, and a 8 comparison then was made between behavior projected for the l 9 inputs of that range. And whether the peak dose rate l l
10 difference then from doing that was greater than fivefold or ' 11 greater than 50-fold determined whether we assessed this as 12 low, moderate, or high. l I 13 In addition, using the PA model, it was possible 14 to specify when during the postclosure time frame that peak j (~'N ( ,) 15 dose rate occurred. i 16 DR. STEINDLER: Could I ask a question? 17 MR. HARDIN: Of course. 18 DR. STEINDLER: We heard this morning that the 19 results of the TSPA model exercise is iteratively used to 20 make design changes, which in turn give you different output 21 from TSPA. You know, that's obviously an iterative process. 22 Where did you break into that iteration in order to finally 23 get a TSPA model to exercise to do this, and how does that 24 relate to what design changes you might yet make and then 25 thereby change the prioritization? i i [A/ ) ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
._.___..._.._.._____..._.__.-.____._.___._____..__._.....m.. . _.
319 1 MR. HARDIN: The models that were used were those I l( 2 that were available in the June-July time frame.of this L 3 . year,.which corresponded with completion of the draft of the L 4 VA. The analyses were done with respect ~to a reference I 5- design which has been developed over a year or two and was !~ 6 available at that time and corresponded'to the VA reference 7 design. L > l 8 If the design changes, as it may as a result of. L l 9 ~ the ongoing design selection exercise.boing conducted by l j E10 DOE, then it may be advisable to go back and reprioritize Lil the technical work. 12 I'm hedging a little bit there because the design 13 may not change in ways that are different that demand that-14 we go back and reprioritize. That is not clear yet. We're r (f 15 only in the first' phase, i 16 DR. STEINDLER: It may or may not. ; 1 17 MR. HARDIN: It may or may not.
~18 This is the set of. assessments made on 19 . significance-of uncertainty to TSPA. There was quite a bit -
20 of interest in this earlier. 21 This is in your package, of course. For the most 22" part'-- let'me just quickly describe what the subscripts 23 represent, as noted at the bottom. Where subscripted A, 24 'this is where the peak does rate in the analysis occurred in
- . 25 .the first 10,000 years, B was 10 to 10,000 and C was 100,000 l-i
b[ l ANN RILEY & ASSOCIATES, LTD. 3- # Court Reporters
- 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 70036 (202) 842-0034 i
)
320'
.1 to a million. And if there is no subscript, it means the r '2 peak dose. rate effect was observed over all those time i
3- periods. 4 And for the'most part, these -- the significance ._ 1 5 assessments are-intuitive, at least to me. They follow the-6 logic of the PA model, and I could take a.few of these as t 7 examples. . Seepage into drifts'has a high significance of' 8 . uncertainty over all time periods, and that is because in 9 early time periods it contributes to degradation of the 10- waste package in the model, and in late time periods it 11 contributes to dissolution of the fuel. 12 DR. HORNBERGER: What does number 6 mean? Do that 13 as an example for me. 14 MR. HARDIN: Okay. Humidity and temperature at 15 the waste package. That's a good example because it is not l 16 quite as intuitive there. The humidity and temperature are
- 17 performance measures that influence waste package lifetime.
18 However, dripping.is a measure that influences lifetime l 19 'more. The-ultimate time to failure of a package is more 20 sensitive to dripping, or to the existence.of dripping l l 21- conditions at the location of t'ae package. 22 DR. HORNBERGER: What are there subscripts B and l 23 C? I L 24 MR. HARDIN: The waste package is intact at early 25 time. i i ANN RILEY & ASSOCIATES, LTD. Court Reporters i- 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
321 1 DR. HORNBERGER: Oh, Okay.
} 2 MR. HARDIN: Okay. Yeah, that's -- another one 3' that may not be instantly obvious is number 11, integrity of 4 the' cladding. And in this case the evaluation was done by
- 5. turning off the cladding in the model.
16 DR. HORNBERGER: Yes. [ 7 .MR. HARDIN: Current confidence,.an expression of i .
- j. , 8 :the degree of certainty that the current representation of l
O, the principal factor for PA is realistic and captures the l 10 entire range of conditions important to performance. 11 There's a bit of judgment there. 12 This is -- in assessing this, we recognize the 13 limitations of our models.
-14 DR. FAIRHURST: So current confidence with k 15 humidity and temperature means that you are better than you 16' need to be?
-17 MR. HARDIN: That's right. That's right.
18 Clearly, this current confidence is a relative assessment, L19 it is based on a critique of the current models. And in
-20 .every case, the potential for improvement is recognized 21 because we didn't score any of these a 7.
22 I am going to discuss several of the key principal
=23 factors, those which received high priority in this 24 exercise, and we will go through the current confidence and
- .. 25 other assessments in more detail.
i-f I l-ANN RILEY & ASSOCIATES, LTD. _ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 g*7++ > w %m yi-.-W'r *
- P e 4, --
++-myc- F=vf a-----'
. - - -. .._ - ... .,.- . - - . - - - - . .-. - ~ . - - . . _ _ .~. .
322 1 The confidence goal.was to find that level of I,,gV) 2 confidence that is both feasible and desirable at the time 3 of'LA. Feasible means that we'can do this by early_in the 4 year 2000. Desirable means that the principal factor is 5 significant to TSPA and is important to defensibility of the' 6 models. 7- Again, the goals need not be high. We need not 8 score a 7 on every factor-for the intended use of this v I
% i 9 information in a performance allocation.
10 So here are the confidence goal assessments. For 11 all the factors but one, and that was number 6, it is both 12 feasible and desirable to show some improvement in 13 confidence between now and preparation of the LA. And from ! 14 here we -- r (%,) 15 DR. HORNBERGER: For goal 6, you want to get 1 16 worse, right? l 17 MR. HARDIN: Yes. Well, -- yeah. We would like
- 18. to learn less about 6. We identified that irony 19 immediately, and we decided that really has the same meaning 20 as a zero.
21 These are the results of the prioritization. The 22 most highly prioritized have a 3, and those are seepage into l .23 drifts, the effects of -- I'm sorry, the integrity of the 12 4 . CRM, and transport through the UZ. 25 The ones that I have enbolded in this list are the ANN RILEY & ASSOCIATES, LTD.
'v Court Reporters l
1025 Connecticut Avenue, NW, Suite 1014 L Washington, D.C. 20036 l (202) 842-0034 L
323 l 1 ones that I would like to discuss in a little bit more f 2 detail. b/ i And I have some extra materials where I can answer 3 questions on other factors as well. ! 4 For seepage into drifts, the significance of 5 uncertainty is high,_ as we talked about earlier. To put 6 this in perspective, we have a seepage model which is based 7 on a porous medium, approximation to the behavior of 8 fractured rock at one extreme. At the other extreme we have 9 a model in which the percolation flux is intercepted by the 10 cross-sectional area of the drift. And we are collecting 11 information, even as we speak, that will help us to pin down 12 where in that range is the best defensible position 'or us 13 to take in PA. 14 The current confidence was assessed to be low. A
/~%
( ,) 15 couple of the reasons for that are that we are basing this i 16 on -- our seepage model is based on a cylindrical drift { 17 geometry and it is thought that when we go to an irregular ! 18 geometry, that the seepage analysis will change. In 19 addition, we don't take into account film flow effects, 20 which are capable of diverting flux, and a substantial flux 21 compared to some of the numbers that we are working with 22 here, diverting it around the waste package so that it does 23 not become a drip. 24 The current model is calibrated against few data. 25 We have a number of different field tests in the ESF wnich l l l'~ ANN RILEY & ASSOCIATES, LTD.
\- Court Reporters
! 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
_ . - - . . . --- - - . - . . ~ . - - . - = - - - - - - . - .. 324 l 1 help to inform this model. They are all, however, being -- i-2- .have been conducted in the TSW 34, which is the middle [V~)
;3 nonlithophysal tuff. Much of the repository will be in the 4 lower lithophysal tuff, so.that that is our next objective, and we do have some planned or. ongoing field tests in that 5
6 unit that will help broaden the base for this model. 7 DR. HORNBERGER: These are, again, water injection 8 tests, is that right? 9 MR. HARDIN: Yeah. Niche studies. There's an 10 infiltration experiment in the lower lithophysal tuff, and i 11 there was an experiment, which I have not seen the analysis 12 of, done at the point where the cross-drift passed over the 13 main drift. In addition, there is some injection work and 14 transport work planned at that location. Seepage has a high rN ( ,) 15 priority and will be continued to'be supported from now till 16 LA. 17 The integrity of the CRM material also has a high 18 significance of uncertainty to TSPA. It is important to 19 recognize that the current PA model is based on an expert 20 elicitatiin of the corrosion rates for this material, 21 alloy-22. The experts have taken a sort of composite view 22 of the variety of different corrosion test data available 23 for this material, and so they have helped us to deal with 24 uncertainty in the near field environment conditions by 25 forming their assessments using data from corrosion tests L' [ ANN RILEY & ASSOCIATES, LTD. Court Reporters ! 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
F: , 325 l l 1 spanning.a wide range of conditions. F We think that that range of conditions is perhaps
~
2 3 too large to be realistic, so,'in some respects, the model ) 4 may be too conservative. HWe have'more recent laboratory I 5 corrosion test data on alloy-22 that suggests that the rate L 6 of. general corrosion is lower than the rate that is used in. l 7 this model, on.the' order of.a-tenth of a micron per year. 8 On the other: hand, there are some CRM failure l
.9 -modes, such as the oxide wedging hypothesis, that are not i
10 factored into this model. ? I l 11. So, our current confidence assessment is I 12' moderately low, but we are looking for a real improvement in i 13' this area as we accumulate laboratory test data and as we i 14 develop predictive models that allow us to extrapolate those (I ,15 laboratory data to the postclosure performance. So these l 16 would be mechanistic models, and Joe Farmer at Livermore, l 1
-17 for example, has been working out models for passivation. l L 18 And, finally, we are going to see improvement in
-19 the specification of the waste package environment. And 20 here we mean the envelope of hydraulic, thermal and chemical
-21 conditions in which the waste package needs to function.
22 And as we resolve some of the details in the LA reference 23 design, it will become possible for.us to define this 24 envelope. 25- For transport through the unsaturated zone, the i I l I-ANN RILEY & ASSOCIATES, LTD. , Court Reporters i 1025 Connecticut Avenue, NW, Suite 1014 Y Washington, D.C. 20036 (202) 842-0034
326 1 significance of uncertainty to PA is high, and this is based j 2
) on an analysis that considers both the UZ and SZ. Let me 3 try to. explain this.
4 The UZ and SZ dispersion assumptions are linked, 5 since the UZ acts as the source for contaminants at the-6 water table. If UZ dispersion is minimized in the model and 7 if the saturated zone dispersion is minimized, you get to 8 one extreme, and if some measure of dispersion is attributed 9 to both of those natural barriers then you get to the other 10 extreme, and the range is a thousand-fold, so we are very 11 much interested in resolving questions of uncertainty of
=12 transport through the UZ.
13 Current confidence'is assessed to be low. The
'14 mixing and dilution' assumptions that are used in the current
() 15- model may actually be nonconservative and we don't take into j E16 account coupled process effects directly or even indirectly 17 really in the transport model. 18~ Some of our reference design thermal calculations 19- -show that the zone of thermal alteration extends to the top
'20. of the Calico Hills and beyond, so we need to look at 21 coupled processes.
l 22 Our confidence goal is moderate. We have ongoing 23 field tests that will be looking at the chemical transport, 24 both at Yucca Mountain and at Busted Butte. The tests at l 25 Busted Butte will look at the transport of chemical tracers [ ANN RILEY & ASSOCIATES, LTD. 4 Court Reporters
- . 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036
) (202) 842-0034 1
l I ! 327 l 1 and colloidal tracers in the upper part of the Calico Hills (m) 2 unit and tests in the ESF facility will be looking at 3 chemical transport of conservative and nonconservative 4 tracers with a view to specifying fracture-matrix 5 interaction and the potential for dilution on the transport 6 pathways in the UZ, so this also has a high priority for 7 technical work. 8 For the effects of heat and excavation on flow, 9 the significance of uncertainty to the PA is moderate and 10 this assessment is based primarily on our uncertainty in the 11 properties that go into the thermal hydrologic models that 12 are used to specify the environment. 13 In addition -- and of course, as I stated before, 14 we have field thermal tests in the middle nonlithophysal (~N
\ ,) 15 test, and those have been very helpful in building 16 confidence in the thermal hydrologic models, so helpful in 17 fact that we would like to run similar such tests in the 18 lower lithophysal tuft in order to build our confidence 19 more.
20 So the current confidence is assessed to be low 21 and one of the other reasons is that we have not included 22 thermal hydrochemical or THC effects in the current model. 23 We have hopes to increase our confidence to moderate and we 24 are going to do that by carrying out the ongoing and planned 25 field tests in the lower lithophysal unit, and we are going l c '-(<
\'
~) ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
l 328 I 1 to advance the state of our THC models.
,m
( ) 2 We rely on models to predict the outcome of N- /
)
3 processes taking place over hundreds or thousands of years. 4 The thermal tests are, while useful, need to be augmented by 5 laboratory testing and so there is an active program 6 underway to build up a database upon which to validate THC I 7 models and the focus of these models and this effort is 8 primarily the rate of flow and the composition of flow into 9 the drift as it will affect the waste package lifetime. 10 For the chemistry of water on the waste packages, i 11 as it turns out Alloy 22 is very resistant material, 12 resistant to corrosion under a very wide range of 13 conditions. 14 The studies that we have done using the PA model ) 7- .
! 15 show low sensitivity of peak dose rate to our uncertainty in 16 chemistry of water, but we think there might be some 17 corrosion processes that we haven't taken into account and 18 for example the accumulation of salts on the package and a 19 micro-environment that could initiate pitting, and so the 20 assessment is done with respect to the current reference 21 design but the implication applies as well to a waste 22 package with CRM on the outside.
23 So current confidence is assessed to be moderately 24 low. We did not take into account introduced material 25 effects except as a sensitivity in the current PA model. ('\- ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 I Washington, D.C. 20036 i (202) 842-0034 i
329 1 In addition, we haven't taken into account coupled ()
~s 2 ,
processes. The coupled processes will tend to, if we get l 3 into a thermal regime where there is a significant amount of l 4 reflux, vapor liquid counterflow in the host rock, we expect 5 that there is a potential for solutes to migrate towards the 6 heat source and this is a matter of some interest and 7 complicates the prediction of water composition entering the 8 drift both during the thermal period and after the thermal 9 period. As water re-enters the system that's dried out it ) 10 can pick up these salts and convey them into the drift, so l l l 11 this is another area where we rely on and will rely on the l 12 predictive models. ! l 13 We have an engineered barrier system prototype ' 14 test program underway which is in its early stages at I (~x l !(_) 15 present, but it is hoped that as the early tests progress to 16 more complex thermal tests with chemical aspects that we 17 will be able to provide more constraint data to-those 18 models. 1 I 19 So our current confidence is moderately low. Our 20 confidence goal is moderate. Through supporting lab tests 21 the predictive models will hopefully give us the ability to l 22 confidently predict water composition in a range that will 23 be used to specify the waste package environment. l 24 For flow and transport in the saturated zone, this l 25 principal factor actually received a lower priority than l ( ANN RILEY & ASSOCIATES, LTD.
\- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l
330 ; 1 some of the others, and I will state from the onset that [ j'v) 6 2 that-is mainly because we recognize limitations on what we 3 are going to be able to accomplish in the time remaining for 4 site characterization before we begin to prepare the license 5 application. The significance of uncertainty to PA is 6 moderate. 7 Based on the conceptual model uncertainty that we 8 have and the way we handled saturated zone flow and 9 transport, we feel that there is somewhere in the five to 10 fifty-fold range of uncertainty in peak dose rate associated 11 with saturated zone models. 12 Our current confidence is assessed to be low and 13 of course this is associated with the flow tube model and 14 our characterization data gap from 5 to 20 kilometers that
/
( 15 Holly spoke of earlier. 16 Our goal is a modest improvement in confidence for 17 SR and LA. We have an ongoing collaboration with Nye County l 1 11 8 that will allow some drilling in this timeframe. We are 19 refining the numerical models for a more productive analysis 20 of dispersion effects. 21 In addition, we are looking at -- we are going to 22 go back and take another look at hydrochemical and isotopic 23 data in the saturated zone to try to constrain the saturated 24 zone model to the extent that we can. 25 We are looking at the conditions that affect l l [~}
\_/
ANN RILEY & ASSOCIATES, LTD. Court Reporters l 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
l 331 1 mobility of Neptunium and also plutonium and technetium. In ! () 2 3 the saturated zone we have some early data which suggests that.we have slightly reducing electrochemical conditions 4 and those indications are being followed through with more
'S data collection and more investigation'in the laboratory of I
- l. 6. neptunium behavior at those conditions in a mixed chemical l 7 system, and in addition since the flow path includes some I i
8 segment of alluvium is that we are going to go back and look I n 9 at the mobility of plutonium, technetium, and neptunium in i 10 alluvial materials. i 11 So in summary, the technical work needed for SR I 12 and LA has been identified using prioritization. Some l 13 things didn't make the cut. We have implemented the 14 prioritization in our program plan. The prioritization j l () 15 considers postclosure performance, also things such as cost l 16 and schedule, and we have applied our judgment in the use of 17 the TSPA model. l 18 We think that prioritization is absolutely, 19 essential to focus the technical part of the program on
'20 needs for SR and LA, and we recognize that some technical 21 work will continue beyond that as part of the performance
-22 confirmation program.
l 23 As a personal observation, this prioritization. 24 exercise is the successor to a couple of other things that 25 have been done since 1989, since the SEP. i
- h
\
ANN RILEY & ASSOCIATES, LTD. Court Reporters . 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l-l
332. 1 This-one differs in a couple ways. It is simpler, j ) 2 it's more direct, ,and'it's been implemented. 3: [ Laughter.] 4' IHl. -HORNBERGER: Thank you, Ernie. I'm sure we l 5- 'have'some questions. Marty, do you want to start?- 1 6 DR. STEINDLER: Yes, I had.a. couple questions. l 7 Ifthought I heard you say that there is a reducing E '8 environment.for neptunium to' operate in. Is'that what you o 9 said? 10 MR. HARDIN: Right. We have' purged-some of the
- '11 older bore holes at the. site and measured the EH using a -
12- compensated electrode and found it in the range of let's say l ,13 zero to 100 millivolts, which is a slightly' reducing 14 condition that favors precipitation, reductive precipitation i 15 of neptunium. 11 6 ' DR. STEINDLER: What does it do to technetium? I l ! 17 MR. HARDIN: I am not sure but I don't think it 18 'does very much. 19- DR.-STEINDLER: And this is still open to the air, ; 1 L '20 however, right? H 21 MR. HARDIN: I am not sure I understand you -- 22L DR. STEINDLER: Don't you have an atmosphere 23 exchange? i
-24 MR. HARDIN: Yes. It think there is some evidence 25' that in the uppermost few meters of the saturated zone is a l-e 5 , ANN RILEY & ASSOCIATES, LTD.
!' Court Reporters 7 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 L \ (202) 842-0034 i c . . - . - -
. . - . ~ . - . ~ - . . - - - . - - . - - . . . . . . .-.~-..
333
.1 zone of water that clearly has a gaseous exchange with a
+'( n) 12 unsaturated zone gas phase, but deeper in the unsaturated , !' ;3 zone. Conditions 1 differ. 1 4 ' Clearly in order.to take performance credit here, 5 we need -- there are a number of hoops that we need to jump
-6 through. We need to understand the observation, look at the l .7 couples, verify it to observations of solid phase --
8 precipitates, . and so on. 9 DR. STEINDLER: That is, I think, an extremely.
,10 important point to check on.
11 The other one I guess I have a. question on, you E 12 ' indicated that'prioritization~seems to be related to cost 13 and schedule. While on the surface, that seems quite f 14 reasonable, if you put yourself in the position of having to l( fq )t 15. defend that decision in front'of a rather nasty Intervenor, 16 standing in front of an Atomic Safety and Licensing Board 17 panel, how are you going to respond when the question is how 18 come.you haven't got this number, which you clearly have 19: indicated you need but you have given it lower priority? L 20 MR. HARDIN: So taking the saturated zone, for
- 21 instance.
22 DR. STEINDLER: Whichever. l - 23 MR. HARDIN: Well, the prioritization that we're L - 24 talking about here is for the purpose of resource 25 allocation, and if the program finds it necessary to i. l'N [ , ANN RILEY & ASSOCIATES, LTD. Court Reporters l 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 ww-,---- ** e-',e e* s +r w . sv % *- iw-- a--*'-vt -s=+ * -e r ,.mev *'n v-- '--r +
334 1 continue studies in any area, the resource allocation can-be ' 2 revised.in the future. We will have a performance 3 confirmation program. I' guess that's my answer. It doesn't ' 4 close the door on technical work ~in any area. 5 DR. STEINDLER: In terms of priority of influence,
'6 apparently. cost and schedule has a higher priority than j 7 technical need. 'Is that an unfair characterization?
L 8 MR. HARDIN: We recognize that'there'are some 1. 9- areas lwithin'the principal factors where further' 10 investigation is possible but may not be feasible. o l _11 DR. HORNBERGER: Charles. I L 12 DR. FAIRHURST: I have several questions. 13 How much of this will you think you can get into l 14 the license application? It looks as though there's a f () 15 tremendous amount of work you want to do, and some of it's 1 16 dependent on getting additional experimental data. You've L 17 got what, a couple of. years? 18 MR '. HARDIN: Right. Well, the first milestone 19 -will be our site recommendation report, and I suspect that 20 some of these new data will not make it into the site 21 recommendation report because of schedule. But the license
- 22 application as drafts of it are prepared and ret' sed in the l 23- following year will contain more and more of it.
I 24- DR. FAIRHURST: The other thing, did I understand
- 25. you to say that you took some bounds on the infiltration a ANN RILEY & ASSOCIATES, LTD.
- Court Reporters l ' 1025 Connecticut Avenue, NW, Suite 1014
- Washington, D.C. 20036 (202) 842-0034 l
H.
"" =;*'M-rY " - + - e
. _ . __ . _ . _ .~ ~ . _. - .._ _ .m _ _ _ - _ . _ . _ _ _ _ . _ . _ _ _ . _ _ . . _
335 i s 1 into.the drift,.and one was to take the percolation flux 1
) 2 across the entire cr'oss-sectional area. , _3 .MR. HARDIN: Yes.
4 DR. FAIRHURST: Now, playing a little the devil's 5- advocate, if you've got a very long time frame, what is your 6 expectation of the extent of collapse above a tunnel? l 7 MR, HARDIN: My notion of that is the difference 8 in void ratio will determine the eventual outline of the
. 9. collapsed structure.
'10 .DR. FAIRHURST: So if you had a 20-percent 1
11 1 - bulking, you could have something that would go about five i 12 . times the height of the tunnel; right? 13 MR. HARDIN: It's possible that absent backfill 14 ~that we could -- I ,r"N i ( ,) 15 DR. FAIRHURST: That's right -- 16 MR. HARDIN: We can develop -- 17 DR. FAIRHURST: That's exactly the point. Right. 18 MR. HARDIN: It's possible, but, of course, as you
- 19. very well know, this is a -- it's a refractory rock that's 20 been stable for a long time and does contain some large 21 voids.
-22 DR. FAIRHURST: P'ght. And -- no, that's a good 23 point, actually. But that is almost a natural analog, 24 right? What's the largest sort of void you can find in 25 there?
ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202)- 842-0034
. . - . . ..___...~ . . ... .- .. - ..- - _..-. ~.... -.
336 1- MR. HARDIN: The scale of tens of centimeters. () 2 DR. FAIRHURST: Right. Right. 3 MR. HARDIN: Which of course is not -- not 4 controlled by fracture behaviors. 5 DR. FAIRHURST: -This is partly --lbecause I don't 6' .know that part.of'it, but I can almost imagine that you'd 7 have -- somebody'could, like a nasty intervenor,'could.ask
~
8 you about a. series of large rubble chimneys or whatever you , 9 want to call them which you're almost focusing into, getting 10 more than -- but maybe that's not feasible. 11 MR.-HARDIN: A rubble chimney might not -- 12 DR. FAIRHURST: No, it's not quite -- 13- MR. HARDIN: The hydrology of it might be 14 favorable. (f 15 DR. FAIRHURST: Well, I don't know. I'm just 16 asking. l 17 lMR. HARDIN: It might be. And that's one of the
- l. 18 things that-they'll be'looking at.
19 DR. FAIRHURST: Then you said that you were going l
'20 to go'into the lower lithophysal and conduct some more sort 21: of thermal mechanical -- thermal --
22 MR. HARDIN: Thermal test. 23 DR. FAIRHURST: Really, some large ones. Is it 24 equivalent to the current -- 25~ MR. HARDIN: No. ANN RILEY & ASSOCIATES, LTD. l Court Reporters - 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 i' i (202) 842-0034
. . . - . . _ _ _ _ . . . _ . . . _,-._m _ . . . _ . . _ . . _ _ _ _ . . . . - _ _ _ . _ _ . _. _ . . . . _ . - - l l 337 l 1 -DR. FAIRHURST: Drift experiment, i 2 MR. HARDIN: Not at all. f) It's somewhere perhaps 3 just slightly more elaborate than the single heater test.
- 4. DR. FAIRHURST: Um-hum. .Okay.
l '5- MR. HARDIN: As slight as we were able to manage.
'6 DR. FAIRHURST: And sorry, can I have just one
!E 7 'more? l l 8' MR. HARDIN: 'Yes. 9 DR. FAIRHURST; I've done a lot of thinking about 10 this,myself,-but'the effort being given on the thermal 11 cycle, you know, all the amount of looking at it, in the 12 postclosure period beyond the -- what's the net overall 13 effect of the thermal cycle? It's obviously going to stress 14 the rock, it's going to damage it.
.15 MR. HARDIN: What are the durable effects?
-16 DR. FAIRHURST: Right.
17 MR. HARDIN: Well,-let's start with the mechanical 18 effects. We're going to develop somewhere in the 19 neighborhood of 20 to 30 MPA of horizontal thermal stresses 20 .in the rock mass, and we'll have to see a kind of
- 21' first-cycle loading effect, hysteresis. So on cooldown I 22 expect joint dilation, and so at some stage of cooldown will
'23 .be.the mechanical event.that captures the response of the
! 24 . system, and it may. involve collapse of some drift openings 25- or it may not. [;
, ANN RILEY-& ASSOCIATES, LTD.
l' . Court Reporters L 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
. - _ , . _ _. . _ . _ _ _ . . - _ _ _ . . _ . ~ _ _ . _ _ _ _ _ . _ _ . ~ . . _ . . _ . _ _ . _ _ _ . _ _ _ _ -
.]
l 338 1 In the chemical area, that's a subject of a great () 2 -deal of ongoing effort and concern. We will certainly J 3- migrate solutes toward the heat source, and there is, even 4 if we were to age the waste to. expend.some of its heat 5 . output prior to emplacement, when.you put the waste' packages 1 6 .into rock, a low-thermal-conductivity material, some heat l 7- rise is expected. We might be able to control that to 70 or 8- 80 degrees, which would go a long way towards simplifying 9 the phenomenology of thermo-hydro-chemical response.to waste 10 emplacement. 11 So that's -- it depends on the repository design. . l 12 In the extreme case, in a -- we have -- our preliminary 13 calculations have shown that if we heat the rock up enough ! 14 to coalesce the boiling zone and t eate a great, big tabular ] () 15 boiling zone, that we can build up a great deal of water l
- 16. above that, and we can inhibit the percolation of the !
1 17 ambient flux through that zone for some period of time. And i 18 that increases the magnitude of the chemical effects that
-19 result from heating. And so we're actually --
20 DR. FAIRHURST: A potential ceiling? 21 MR. HARDIN: Potential ceiling or concentration of L22- halides, for example, things that might be adverse to 23 package lifetime. Those are considerations in the design 24 selection process.
-25 DR. HORNBERGER: Ray?
l , ANN RILEY & ASSOCIATES, LTD. [ Court Reporters i 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
I l 339 l
- 1 DR WYMER
- Yes, I'd like to nitpick you on three
/~'N I i j 2 basically chemical questions.
l 3 One is, have you taken into account any secondary 4 precipitation effects on the solubility of neptunium? l 5 MR. HARDIN: No, not in the current model, but i 6 there actually are experimental programs under way to look l 7 at that. t 8 DR. WYMER: Okay. And the second one is, I notice 9 you don't have technetium listed here, and that's probably 10 because you say it's going to be moved, and that's the way 11 it is, and there's no reason fooling around with it. But l l 12 again you had the question of secondary precipitation. This 13 is technetium. It's so important it seems to me it might be 14 worthwhile to take a look at that. I f'h \ k ,) m 15 MR. HARDIN: We are evaluating the potential for a 16 surface complexation of pertechnetate on iron oxide and 17 looking for other possible getters. And the idea here is l 18 that it would be used -- I'm kind of exceeding the scope of 19 this talk, but it is possible -- it's plausible that we 20 might be able to come up with a design that might control 21 early releases of technetium. 22 DR. WYMER: And my third question is you said 23 someplace in your discussion that you assumed that 24 radionuclides will migrate toward the high-temperature
- 25 region, I presume based on increased solubility at the n
, ( ANN RILEY & ASSOCIATES, LTD.
\ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
(
1 340 1 higher temperatures. But if the migrating species is a i ew ( mJ
) 2 complex which dissociates at higher temperatures, then of I l
3 course you get the opposite effect, and the complexes might l 4 be such things as chlorides. 5 There are mechanisms for concentrating chloride or 6 carbon dioxide or, you know, carbonate. And also there are 7 temperature effects on redox, and if you do in fact have 8 some mild reducing conditions somewhere, and you don't know 1 9 what the reductant is, then perhaps there'd be a temperature 10 effect on the redox which could act in the opposite 11 direction from the one that you're assuming. l 12 I just wonder what kind of attention you paid to ! 13 these kind of subtleties of the chemistry. 14 MR. HARDIN: When I spoke of migration of solute i (_) 15 towards the heat source, I wasn't thinking about ! 16 radionuclides, I was thinking about species derived from the 17 host rock. 18 DR. WYMER: Oh. 1 I 19 MR. HARDIN: And for the most part the releases i 20 from the waste packages are not expected until after 21 cooldown. So that to the extent that that's true, that 22 simplifies the analysis of transport. 23 DR. WYMER: Okay. 24 MR. RARDIN: No comment on the idea that there may l 25 be temperature effects on reduction or redox. The i [) V ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 l (202) 842-0034 l l
341 1 conditions inside the waste package are really very l t 2 interesting and difficult to predict. If we came up with a w 3 waste package in which the spent fuel would remain in a j 4 reducing environment even after the package was breached, 5 that would have obvious benefits. 6 DR. WYMER: Obviously it would have a major 7 effect; yes. 8 DR. HORNBERGER: John? 9 DR. WYMER: You might even want to work for that. 10 CHAIRMAN GARRICK: Yes. I think I'm impressed 11 with what you did if I could really convince myself I know 12 what you did. 13 DR. HORNBERGER: John's working at his left-handed 14 compliments.
/"'
( )T 15 [ Laughter.] l 16 MS. DOCKERY: He doesn't need to. l l 17 CHAIRMAN GARRICK: You talk about prioritizing 18 these factors, and you indicate that this prioritization l 19 process incorporates things like uncertainties of l 20 performance, et cetera, and then you say other
'21 considerations involved operational policy. Of course l 22 that's extremely vague and not clear how you in fact 23' incorporated that into the prioritization. And then the 24 catchall of environmental health and safety. And then you 25 close the session with cost and schedule, which is a real l ,m ANN RILEY & ASSOCIATES, LTD.
(
-}
Court Reporters i 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
l 342 1 litany of things that are hard for me to get good resolution r~n 2 () on as to how they were really considered. 1 3 Let me ask this question, if you did the 4 prioritization and you concluded by saying that the 5 technical work needed is identified by using prioritization 6 of principal factors, if you based the prioritization just 7 on, say, the performance assessment and didn't consider I 8 these other considerations, et cetera, et cetera, and the ' 9 performance assessment certainly considers uncertainty, how 1 10 would it differ from this? 11 MR. HARDIN: If we had more time, we would drill 12 more bore holes in the saturated zone. If we had more time, 1 13 our engineered barrier system prototype tests would be more 14 advanced by the time of LA. We would have a better
) 15 empirical basis for the predictive models that will we used 16 for in-drift hydrology, thermal hydrochemistry. Those are 17 areas that I expect there will be work continuing.
18 DR. HORNBERGER: But that answer then is that you 19 wouldn't do anything different, you would just do it more. 20 MR. HARDIN: Yeah, I think that is a valid l 21 interpretation. I mean we have exercised our technical 22 know-how and we know what we think we need to do. In some 23 cases, we just know that we have to make calls and that some 24 of those things that we like to do, because we think they l l 25 could be very useful in constraining the PA model, may have l [ \% ANN RILEY & ASSOCIATES, LTD.
\~ I Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 i Washington, D.C. 20036 l (202) 842-0034
l l 343 11 to wait. Or we may find down the road, a year or two years l
-{ 2- from'now, that we have achieved reasonable assurance through l 3 changes in the repository reference design or the repository ;
! l 4 safety strategy, and we no longer feel it is quite as 5 important-to undertake that set of activities. 6 CHAIRMAN GARRICK: And is this, the results of I l lF 7- this going to actually be used? Are you really -- is this ! 8 'really going to be the basis of allocating resources, for 9 example?
- 10. MR. HARDIN: Yes. Without question. The budget 11 this year is very different from -- in this area, is very 12 different.
13E DR. HORNBERGER: Andy, do you have anything? 14 DR. CAMPBELL: Yes. When you talked about -- or, () 15 actually, when Holly talked about backfill, and reading 16 through the technical basis document, I get the impression 17 that you actually haven't analyzed backfill, at least 18 insofar as how it would affect the moisture movement in the
- 19. drifts. Is that-a correct characterization? Because that 20 would be a very different conceptual model.
21 I mean, basically, your conceptual model that you
' 22' .have drips onto the waste package, which has an enhanced 23' corrosive effect relative to just humidity, for example.
24 And then you develop patches and that allows a certain 25 ' amount of water in based upon the size of the patch. And l' ,\/ /) ANN RILEY & ASSOCIATES, LTD. , Court Reporters L 1025 Connecticut Avenue,. NW, Suite 1014 Washington, D.C. 20036 ls (202) 842-0034
. w , - , - , + - - - c ,,..+yr--+. r - . , ~ e. -. -,-.1--r , e- m-
~
V , L 344 l 1 everything is driven by that. And so that is where I get () 2 back to this idea.I raised earlier about, are you looking at l 3 alternative conceptual models? 4 It goes back to the EBS working group we had in 5 . June, where this concept of a " robust," in quotes, L 6 repositor
- approach was-put forth, and has been put forth l
-7 many times, in which you have an-area of-very large 8 uncertainty and large impact on your result. Do-you then 9 step out of the model and say, how could we nullify that 10 effect or change it completely? And it would seem that a 11 full analysis of backfill in terms of its effect on moisture 12 flow, as well as maybe some chemical effects, might be
.13 worthwhile. Because your ceramic coating and your drip 14 shield, really, are just another layer of C-22 or some other 15 version of a barrier, and it shifts things out, but it
-16 doesn't really change things.
17 'MR. HARDIN: Yeah, I agree. 18 DR. CAMPBELL: Okay. That is an observation. Is 19 there any intent to truly do an analysis of effects of j
-20 backfill in terms of the moisture field and how that would 21 affect water getting in or possibly not getting into the 22 uaste. packages?
23 MR. HARDIN: Yes, there is. Right now there are 24 scaled prototype tests under construction that will drip 25 water into a cylinder that has a simulated waste packaos ANN RILEY & ASSOCIATES, LTD.
+ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
1 345 1 into which backfill has been emplaced. P () f% 2 DR. FAIRHURST: Is that just regular backfill or 3 is that enriched? 4 MR. HARDIN: Well, the plant calls for both. The 5 backfill behavior depends on some -- you know, the lateral 6 dispersion of the unsaturated flow, even as a steady state 7 process, it depends on properties of the backfill. And that 8 is kind of an important issue and the EP3 group is looking 9 at that also. 10 MS. DOCKERY: Ernie, can I -- 11 MR. HARDIN: I was going to ask you a question, if 12 I might pose it to you. In the sensitivity -- or in the 13 analysis of-options that was done for TSPA-VA, what was the 14 dripping model used for backfill? I don't recall.
\ss) 15 MS. DOCKERY: That's a good question. I don't --
16 I would have to go and look at that one. 17 MR. HARDIN: I don't think there was a difference. 18 MS. DOCKERY: I don't think there is a whole lot 19 of -- 20 MR. HARDIN: In which case, the effective backfill 21 is primarily thermal, thermal -- 22 MS. DOCKERY: Yeah, 1 believe that you were right 23 on that. 24 MR. HARDIN: It insulates the package and keeps it 25 drier and hotter for a longer time. r () ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
_ _ _ . _ . - _ _ __ _ . _ . _ _ - . . - ~ _ ._._,_. _ __._ _ _. -. _ . _ . . - 346 1 DR. CAMPBELL: _Right. But you have got a a f~%. t
) 2 . conceptual discontinuity there because you have got effects
'3 on the flow of moisture by the physical presence of the 4 backfill and the voids, and the permeability and everything. !
5 But, in essence, what you are saying 4 3 you still have 6 advective dripping of water onto a waste package other than 7 the thermal effects of the backfill. 8 MR. HARDIN: And we haven't taken credit for the 9 lateral dispersion of a concentrated source of water. 10 DR. CAMPBELL: Right. 11 MS._DOCKERY: One of the things that I wanted to l 12 add is that all the analysis that we saw before were done on 13 a reference design which did not include backfill. And, so, 14 then, when the design' options were done, a quick and dirty
/y ,
(s,/ 15 analysis of kind of a "what if" was added. We do -- 16 certainly, if design selection comes in that we should put 17 ~ backfill in, we will have to do a much better job. And we
- 18 also recognize that we are going to have to do an analysis 19 that might look an awful lot like backfill as the drift
'20 self-backfills. If the drift were open and it starts to 21 collapse in on itself, and some of the models say as early 22 as a thousand years after closure, you may basically have 23 the rubble filling up the entire drift. In that case, we 24 have to look at the contact modes with the waste package and I 25 how that affects the degradation model, and that is part of
! ANN RILEY & ASSOCIATES, LTD. F '- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
347 1 our planning for the next go-round on.a TSPA. () 2 MR. VAN LUIK: This is Abe Van Luik, DOE. Is it 3 possible' for to interject something also? 4 DR. HORNBERGER: Please. 5 MR. VAN LUIK: I wanted to defend Ernie, I think 6 he has done a wonderful job in explaining how we prioritized 7 for this year. We have actually implemented the 8 prioritization in our planning, and it was a call that said 9 we are going to do a license application sometime in the 10 year 2002. Where is our vulnerabilities, and what can we do j 11 with our limited time and resources to fill those
- 12. vulnerabilities to where we could fill them? And that is 13 exactly what you are seeing here is the results of that 14 effort.
r (_ ,/ 15 I think it is important to know that, you know, at 16 one time, in fact, the current schedule, if you look at it, 17 says that the SR and the LA, TSPA are the same analysis. 18 After we did this, we said holy cow, that is not possible, 19 because then we can't do any of this work because the SR is 20 done this coming year. So we were insightful and flexible 21- enough, probably not enough to satisfy everyone, that we are 22 now creating an update of the whole TSPA for the LA 23- specifically to include the important points coming out of i 24 this work. So, I think, you know, there's something to keep
'25 in mind in case you are still looking at the old schedule.
[\- ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
l l 348 l'- DR. HORNBERGER: Ernie paid appropriate lip () 2 service to natural analogs and.I. appreciate that. I do have l 3 a question on natural analogs, though. It strikes me that, ~ I I l 4 again, everything that we have been seeing, the performance, I l '5 and in your priority-setting,.it is clear that the' i l 6 performance of the canister, the corrosion of the canister 7- itself is quite important. Is there any chance, is there 8 anything~out there that could be natural analog to alloy-227
-9 I just got used to saying C-22 and now you guys have changed 10 it on me again. Are there any natural analogs that could be 11 used?
12 .And the second question, if the answer to that is 13 a simple no,'can Pena Blanca, does it offer good l 14 possibilities for building public confidence in Yucca !
.15 Mountain?
16L MR. HARDIN: I don't know the answer to your j 17 question about alloy-22. I know that the alloy has only i 18 been in existence for a couple of decades. It derives its 19 corrosion resistance from aspects that are common to other 20 alloys. It maintains those characteristics over a wider 21 range of conditions. I don't know whether there are any
-22 good natural analogs.
23 Pena Blanca is of great interest to us because -- 24 and we haven't done this, but we think that if we applied 25 the conceptual models and the actual model, PA model, to l-I- ANN RILEY & ASSOCIATES, LTD. j - Court Reporters L 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
f-o _ 349 1: Pena Blanca,. it would show that the various' uranyl species
). 2. should have moved out a long time ago, and we would like to 3 --understand that system in that context.
4 I think that it will go a long way towards i 5 supporting the secondary phase part of the source term model 6 that we currently don't_have. 7- DR. HORNBERGER: Tim.
-8 MR. McCARTEN: Yes. Tim McCarten, NRC staff. I 9 believe, and I can get the articles for you, that there is a j 10 mineral, josephinite, that might have some promise of being 11- similar in some characteristics to C-22, but I can get the 12 -- there are some articles on it that we can provide both to 13 you and to -- now, it is my interpretation of Michael 14 LMcNeil, so, I am not a metallurgist,.so I may be
) 15- misrepresenting it, but that is my. understanding. ,I can get 11 6 the articles, though, both to DOE and to the Committee.
17 DR. HORNBERGER: Thank you, Tim. l' 18 DR. CAMPBELL: Josephinite is a phase derived from 19 certain meteorites that have extremely high nickel content. l 20 That's the analog. I l 21 DR. HORNBERGER: Okay. Any other questions? 22- [No response.] 23 DR. HORNBERGER: Thank you very much, Ernie. We ?. l 24 -will' move on then to postclosure and defense-in-depth. And, i l .. 12 5 Jack, you are back on. l-1 ANN RILE. & ASSOCIATES, LTD. '3' Court Reporters
' 1025 Connecticut Avenue, NW, Suite 1014 E Washington, D.C. 20036 l
(202) 842-0034 e e ,-n , , . - - - ,-g--. , - - ,
. -,.n. . - . - . . - -.. - . ~ . . - - . - -...... - . - .._. - ..... . _ - . - .
i 350
-1~ MR. BAILEY: Good afternoon. Thank you for having 1
i i p' (j
~
2' me on again. I'm going to talk about postclosure defense in
.3 depth, and I'm' going to see'if standing makes working that
- 4. machine a little bit easier.
i 5 Two or three things before I actually get into the
~
6' . talk. First is that-what you're going to see today is an-7 early version of some work that we've done and shown in i 1 8 preliminary fashion to the DOE. I have a very nice paper 1 9 which is not ready for publication and is my. brains on this l
~
10 subject, and I may have to refer to it if we get too deep in l i 11 the questions. But I'm going to describe a method 12 associated with approaching defense in depth which I'll go I 1 13 through inside of here. H 14 Now there were two other issues which I thought 15' have come up a couple of times and were handled.again, but I I 16 -felt compelled to say a couple words about it, related to 17 defense in depth in my mind, and I'll tie them together in I 18 case you don't happen to see the same tie that I do as we go l 19 through. 20 Dr. Garrick asked a question about in putting 21 these models together did we put together conservative 22 models or realistic models or what we called optimistic 23 models. And in direct answer to that about a year ago we l 24 sat down as a team and tried to do a subjective evaluation j i l i 25 of is each one of these models optimistic, conservative, or i 1 I 1 L e i ANN RILEY & ASSOCIATES, LTD. i l.N Court Repcrters . 1025 Connecticut Avenue, NW, Suite 1014 ) Washington, D.C. 20036 (202) 842-0034 , l
~_. _ __ . ....-
. -- _ _ _ . = _
351 1 realistic, optimistic meaning that it gives us results which (i* w) 2 frankly are favorable to the dose that we see, but the 3 family, the project would probably not agree in whole with 4 this. And by the way, I subscribe to your theory, Dr. 5 Garrick, that the data should drive the probability 6 distribution function. 7 So we looked and did that and set as goals for 8 ourselves that we would try and make everything either 9 realistic or to the conservative side of realistic, so that 10 we could use these models without concerning ourselves with 11 whether or not an answer that we're getting is being driven 12 by a nonconservative aspect of the model. That was one of 13 the goals unat we set out to accomplish.
, 14 What we found is that we had a couple of f 1
(/ 15 nonconservative areas, at least from our critics. Cladding 16 was felt to be nonconservative. We couldn't justify it. 17 And the other was that in the early stages, the early 18 10,000-year-type period, the unsaturated zcne transport 19 below the waste package was in fact nonconservative and 20 didn't treat the transport of the radionuclides correctly. 21 Both of these were clearly pointed out by our TSPA 22 peer review committee, and we were aware of those as well as 23 we went through this, but we felt that it was what we could 24 do in the modeling in the time frame that we had. And so we 25 had two nonconservative areas which we're very careful to O () ' ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
._ . _ _ . _._ _. _ __ ______.._-..__.._..-._.___..m es 352 T1" ' deal with.
~
(f 2 We believe that pretty much everything else was , , -3' somewhere in the-realistic-to-conservative stage, based on f I l 4 the data that we had when we took our' snapshot and began our 5- model exercise about a year-ago and-tried to work through it i-6 -- .to the final' pieces.that we had. So we in. fact considered , n .7 exactly what=your question was, is.can we keep.this thing
'8 preferably'in realistic so we don't mask effects of the 9 system by overly' conservative or' bounding pieces, and don't 10 create nonconservative results which mask what the system is
. 11 really doing.as well.
12 That allows us to have the PA for use as a tool in 13 some pieces so that we can have some confidence in results
-14 and how well it works.
l ' 15- Now Ernie answered a question a couple of minutes 16 ago which had to do with what we'd do if there were more 17' ' time,-and I think that Dr. Hornberger explained well, you 18 would do more. And I think the excellent example of that is I 19 the saturated zone. The saturated zone is almost uniformly 20 reviewed as being extremely conservative. It is a 21 conservative rendition of what probably happens in a ! 22- saturated zone. The more work that we do will probably make 23 the saturated zone processes cause the end result, the dose 24 at the boundary, if you will, to become less. We probably 25 can get more dilution. We may have the question on the i l- ANN RILEY & ASSOCIATES, LTD. Court Reporters
'1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
. . . - . - . - . - - . . ~ . . - . . . - . - . . . - , - . . . - . . . - . - . . . - . - . - . . ~
l 353 l- [ 1 reducing capability. We may learn more as_to what the
-~
! 1 y j' 2 alluvial aquifer does as opposed to what the tough aquifer 3 does. And so that's-probably in all cases a conservative i 4 result. 5 Now we can be criticized for having a conservative 6- or a bounding result, and does it mask something else in a l 7 system, and we have to be careful to that. On the other 8 hand, if we have a conservative approach or a conservative ; L 9 process and we are able to satisfy all the aspects of the 10 postclosure' licensing case, then that may very well be an 11 adequate representation of what we need in order to make our 12 arguments. l 13 And so the_ question, and I talkedfabout it L 14 earlier, I don't talk about necessarily reducing 5
.\ 15 uncertainties, I talk about understanding uncertainties. If t 16 we can adequately bound our uncertainties in a model such !
j ! 17. that that bounding does not cause a hiding or a change in a ! 18 behavior in another part of the system, which means you want 19 to get them in the right ballpark, then that may be good
.20 enough to move forward. And I'm afraid only God knows l
21 exactly what the uncertainty is on any particular area, and j 22 how many more times you can reduce it, it's hard to say. l' 23 As I think I said earlier, the idea of being able i 24 to predict deterministically exactly what's going to happen l 25 . with seepage, at what point in what tunnel at what time and i i
't -
ANN RILEY & ASSOCIATES, LTD. Court Reporters y 1025 Connecticut Avenue, NW, Suite 1014 ( Washington, D.C. 20036 (202) 842-0034 I
354 l 1 how much,.is probably an unrealistic use of resources and an ' j ) :2 unreal'istic belief on our part as opposed to establishing a 3 range or something that we can work with there. 4 Now those-are two questions that came up in the 5 course of this, and I'm going to talk to the postclosure l 6 defense in depth. How are they related? They're related at 7 least in my mind because we want to use the PA as a tool to 8 identify to us how the system performs, and then we want to 9- slice and dice colloquially the system in lots of different 10- directio".s to understand what the performance of the system l 11- .is given certain perturbation. And that's what the defense 12 in depth is. And we need a tool, if you will, that doesn't
- 13 have nonconservatisms that mask what's really going on -- we 14 want'it to be as realistic as possible -- as long as we're i 1
l /N l) s 15 _ aware of the conservatisms and what they might do to us.
- 16. So with that, I'll talk about defense in depth.
I 17 And you will recall from my diagram of the slice of the 18 engineered barrier system that I gave a year and a half ago 19 almost t-o the day in this room, where I talked about it from 20 a purely engineering point of view, we looked at -- if I can 21 get all my charts together here -- we used defense in depth 1 22 to assure safety, as it says, when the quantitative 23 assessment includes significant uncertainties. If we're not 24 sure exactly what's going to happen in the deterministic 25 'world -- thank you; I saw it -- in the deterministic world t 6 ANN RILEY & ASSOCIATES, LTD. l ('-) Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 ! Washington, D.C. 20036 l (202) 842-0034 I ._
355 1 perhaps for those things which are very low probability but 2 ( ) may have a big effect on us. And so we consider those and 3- make sure that they as an event can be handled.
'4 Now we've made defense in depth an essential 5 element of the postclosure safety case. The philosophy 6 typically-employs multiple protective measures, so that if 7 any one of them fails, something else is there to bootstrap 8 it up and not make it a problem.
9 I talked earlier about vulnerability, not having 10 one thing that protects you, and everything else is 11 secondary, but try and even those pieces out. In general 12 it's multiple barriers, it's conservatism, make it thicker 13 than you have to, if that's an appropriate piece -- making
.14 the wing thicker on an airplane may not be a good idea to O( ,/ 15 make it fly better, You may want to build it into 16 flexibility instead so there's some design consideration.
17 You want to look at redundancy, having two of the 18 same thing. Diversity, two of the same thing but they 19 function in different manners, a la two different kinds of 20 instrument, an electric-driven instrument or a 21 direct-reading instrument. So you may want to consider
- 22 different failure mechanisms, if you will, in the system 23 design. And of course there's other measures, and that's in 24 the preclosure world quality assurance, technical 25 specifications to ensure the analytical bases are met, and l
\ IO ANN RILEY & ASSOCIATES, LTD.
\~Y Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
-m n --,
.. - - . - - . - - - . - - - . ~ - - - - . . - ~ . . - - ~ . . - . -
N 356 l
~1z perhaps something like emergency plans, which allows for L2 interdiction should you have some kind of a problem, but the
. 1 defense in depth is in fact the interdiction by the human.
3' ; 4 Now in'postclosure -- if I go't it all on there -- in the postclosure world we have a passive system. l 5 The ; l L6 -principal safety issues are exposure at a-boundary.a long < 7 time from now, and movement of the mobilized radionuclides 1 8 to the accessible environment. It's demonstrated through l ! l 9 analytical predictions that are based on models of natural '
' 10 processes, and those natural processes are both the natural 11 -system and what happens to the engineered system, each one 12 of those processes containing uncertainty. So it's
! 13- important that we look at how'are we figuring out what's 14 going to happen that's through this modeling process as l 15 opposed to something that we can go and break and test and 16 know that it's going to work this way. l 17 Therefore, we rely heavily on multiple barriers, 1 L 18 and in our mind it may be redundant and diverse to the 19 movement of water. I believe it was stated earlier that it 20 'looks.like water is the enemy, and the answer is yes, and if l 21 you can find ways to interdict the water, then you've 22 probably got yourself into pretty good shape. 23 And the ocher is you may want to use barriers to 24
- address the uncertainty and the limitations on your models,
' 25- and I go back to seepage. If we can't figure out exactly f-ANN RILEY & ASSOCIATES, LTD.
f() ' Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 ( Washington, D.C. 20036 (202) 842-0034
l 357 l' what. seepage looks-like, then maybe we want to put something () 2 .like'a drip shield in that reduces the importance of'the L 3 uncertainty associated with that. seepage. So we can put 4 these in in two different ways. l 5 Now missing from this slide, and it shouldn't be, 6 is what was at the bottom-of the first slide, and that is, !~ 7 is that quality assurance processes, as we put together our 8 process,-as we gather our data, as weido our engineering, as i 9- we'dofour fabrication, potentiallyl10,000 waste packages of l 10 a very large size device, which is subject to corrosion 11 : failure, the quality process associated with that, the 12 installation, the technical specifications, so the quality 1
.13 processes and those human factors that can be controlled 14 --prior to closure of the facility have to be considered as i
) .15 part of the defense in depth in the long run.
16 Now, Part 63, in draft, says that the NRC relies l 17 'on defense in depth to assure safety when the quantitative l 18 assessments include significant uncertainty. In the past
'19 that was done in a set of subsystem performance requirements L
i 20 which were made before the system was fully understood, and (~ 21 so maybe weren't completely appropriate as selected with i F 22 quantitative measures. But what the NRC has now said is use 23 the multiple barriers to enhance the performance of the j 24 system. I 25 And what they want us to do, there's good news and i i () ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 i Washington, D.C. 20036 l (202) 842-0034 i I.
i 1 358 i 1 bad news, the good. news is that they asked us to do it, the
'l [ 2 bad news, they didn't-tell us how. So we'have to make it
- 3. up, that's both good and bad, we have to identify the 4 principal barriers.to this system. How do we decide what 5 'they are? And we're going to talk about that in a minute, 6 provide a transparent assessment of the contribution of each 7 barrier -- this is what Part 63 colloquially tells us to l
8- .do -- and indicate the system performance enhancement 9 provided by that barrier, and whether system performance l 10 hinges on any single barrier performing as anticipated. 11 We're back into the vulnershility analysis, what are the 12 real barriers and how do they work? 13 That's what I am going to talk about. So our 14 approach, and you are going to see this several times -- yS) (_ 15 evaluate the expected performance and identify the principal < 16 barriers.
- 17. You have listened most of the day to what the 18 performance assessment told us about a reference design and i
19 you haven't heard what the principal barriers are yet. I am 20 going to go through a different way of looking at the same 21 system -- what are the threats to performance, and then what l 22 we call selecting neutralizations, and instead of running a L 23 simple sensitivity study, and I shouldn't say simple -- they l 24 are not easy sensitivity studies, as you guys ask lots of 25 ' questions -- but rather than just run a sensitivity study of s I h ANN RILEY & ASSOCIATES, LTD.
'As# Court Reporters
, 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 (202) 842-0034 1, I
359 1 extending something'necessarily to its l'ait of uncertainty
.() 2 'it.in' fact is somehow changing the system, and part of what 3 we did here is we went back and ran the PA by changing a 4 part of the system.
5 Of course, the models'weren't set up to not l L 6 represent the system. -They were set up to represent the ! 17 ' system and so another reason that this work is in draft is 8- we look at how we made modifications to the TSPA runs to-9 accomplish this, and I will point that out.
.I
'10 We need to neutralize the barriers that reflect 1 11 the likely threats and determine the' contribution of each 12 ~ barrier to the total system performance, I might give you 13 just a couple of examples, and then evaluate the overall 14 postclosure defense-in-depth evaluated by the system.
() 15 In boxology, which you can't read --
)
16 DR. CAMPBELL: That's as good as it is going to 17 get. 18 MR. BAILEY: As good as it's going to get? All 19 right. I'll have to read it then. 20 Very simply, I already said it -- identify and 21 assess the roles, evaluate the expected performance, 22 identify barriers, determine the threats, conduct a 23 neutralization analysis, and then do a comparison, and I am l 24 going to show you an example of some of these is what we are
- 25 going to walk through here, and you will see this chart a
'4 -(~si
' ANN RILEY & ASSOCIATES, LTD.
2 Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 l . Washington, D.C. 20036
- l. (202) 842-0034 1
360 l 1 couple more' times, l , p, - 2~ The barriers have been identified not so much in p 3 the TSPA analysis as by the TSPA analysis, and we set up two 4 criteria in terms of identifying what we believe the 5 barriers should be, and you.can see down here that we looked 6 at what is a barrier to water, what's a barrier to 7 radionuclide mobilization, and what is a barrier to ! l 8 radionuclide transport, what keeps it from going anywhere, and'what keeps it from getting entrained in the water or 1 9 ' 1 10 allowing the water to do its dirty deed anyway?- ' 11 Those are the - that is where we ended up, and we l 12 looked at that from two standpoints. Was there a barrier or i 1 13 a containment that appeared to last for 1000 years? We just 14 took that. Is there something there that looks like it (O _,/ 15 would provide us 1000 years of containment, so we chose what 16 looked like it had the potential for 1000 years, and the 17 second is we took a look for the transport or the 18 mobilization of what gave us a 10 to the minus 4th reduction 19- in transport or mobilization. Could it just slow it down? l 20 Could it retard it or slow it down in a timeframe, not 21 necessarily -- so it is not actually a barrier in the sense l 22 of containment but it is a barrier in a sense of slowing it l 23 down and reducing it. 24 So we went back to the Total System Performance 25 Assessment that was done and we came up with these and we ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
. _- __ _ ~ _ .-
. - . _.- -. ..- - . . - . _ _ - . - . - _ ~ . - _ - . - - . . . - _ . . .
361 1- said -- and remember we had attributes earlier. We had the (} 2 four attributes, which are to keep the water off, keep the
-3 waste package-intact, slow the release rate, and then 4 dilute, and those are not too much different than what I 5 described before of-keep the water away, containment, or 6 have a reduction, but we looked at the whole system and we j 7 gathered the system together and concluded that the 8 overlying rock units create that environment.for the waste I 9 package which has a small amount of moisture in it, and that 10 is dependent upon the climate, what makes it thr~ ugh the --
11 first through the rock units above the waste package, and I 12 the seepage into the drift. 13 The second is the waste package itself. How will I 14 the' waste package perform? l j () .15 The third is the cladding itself, what happens 16 with the cladding if we are wrong about the cladding or the 17 cladding doesn't work in the manner -- another thousand year 18 barrier. l 19 The waste form itself, a barrier to radionuclide
- 20. mobilization, and that is the ceramic associated with the i
21 pellet or the glass doesn't exactly turn into something that 22- transports very readily. It does it very slowly. It l 23 actually doesn't meet the 10 to the minus 4th criteria, but 24 we put it in there to take a look at it anyway to see what 25 kind of an effect it would have. I i O
- t. ; NJ ANN RILEY & ASSOCIATES, LTD.
Court Reporters
'1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 lJ (202) 842-0034 L - .
. . - . ~ - . . - --.~. .-.- . - . - - - . - - - - _ _ . _ - - .
362 . l -1 We looked at the drift invert, and what we did in
- r. ,
i [ss\ 2~ the invert is we treat the. intact concrete and the rubblized 3 concrete as a diffusion barrier and so you have to diffuse l 4- through-that invert as it comes out of the package and it j 5 can pick up your 10 to the minus 4th reduction as it moves l 6 through that invert, and so we looked at that, but we have
'7 some uncertainty there.
8 We then looked at the unsaturated zone below the
-9 package, which has been talked'about, and we looked at the l l
t 10 saturated zone as it moves out to the biosphere, and you l-11 will notice the biosphere is not included in that we didn't l 12 see that the biosphere could provide that large an amount of 13 reduction nor was it-a containment type barrier. 14 So we laid ourselves out these numbers, and this I ( 15 is preliminary work, but this was the approach that we took 16- of what are the real barriers that we have here. 17 Now this is design-dependent. If I want to put in 1 18 a drip shield, you will end up with'another barrier in here 19 for the drip shield, which although not containment does 20 keep a barrier to water or a ceramic coating could be added. 21 I talked earlier about one of the things that you i L 22 could consider as a criteria is trying to get barriers into 23 each attribute. Well, we took this attribute of the l 24 overlying rock units as a barrier to waste which ran from 25 the climate all the way through seepage. If we chose to !\s h ANN RILEY & ASSOCIATES, LTD. Court Reporters i 1025 Connecticut Avenue, NW, Suite 1014 [ Washington, D.C. 20036 l' (202) 842-0034 L !~ l
,. -_ _ _ . . . -_ ~ . . _ _ . . . . _ _ _ . . _ _._.._...__ ..._____.. _ _._.
I 1 363 1 pave the top of the mountain, do'some type of a surface
-( ) 2 modification cn: build ~the forest up there and maintain the )
r 3 forest so evapotransporation could in fact carry off lots of l 2' '4 - moisture, we.could possibly even enact another barrier 1 5 should.we be, forced'to, going back to the slice of the l 6 tunnel at one point. i
.7 So this is what we came up with with regard to the 8 reference design that we have been talking about all day, 9 but again as you start to look at features, and remember we 10 :are'doing'this features evaluation, alternatives -- I i
11 believe we have six alternatives and some 27 or 29 type of' 12 design features that we're considering to help us. We can 13 use this same kind of a process to see how many different i 14 barriers there'are and start moving around what the worth of
) 15 each one of the barriers is.
16 The'second step -- determine the threats to l 17 performance - we can evaluate the threats to the' individual u
-18 barriers, and they are pretty straightforward -- they don't 19 work -- so those are seven pretty straightforward threats 20 fore the barriers.
21 The second way to consider this is is there a 22- threat that we have got a common model, a common error that l 23 in fact it behaves in some manner and we don't have it j 24 right? 13o we looked at do we have a common conceptual model 25- that if it doesn't work it isn't going to work the same way [ ANN RILEY & ASSOCIATES, LTD. , Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 b . , . _ ,
L 364 l and maybe we_need to zero, if you will, both of those out at l Td \ 2 the same time to look, and so we have taken that as a first 3 ' step. 4 The second is common sources of~ uncertainty. An 5- example of that'is theEwaste package and the zirconium. 6 Althoughfthey are different metals, they are still metals, 7 and our approach to dripping water and it corrodes at so 8
,many milligrams per decigram per year may in fact be flawed L 9 and the whole approach, and so we looked at maybe we should 10 do both'of those at once as'well.
11 Then a failure of one barrier affecting another -- 12 we didn't find one of those offhand. That would be the type l .13 of.a thing where the carbon steel waste package exacerbates 14' the corrosion of the C-22 which right now is being (f. 15 investigated, so we didn't do it, but you might look at 16 something where you have a consequential failure, where-one l
- 17. -causes something else to fail.
18 So that was how we came up with what the cases are 19 we should run. 20 And I am not going to show you all the cases we 21 have run. l 22' The third step is conduct your barrier
'23 neutralization analysis. And, very simply, we looked at !
24 ' conducting analyse.s one by one. We took the threats in the 25 common modes and we neutralized them, and I will talk about i 1 l 4 i-ANN RILEY & ASSOCIATES, LTD. (*9
~/ . Court Reporters ;
, 1025 Connecticut Avenue, NW, ouite 1014 L Washington, D.C. 20036 I 1 (202) 842-0034 i
.. . - - - _ .. _- . =- . . . - , . - - _ - - ~ ~ .
365
- 1 'what neutralization n.eans in a minute.
.2
) 'And it says here,.you can do it the full TSPA 3= methodology, or you can come up with a simpler calculation
)
1 4 . tool to give yourself an idea of how it is going to respond, 5 to determine whether or not you want to tweak the entire 6 TSPA-to come up with an answer. So we have developed some 7 : simpler tools to try and get us in the ballpark with what 8 these mean. 1 9~ And I am going to show you any of those results 10 today, but I am going to let you know that we don't have to 11 go -- because there has been a lot of resource questions. 12 We don't-have to necessarily go and change.the whole TSPA to 13 try one of these. We can, in fact, do some much smaller
'14 computational methods to find out if it looks like it has a (O_j- 15 : benefit before we make the decision to move to TSPA.
'16 Now, a very simple result, and I have picked two 17 examples, and I see I get no color. We took dose as a 18 result and we took time, year after closure. Because of 19 preliminary results, I took the actual dose and I took :he 20 timeframe off of it, just so that you can see the shape of 21 the curve. And what we did in this is we neutralized the 22 overlying rock units.
'23 Now, what does that.mean? It is amazing I got 24 this far without my notes. See if I can read. If you 25 recall, we had climate. We had the six different regions, l
l
- ANN RILEY & ASSOCIATES, LTD.
N- Court Reporters l 1025 Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 i (202) 842-0034 i E
366 1 'which.I believe Holly referred to earlier. There is a [m)] 2 certain~ amount of timing associated with the water moving 3 the 200 to 300 meters from the surface to the repository, 4 although we have looked at fast paths, which is what we 5 really addressing with the neutralization. There may be 6 fast paths, but we have looked at the timing associated with 7 that. 8 Seepage, I guess the mean is about 5 percent of 9 the package will see seepage in the first 5,000 years is the 10 way the seepage model works out, and about 25 percent see it 11 when you hit the long-term average type conditions. 12 Basically, all seepage into the drift at the point of a 13 waste package, hits the waste package, in the current model, 14 and there is an enhancement of how much of the water 1/~)N 15 continuing to hit the same spot, in other words, we up the 16 probability it is going to hit the same spot, in order to 17 account for focusing of flow. That is how the analysis is 18 done for the base case. 19 In the neutralization case, what we did is we said l 20 that the percolation, or the infiltration, take your choice, 21 in fact, all of them, equal precipitation. We said that, 22 basically, it is raining inside the drift for the 23 cross-sectional area of the drift. There is no time delay. 24 Every package is wetted and the focused effects remain, that l ! 25 we had talked about earlier with regard to potential [) (> ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
1 367 1 l focusing effect. ( 2 So this neutralization is not just go to the
)
3 higher end of what'the uncertainties say, the .1 to 30, I 4 think is what ina use with a mean of around 6. Instead of
- 5' just going to 30, we went all the way to what the 6 precipitation is for the year. We took that over the area
.7 of the waste package. No time delay, and every package sees 8 it instead of just one package' sees it.
9 And so from a neutralization point of view, we , in 10' fact, are taking a much harsher review in this approach than 11 we would on just running a sensitivity study of what is 12 going to happen. ! 13 And you can see that things happen a little i 14- sooner, and things happen with a'somewhat higher result. 15 And'if you can read this over here, you end up with a one to L 16 .two order of magnitude difference associated with that ! 17 neutralization of the overlying rock units. Which says that l 18 if we are completely wrong and it is raining in the drifts, 19 this is the kind of a result that we are going to get. That 20 provides us with some understanding as to what can happen l 21 with the system now, and what types of things we can be 22 looking'at to offset that if we need to offset it, and I 23 haven't said that we have yet. l 24 Now, the second one that we did, and we actually 25' did all of them, but I only brought two, and the reason that l I ANN RILEY & ASSOCIATES, LTD. I \ Court Reporters
- 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 i.
368 1 1 I only brought two is that some of these are hard to ! l (~ ,m) 2 explain, quite frankly. Some of these are hatd to explain v 3 and we are still working through whether they make sense, 4' and I will have trouble with this one. Oops, off the chart. l 5 Here we neutralized the drift invert. And what I 6 that means is that the way that we handle the drift invert 7 is its modeled as either a solid or a rubbleized diffusion 8 barrier, depending on the timeframe. And what we did in 9 this neutralization is we took no travel time, so there is 10 -- the diffusion holdup is just advective flow, and we took l 11 no dispersion associated with what the invert could do for l 12 us. And in this case you can see the timing didn't change 13 very much, and there are some differences in the short-term l 14 and the long-term. Again, a one to two order of magnitude j ,) 15 difference, by not understanding fully the uncertainties 16 asscciated with the invert. 17 Now, I have got you two examples. The next step 18 is to determine the change, and I am only going to show the 19 two. The final on the chart is determine the maximum change 20 in the barrier neutralizations. This is a little harder. 21 This is partially why it is still preliminary, in my mind. 22 There are some simple measures that you could use. What is 23 just the maximum change in any individual neutralization 24 analysis? Go through, if you believe we got the barriers 25 right, then go through and see what each barrier is worth. ()
's '
ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
_- -. . . - . ~ . - . . - . - . - . - . . . - . - . - . - . . . . - . f 369 1 And I showed you two here, and you can break them into e:, ( 2 timeframes, for example, and you will see one was_probably a 3 factor of 10 and one was probably a factor of 50, in the 4 first 10,000 years, for example. And we can do all seven.of i L 5 them and see, if they are all between 10 to 50, then none of l 6 them really drive the system badly, so we could do just a 7' simple comparison. , 8 We can get into advanced math and so some mean 9 square average of all the neutralizations and force all of 10 them within 20 percent of the root mean squared. You know, i 11 make some less effective and make some more effective, or 12 leave the ones that are effective more effective and force l 13 the other ones-down somehow. How do you force them down? L 14- Well, you force them down through the addition of other () 15 factors, through other design features, or by learning more 16- about the. system and reducing the uncertainties perhaps. 17 The measure'should indicate whether any principal 1 18 barrier is not compensated by others. Again, the idea being 19 -- and, in fact, the draft regulation that the NRC has put 1 20 out, basically says take each barrier, if you will, identify I 21 them, take each barrier and eliminate it and see how much of 22 a change it causes. It doesn't say what the answer should ! 23 be, but it says find out what those numbers are. So the 24 idea is to find this means of a comparison. j - 25 And another means of comparison is, does the l l ANN RILEY & ASSOCIATES, LTD. O-r Court Reporters l 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l r 4 - , ,m.c , . . . . _ . . . _.g. __ , . . - , . . , . - _ ,,
370 i 'l measure,'if you take'it out, do you still meet -- we chose a
/~w}
2 25 millirem, which has no significance, just a number that 3 we'put in. 4 Good day, Dr. Garrick, 1 5 CHAIRMAN GARRICK: You haven't driven me away. 6 MR. BAILEY: I didn't think'I had, sir.
)
- i. 7 And so you can look at it with regard to the l 1
8 overall performance measure. I mean there could be a l 9 measure that says you have to -- a neutralization of each 1 10 barrier, neutralizing a barrier, you have to still meet the 11 standard. And we have made up a standard here, there is no l 12 meaning to this standard except to throw a number on the 1 13 board. But maybe you have to meet it. 14 On the other hand, maybe you don't even have to p (,) 15 meet it. We are talking about a very extreme case. We are 16 not talking about_what we had talked about previous today I 17 with regard to running the probabilistic evaluation and 18 finding the mean and using the mean, we are now driving 19 ourselves, in one case, to the extreme of what sort portion 20 of the system might be able to do. And I will go back and 21 review what happened, as I said, in that overlying rock 22 units, we,.in fact, changed a half a dozen parameters that 23 are now more conservative than they would normally be in the 24 way you do it. So we may not even believe that this has to 25 be met, and we haven't concluded that. And so that is one
- _ ANN RILEY & ASSOCIATES, LTD.
t 's- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
l 371 l 1 of the reasons why we are still in draft here as to how to (rm) 2 do it. 3 And as I said before, we are going through this 4 alternating design and feature selection process and we're 1 5 'using this methodology as part of our guidance to try and 6 choose systems. 7 Now'how does that happen? Well, if we choose a 8 system that doesn't have -- let's pretend for a moment, and 9 I think we all recognize it, that the waste package, had I 10 showed that chart, probably has a factor of five orders of 11 magnitude difference. And we can go back to the TSPA and 12 probably figure it out. But it might be four or five orders 13 of magnitude. Very big one. Now if we went to, let's say, 14 a low-thermal-load repository, and the kinetics of corrosion (~N ( ,) 15 are different and now it's only two-and-a-half, that might 16 be a basis to consider that may be a better design. Or you 17 might want to put in a drip shield and a ceramic coating and 18 drive all of them to the same level again. 19 So there's several ways to skin the cat with 20 regard to trying not to have a single-point vulnerability. 21 And that has to be within the purview of the Department of 22 Energy to make those decisions based on these types of 23 evaluations. 24 So what we did, we believe, is this approach that 25 we've taken leads to an identification of principal l [ j ANN RILEY & ASSOCIATES, LTD. ' k s' Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
I' i! 372 1 barriers, and what each does, in a neutralization point of: i
} 2' view, to the system performance. It illust' rates redundancy 3- fof' barriers,'and I like the~ word resilience of the system to 4' potential threats to performance.
l l5 You know, I'm not sure everyone would have guessed 6 that taking away~the entire overlying rock unit.would have t 7 only changed the result by that factor. That's not 8 something that's apparent. And these are draft results. So 9- we have to go back and look. But that's the-kind of results 10, that we were finding. l 11 And it helps us get an understanding of the system 12 once again, and understanding the system is what we're 13 trying to do. It evaluates the postclosure defense in depth 14' of the system within the context of the total system O 15 , ( ) performance assessment. We can use it along with expected 16 performance cost, other factors to compare design. As we 17 said before, we have to keep those in mind as working a 18 project. And what we do is traceable and repeatable so that 19 you can see how those pieces come together. 20 Preliminary presentation. I'll take questions. 21 DR. HORNBERGER: Jack, I have an observation to 22' start off with. 23' MR. BAILEY: Shoot. 24 DR. HORNBERGER: The ACNW recently did a trip to 25' ' Germany, where we met with some international people, and l j O i(> ANN RILEY & ASSOCIATES, LTD. Court Reporters l 1025. Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l' '
_ -. _ __ _ m.. . - - , _ . - - _ - _ . _ _ _ - - . _ . _ _ . _ . - - . . _ . _ . - . _ - _ _ i 373 1 our French friend, Monsieur Brigaud -- is that his name? -- () 2 3 took us to task -- well, took John Garrick to task for using the term " defense in' depth," and he pointed out that we 4 really didn't know what we meant when we used the term, and 5- that we're adding confusion because in the reactor business, l 6 some people at least think they know what they're talking 7 about. 8 And the ACNW, at least, we have strived to try to 9 stick with multiple barriers when what we're talking about
- 10. is multiple barriers. That's just an observation, but I was 11 going to invite Jack Sorenson, who has thought long and 12 deeply about this, whether he had any comments that he would 13 like to make.
14 MR. SORENSON: Not at the moment. I'm still f\ 15 trying to digest this. 6G 16 I did have one observation, and that is the ! 17 ' quality assurance, the reactor side is not an accepted j 18 element of defense in depth. It's part of the condition you 19 operate in, not a barrier in the sense. 20 DR. HORNBERGER: I think Brigaud's point is that 21 we may tend to do ourselves harm by using the term in a very 22 poorly defined sense. 23 MR. BAILEY: If I may interrupt for a moment. 24 DR. HORNBERGER: Please. 25 MR. BAILEY: The TSPA peer review panel provides O ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washi"gton, D.C. 20036 (202) 842-0034
374 1 us with the same commentary that it is a term of art from ( 2 l [v) the deterministic reactor world, and it clearly isn't the 3 same, which is why I tried to make a distinction between it. 4 However, the regulation does define it that way, and so I l l 5 have to try and find some means of definition, since they 6 allowed me to do that. 7 DR. HORNBERGER: Marty, do you have questions for 8 Jack? 9 Charles? 10 DR. FAIRHURST: Not really a question. I have a 11 comment. I actually like it. I like this. 12 MR. BAILEY: Thank you. 13 DR. FAIRHURST: No, no. But it's useful, very 14 useful, to get an understanding of what is it, you know, ()
. r .,
15 what happens if you just decide that that's such an 16 uncertain parameter, I'm just going to pretend it has no 17 influence whatsoever. And gradually go through it and as 18 you say taking off the heat, stop 300 meters of rock 19 essentially is what you did. Well, how would it do? Well, 20 there's another side to it that perhaps look at adding some 21 barriers -- 22 MR. BAILEY: Oh , yes. 23 DR. FAIRHURST: Whatever, and say, you know, how 24 much could we get out of that. 25 MR. BAILEY: Yes, and how does it affect the l l ANN RILEY & ASSOCIATES, LTD. l,/~)T , ( Court Reporters l 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
1 j 375 l 1 relative importance of the others. } (( j ' 2 DR. FAIRHURST: Sure. l 3 MR. BAILEY: Maybe I didn't stress that in here. l 4 If I go back to one of the slides. Maybe I went past it too 5 quickly, because I'm too close to it. On this particular 6 slide -- I didn't show a summary of all of them because it i 7 isn't complete, and I didn't feel that it was appropriate to , 1 8 try and give you an overview of it. I l 9 But basically we went through and did these seven, l 10 and we know the relative ratings of what each one of these ! 11 is worth, and now once you put one more in there, for 12 example, the drip shield, which you saw some performance 13 from Holly earlier today, the relative worth of all of these 14 barriers changes a lot. (_,) 15 DR. FAIRHURST: Right. 16 MR. BAILEY: They all change. And so it's like a 17 principal factor table, you put a new element in, the 18 importance of each factor gets shifted based on which i 19 process or which threat in this particular case you have now 20 I interdicted in another manner. And so all of those things 1 21 can change, and you can end up with a different result. 22 And so that's exactly what we're doing in this 23 design process that we're going through is looking at what 24 do you get from base designs and what kind of natural, if 25 you will, defense in depth with this kind of a methodology i; ANN RILEY & ASSOCIATES, LTD.
\- Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
.-, m._ . .m._. . . . _ . . . . . . .___.._ _ __ _ __.- _ .. -. . . _ ......_. _ _ _. __
376 1' results, and then as you start putting in and taking out ) IY 2 different features, how does all of this change and how does V 3- the relative worth of each one of these areas change. 1 4 DR. FAIRHURST: With, for example, this drift
~
P . 5 invert, if you don't use concrete, and don't put in an j
-6 invert, you never have it, but you may say hey, there's 7 enough benefit in there,-we should throw some stuff in 8 there,.it'll do the same thing.
9 MR. BAILEY: Yes. .Yes, very much so. I mean, 10 gravel may very well do the same thing for us in terms of g 11 creating a barrier that diffuses and disperses without the , 12 harmful effects, if there are any, to the chemistry 13 associated with the concrete. And so, you're right, you can 14- exchange and see what'happens as well. () 15 DR. HORNBERGER: Back up to Marty. 16 MR. BAILEY: I got him. 17 DR. STEINDLER: I'm trying to distinguish the 18 benefits derived'from design changes and analysis by the 19 ' output of a decent analysis , TSPA-type analysis, from what I L20 sense to'be in some areas simply a requirement currently 21 written in Part 63 that DOE in fact analyze the impact of 22' barriers. l 23 I think you're mixing, at least I sense that you i 1 are mixing the two. And while it's quite obvious that the
~
' 2 4' 25 impact of design changes which you can call different l-
- I \ ANN RILEY & ASSOCIATES, LTD.
L \% s/ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 , Washington, D.C. 20036 l (202) 842-0034 )
I' l 1 377 1 1 barriers is a useful sort of thing,.the other facet of that ? t(~%
- p. 2 is a kind of a bureaucratic exercise. In other words, what 3' are you going to do with the information that tells you that 4 if you allow it tol rain inside the drift, you get a L 5 particular kind of answer? That's not, it seems to me, L 6 other than being able to say well, we've analyzed it, and l 7J therefore we have satisfied section 63 point something, l
L 8 that's not a very useful sort of thing. i { !- 9 Whereas, on the other hand, if you say well, we
' ' 10 have analyzed the changing composition of a backfill, that 11 is a, you know, that comes under the heading of optimization i
11 2 of design. i l 13 Am I drawing an artificial distinction? I I 14 MR. BAILEY: No. No, I understand your question, l 15' and I don't know'that my answer will take the artificiality i 16 of the approach away, but'let me try. 17 If we: harken back, and I know I don't have the i 18 presentation in front of me that we did a year ago, or a
!19 year-and-a-half ago, we looked at all the different kinds of
'20 barriers we had and what kind of threats, and at that time 21 we said well, we can count. How many different things catch 22- water? How many different things catch radionuclides? And 23 how many ways are there so that we have confidence that if l.
l -24 'something isn't working the way that we think, we can 25 capture them? Okay? ANN RILEY.& ASSOCIATES, LTD. i Court' Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
. - . - . - . . . . . . , - . - . . _ ~ - . - . . ~ . _ . - . . - - - - - - - . . . . . - . -
i L 378 l 1 So first of all we've always thought that'we () l
-2 should find'yet another way to think through-the' performance 3 assessment.
l l 4 To make'sure we understand the vulnerabilities of ,
~5 uthe analysis because~you have performance assessment that 6- gives you a good mean value, but you may find that the
- 7 ' primary contributor to-that mean.value is one particular 8 . process or one particular action and you want to try and 9 find those that you can see'-- can I do something'to resolve 10 that' issue in a responsible manner that maybe I can make 11 this into.a more robust system by not having so much 12 dependence in one area. That's the first thing.
13 Now the'second is that looking.at whether or not 14 it is raining directly inside of the drifts, and it is worth ( 15 two orders of magnitude or an order of magnitude difference
' 16 ' in what you do, we had a fairly lengthy discussion earlier 17 about how much more work should we be doing and where should !
18 we be doing our work. ) l 19 If we find that the characterization of that 200 20 meters or 300 meters above and if we know it perfectly, it 21 doesn't make a whole lot of difference'-- it makes two i 22 orders of magnitude difference -- then it helps us again in 23 'that'prioritization of where we should be putting our 24- energies. 25 If in. fact for point three, when you do each one L I l ANN RILEY & ASSOCIATES, LTD. g / Court Reporters
.1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 ,
1 1 4 u .- .I
, , ._ _. _ . - . - - ~
379 . I i 1 of these barriers, and I have got to come back to that, when '
.D 2 you do.each one of these barriers and you do your l (O l 3 comparison, and you find out that in your contribution to 4 performance that the overlying 300 meters, and I don't know 5 that it is, but let's say that it's the number two player, l
6 the number two from the bottom player in overall 7 performance, and you have got three over here that are three 8 times as big or 10 to the 3 bigger, then maybe that is where 9 you ought to do your resources again, so you have that 10 ability to make these comparisons and understand with the 11 system as to how well do you have to know what you are 12 doing. 13 I harken to something I am usually told not to do, 14 but I will do it, and the question is if I drop this piece rT 15 of paper will it hit the table. (_). The answer is obviously 16 yes, but if the answer is how -- oh, that is different every 17 time. 18 DR. HORNBERGER: You almost missed. 19 MR. BAILEY: Do I need to know how? Do I have to 20 reduce uncertainties to how or is yes enough? We have to be 21 able to make those trade-offs, and I am not saying that the 22 answer is going to come down as easily to yes or how in 23 these particular items, but it gives another insight into 24 what is important, why is it important. That is what we 25 want to use these for, so aside from the bureaucratic i c [i ANN RILEY & ASSOCIATES, LTD. (s / Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
j 380 1 exercise, I think (a) it needs to be done. I think
\
{,-~ } 2 defense-in-depth says you don't fall off a cliff. l
\~J You don't 3 have something and then when something happens you fal' off 4 the cliff I guess in the charts that we draw. You don't fall 5 off the cliff and you are suddenly here and something goes 1 6 wrong and, poof, you are way up -- you want a little jig if 7 something goes wrong in the way that you don't think it is, 8' ao you want to look for that as well.
9 Did that help? I'll try again if you say no. 1 10 DR. STEINDLER: Just a little. ' 11 MR. BAILEY: Okay. 12 DR. HORNBERGER: Enough so that he doesn't want i 13 you to try again. i 14 [ Laughter . ] O) (_, 15 DR. HORNBERGER: Ray? 16 DR. WYMER: No, I'm with Charles. I thought that 17 it was a really rational approach. I think it is one that 18 we have all talked about over the years -- you know, drop 19 them out one at a time and see what happens, see what is 20 important. 21 I am looking forward to the time when you put 22 numbers on these -- 23 MR. BAILEY: I can show you the results, yes. 24 DR. WYMER: -- coordinates and do all the cases. 25 MR. BAILEY: I really hope to get a little i f' ANN RILEY & ASSOCIATES, LTD. ! \,_]/ Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
i l l 381 i- 1 conversation here with regard to what we showed as opposed () 2l
~
to have a definitive statement. 3 DR. WYMER: And I am also interested,in the sort 4 of. fundamental changes in design, like several people have 1 5 alluded to - put in the backfill, which seems to me to be a i 6 major factor both with respect to the roof dropping in and i 7 ' diverting water if you make a Richards barrier out of it and
~8' other ways you might chemically alter it.
I L 9 MR. BAILEY: Yes. j 10 DR. HORNBERGER: In fact I thought that Ray would 11 get excited when he saw the neutralization because he 12 thought you were talking about chemistry. i I 13 [ Laughter.]
.14 MR. BAILEY: I was actually worried about that.
15 DR. WYMER: I caught on pretty quickly. l 16 '. DR. HORNBERGER: Andy? 17 DR. CAMPBELL: Back in July, and Norm is here so 18 he can correct me if I am wrong, but we got a presentation L19 from NRC' Staff on importance measures,.and in that 20 presentation they talked about neutralizing components in 21 doing an importance measures analysis, and they do it I 22 mathematically rather than changing the model around, if 23- 'that's correct. 24 Are you guys aware of that and the Committee 25' actually worked on a letter yesterday and voted on it, on l ANN RILEY & ASSOCIATES, LTD. 1 Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
l 382 1 importance measures, and have you been communicating with
,m
( ) 2 the staff on this? 3 MR. BAILEY: I am not personally on top of the { l 4 issue. I understand -- I think what they are doing and I i 5 don't the specific measures that were chosen. Personally I 6 can't speak to it at this point. 7 DR. CAMPBELL: Then the follow-on question then is 8 more a regulatory question which maybe I shouldn't be asking 9 you but what criteria then do you use to determine whether 10 or not you have a sufficient number of barriers in place to 11 meet defense-in-depth and does it have to be a numerical 12 value? 13 MR. McCARTIN: I have been waiting for the right 14 moment to jump in. O) \, 15 MR. BAILEY: I am sure you have. 16 [ Laughter . ] 17 DR. CAMPBELL: That was a gift to you, Tim. 18 MR. BAILEY: You asked me a question, Tim, so I l 19 want to go first and the reason I want to go first is -- 20 just bear with me -- 21 (Laughter.) 22 MR. BAILEY: -- if I can find my page -- the . 1 23 answer is I don't know. That is one of the reasons we are ! 24 still in preliminary. 25 As I said, I showed three different ways -- simple l [\~/) ANN RILEY & ASSOCIATES, LTD. Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
I 383 l - l' . measure comparison, root means square which was just a 2 mathematical means, maybe it's the standard. We don't know l 3 .the answer to that question and we don't know what we would i 4 propose at this point in time. ! ! That is why I threw several l l- 5 out. l 6 MR. McCARTIN: In writing Part 63 we went to as 7: great lengt.hs as 'we could to not requiring any separate 8 analysis by the DOE. 9 MR. BAILEY: I understand. 10 MR. McCARTIN: Providing flexibility to the DOE to 11 demonstrate multiple barriers whichever way you can, 12 Now I know you did imply and I think, hopefully j 1 13 you didn't mean it, but I would like to point out was l 14 incorrect that it was required for you to remove these 15 barriers and neutralize them by the regulation. 16 At one point you said that is required in Part 63. 17 Absolutely not. 18 MR. BAILEY: That's two paragraphs we need to talk 19 about later then. l, '20 MR.'McCARTIN: Well, it certainly was not the 21 intent and all we were asking for was to identify the 22 barriers, talk to how they contributed to performance, and 23 what the basis for the performance was, and that was it. 24 Now there are a lot of different ways one can do
~25- that.
ANN RILEY & ASSOCIATES, LTD. 'D Court Reporters 1025 Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034 l
l 384
- 1. MR. BAILEY: Sure.
l' ( ) 2 MR. McCARTIN: There is one-off, or the 3L neutralization is one way potentially, but we certainly did 4 not try to require anything other than the performance ; 5 calculation and we are'looking, as Norm has been doing some ' 6 . work with people'here-and at the Center, we have thought
'7 about providing additional guidance in this area, but I 8 think from a requirements standpoint and that is why 9 hopefully we will go out for public comment in the near term )
l ' 1110 on the rule, but we certainly did not try to imply in there, . l 11 and I would be interested to hear where are the words that j 12 we were requiring analysis like this, but -- 13 MR. BAILEY: Okay. Thank you. That's helpful as 14 a clarification and I am not unhappy that you didn't provide
) 15' more guidance despite my comment. It is giving us the 1
16 opportunity to evaluate'and understand the system and ! 17 provide our rationale is I think an appropriate approach. 18 DR. HORNBERGER: Other questions or comments? 19 Norm? 20 MR. EISENBERG: Let me just say that I think there 21 is a remarkable similarity between the work that we 22 presented in July and what DOE has presented today. 1 23 They are not exactly the same but I think we are ' 24 thinking along similar lines for similar uses -- 25 prioritization of the effort and some idea of where the j f l l(' ANN RILEY & ASSOCIATES, LTD., Court Reporters 1025: Connecticut Avenue, NW, Suite 1014 Washington, D.C. 20036 (202) 842-0034
)
1 l 385
.1 vulnerabilities'in the system are.
() 2 3 MR. BAILEY: DR. HORNBERGER: Right. Good. Okay. Thank you very much, Jack. 4 MR-. BAILEY: You're welcome. 5' 'DR. HORNBERGER: We are now at a'part of our
- 6 agenda where we are going to take a break and when we
! 4 l 7 reconvene we do not need the Reporter anymore. ! 1 8 [Whereupon, at 3:10 p.m., the recorded portion of
~9 the meeting was concluded.]
10 11 l 12 1 13 14 15 16 17
- 18
'19 20
~21
-22 l 23 l
24 25 , i [' ANN RILEY & ASSOCIATES, LTD.
, Court Reporters i
,.. 1025-Connecticut Avenue, NW, Suite 1014 l Washington, D.C. 20036 ; (202) 842-0034 I
REPORTER'S CERTIFICATE i This is to certify that the attached proceedings L}A. j
) before.the. United States; Nuclear Regulatory Commission in the' matter'of:
i NAME OF. PROCEEDING: 10STH ADVISORY COMMITTEE
'ON. NUCLEAR. WASTE (ACNW)
DOCKET NUMBER: L. PLACE OF PROCEEDING: Rockville, MD L I~') . were held.as herein appears, and that this is the original l- ( _ / 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 transcript is a true and accurate record of the foregoing proceedings. [%CMh Mark Mahoney k~N l-Official Reporter Ann Riley & Associates, Ltd. e O
O O O YUCCA MOUNTAIN PROJECT Prioritization of Tecanica.. Wor 1 Neec.ec. to
~
Comp..ete the Postc osure Safety Case Presented to: Advisory Committee on Nuclear Waste 105th Meeting Presented by: Dr. Ernest L. Hardin Technical Lead, Near-Field Models .- CRWMS Management & Operating Contractor , U.S. Department oflinergy Office of Civilian Radioactive December 16,1998 waste Management
o o O Elements of the Postclosure Safety Case
- 1. Expected postclosure performance (and supporting evidence)
- 2. Design margin and defense-in-depth
- 3. Consideration of disruptive processes and events
- 4. Insights from natural and man-made analogs
- 5. Performance confirmation plan HARDIN-121698. PPT 2
O O O Repository Safety Strategy (Rev. 2) Approach to Postclosure Performance - Principal Factors ; Attributes of the r Repository Safety Strategy Principal Factors Limited water 1. Precipitation and infiltration into the mountain ' contacting 2. Percolation to depth waste packages 3. Seepage into drifts 4A. Effects of heat and excavation on flow (mountain scale) 4B. Effects of heat and excavation on flow (drift scale)
- 5. Dripping onto waste package !
- 6. Humidity and temperature at waste package Long waste package 7. Chemistry of water on waste package t
lifetime 8. Integrity of outer carbon steel waste package barrier ,
- 9. Integrity of inner corrosion-resistant waste package barrier Low rate of release 10. Seepage into waste package of radionuclides 11. Integrity of spent fuel cladding i from breached waste 12. Dissolution of spent fuel and glass waste forms packages 13. Neptunium solubility i
- 14. Formation and transport of radionuclide-bearing colloids i Concentration reduction 15. Transport through and out of EBS during transport from 16. Transport through unsaturated zone '
waste packages 17. Flow and transport in saturated zone
- 18. Dilution from pumping
- 19. Biosphere transport and uptake HARDIN-121698. PPT 3 l
o o O Technical Work Plans o Prioritization of additional technical work for Element i 1 of the postclosure safety strategy:
-+ 1. Expected postclosure performance
- 2. Design margin / defense-in-depth '
- 3. Consideration of disruptive processes & events
- 4. Natural and man-made analogs
- 5. Performance confirmation plan o Other technical work will address Elements 2 to 5 HARDIN-121698. PPT 4
---------------_------------------_-----------_-------J
o o O Objectives for Prioritization o Determine additional technical work needed for SR/LA o identify how technical work will contribute to TSPA o Incorporate the uncertainty of performance projected for the reference design o Possible future re-prioritization based on any changes in strategy and reference design 4 a HARDIN-121698. PPT 5 l
o o O i i Other Considerations in Planning & Prioritizing Technical Work
- o Operational i
o Policy o Environment, Safety & Health t i i HARDIN-121698. PPT 6 , i
o e O Prioritization of Principal Factors Conduct of Activity o A multi-disciplinary, multi-organizational team was
- assembled by the DOE o Principal investigators were included a
A consultant facilitator was included for control of methodology and assessments o A formal record was generated, and a report produced HARDIN-121698. PPT 7
o O O 0 0 Prioritnization of Principal Factors Method l Priority = Confidence Goal - Current Confidence I l Assessments o Significance of uncertainty to TSPA (L,M, or H rating from PA sensitivity analysis) o Current confidence (1= Low,7=High) o Confidence goal for LA (1= Low,7=High) I HARDIN-121698. PPT 8 ___________-________________________________________-____________-_____s
o o o ' Significance of Uncertainty to TSPA Evaluated using the TSPA model to exercise the ranges of uncertainty recognized for model inputs. Ratings were assigned to each principal
- factor
Low < 5X change in peak dose rate Moderate > 5X but < 50X change l High > 50X change ; and specified as to when the effect occurs: i 0 to 10 kyr 10 to 100 kyr 100 kyr to 1 Myr HARDIN-121698 PPT 9 i
o o o Significance of Uncertainty to TSPA Attributes of the Significance of Repository Safety Strategy Principal Factors Uncertainty
- Limited water 1. Precipitation and infiltration into the mountain M contacting 2. Percolation to depth M waste packages 3. Seepage into drifts H 4A. Effects of heat / excavation on flow (mountain) Ms 4B. Effects of heat / excavation on flow (drift scale) Mn
- 5. Dripping onto waste package M
- 6. Humidity and temperature at waste package La.e Long waste package 7. Chemistry of water on waste package M lifetime 8. Integrity of outer carbon steel WP barrier M.
- 9. Integrity of inner corrosion-resistant WP barrier H._n Low rate of release 10. Seepage into waste package M of radionuclides 11. Integrity of spent fuel cladding H.
from breached waste 12. Dissolution of spent fuel & glass waste forms Mn.c > packages 13. Neptunium solubility Ms.c
- 14. Formation & transport of radionuclide-colloids Msc Concentration reduction 15. Transport through and out of EBS Ms.e -
during transport from 16. Transport through unsaturated zone H. waste packages 17. Flow and transport in saturated zone M
- 18. Dilution from pumping M
- 19. Biosphere transport and uptake L
- Subscripts: "a" = 0 to 10 kyr, 't" = 10 to 100 kyr, and "c" = 100 kyr to 1 Myr (no subscript = 0 to 1 Myr)
HARDIN-121698. PPT 10
o o O Current Confidence An expression of the degree of certainty that the j l current representation of the principal factor for l TSPA is realistic, and captures the entire range of conditions important to performance l I l HARDIN-121698PP1 11
o o O Current Confidence Assessments ' i RSS Attributes Principal Factors
- Signif. of Current i Uncertainty Conf.
Limited 1. Precipitation and infiltration into the mountain M 4 water 2. Percolation to depth M 3 contacting 3. Seepage into drifts H 2 waste 4A. Effects of heat / excavation on flow (mountain) Mn 1 , packages 4B. Effects of heat / excavation on flow (drift scale) Ma 2
- 5. Dripping onto waste package M 2
- 6. Humidity and temperature at waste package La.c 5 Long waste 7. Chemistry of water on waste package M 3 package 8. Integrity of outer carbon steel WP barrier M. 4 lifetime 9. Integrity of inner corrosion-resistant WP barrier H..a 3 Low rates of 10. Seepage into waste package M 3 i release for 11. Integrity of spent fuel cladding H. 3 radionulides 12. Dissolution of spent fuel & glass waste forms Mac 4 from breached 13. Neptunium solubility Ma.c 4 waste packages 14. Formation & transport of radionuclide-colloids Ma.c 2 Concentration 15. Transport through and out of EBS Ms.c 3 i reduction 16. Transport through unsaturated zone H. 2 transport from 17. Flow and transport in saturated zone M 2 waste packages 18. Dilution from pumping M 5
- 19. Biosphere transport and uptake L 5
- Subscripts: "a" = 0 to 10 kyr, "b" = 10 to 100 kyr, and "c" = 100 kyr to 1 Myr (no subscript = 0 to 1 Myr)
HARDIN-121698. PPT 12 '
Confidence Goal o The confidence level that is both feasible and desirable at the time of the LA : o Feasible means the goal can be accomplished by early 2000 o Desirable means the principal factor:
- - is significant to TSPA, and
- is important to defensibility o Goals need not be "high" for total system performance allocation i
1 ! i HARDIN-121698 PPT 13
o o o Confidence Goal Assessments RSS Attributes Principal Factors
- Signif. of Current Conf.
Uncertainty Conf. Goal Limited 1. Precipitation and infiltration into the mountain M 4 5 water 2. Percolation to depth M 3 5 contacting 3. Seepage into drifts H 2 5 waste 4A. Effects of heat / excavation on flow (mountain) Mn 1 2 packages 48. Effects of heat / excavation on flow (drift scale) Mn 2 4
- 5. Dripping onto waste package M 2 4 i
- 6. Humidity and temperature at waste package Ln.c 5 4 i Long waste 7. Chemistry of water on waste package M 3 5 package 8. Integrity of outer carbon steel WP barrier M. 4 5 lifetime 9. Integrity of inner corrosion-resistant WP barrier H..n 3 6 !
Low rates of 10. Seepage into waste package M 3 3 > release for 11. Integrity of spent fuel cladding H. 3 5 radionulides 12. Dissolution of spent fuel & glass waste forms Mn.e 4 5 from breached 13. Neptunium solubility Ma.c 4 5 waste packages 14. Formation & transport of radionuclide-colloids Mec 2 4 Concentration 15. Transport through and out of EBS Ma.c 3 4 , reduction 16. Transport through unsaturated zone H. 2 5 i transport from 17. Flow and transport in saturated zone M 2 3 waste packages 18. Dilution from pumping M 5 5
- 19. Biosphere transport and uptake L 5 5 i
- Subscripts: "a" = 0 to 10 kyr, "b" = 10 to 100 kyr, and "c" = 100 kyr to 1 Myr (no subscript = 0 to 1 Myr) '
HARDiN-121698. PPT 14
O . . O O . Priorit.ization Results - RSS Attributes Principal Factors A Current Conf. Priority Conf. Goal Limited 1. Precipitation and infiltration into the mountain 4 5 1 water 2. Percolation to depth 3 5 2 , contacting 3. Seepage into drifts 2 5 3 j waste 4A. Effects of heat /excavction on flow (mountain) 1 2 1 packages 48. Effects of heat / excavation on flow (drift scale) 2 4 2
- 5. Dripping onto waste package 2 4 2 L
- 6. Humidity and temperature at waste package 5 4 08 Long waste 7. Chemistry of water on waste package 3 5 2 !
package 8. Integrity of outer carbon steel WP barrier 4 5 1 lifetime 9. Integrity of corrosion-resistant WP barrier 3 6 3 Low rates of 10. Seepage into waste papage 3 3 0 release for 11. Integrity of spent fuel ciudding 3 5 2 radionulides 12. Dissolution of spent fuel & glass waste forms 4 5 1 from breached 13. Neptunium solubility 4 5 1 waste packages 14. Formation & transport of radionuclide-colloids 2 4 2 Concentration 15. Transport through and out of EBS 3 4 1 reduction 16. Transport through unsaturated zone 2 5 3 transport from 17. Flow and transport in saturated zone 2 3 1 waste packages 18. Dilution from pumping 5 5 0
- 19. Biosphere transport and uptake 5 5 1 0
^ Emboldened factors discussed in detail. 8The calculated priority for this factor has the same meaning as zero.
HARDIN-121698. PPT 15
o o o Seepage into Drifts Significance of uncertainty to TSPA = High o Dose rate uncertainty of > 50X is associated with model resolution and property estimates Current confidence assessment = 2 o Current models simplify postclosure drift geometry & neglect film flow diversion o Current model calibrated against few data Confidence goal = 5 o Planned or ongoing field tests will measure seepage and UZ flow behavior in the middle nonlithophysal and lower lithophysal host rock units Priority = 3 __
integrity of Inner Corrosion Resistant ! Material Waste Package Barrier Significance of uncertainty to TSPA = High o Dose rate uncertainty of > 100X is associated with the : current CRM model Current confidence assessment = 3 o Model may be too conservative in some aspects (e.g.,
- general corrosion rates) o CRM failure modes may not be comprehensive in TSPA Confidence goal = 6 o Ongoing laboratory testing of WP materials o Mechanistic modeling of material degradation o Specification of waste package environment Priority = 3 HARDIN-121698. PPT 17
3 e o ! Transport through the Unsaturated Zone Significance of uncertainty to TSPA = High o Peak dose rate uncertainty of 1,000X with different Unsaturated Zone and Saturated Zone dispersion assumptions (especially for 0 to 10 kyr) Current confidence assessment = 2 o UZ mixing / dilution assumptions may be non-conservative o Coupled process effects on transport are not incorporated in TSPA Confidence goal = 5 o Ongoing transport tests at YM and Busted Butte o include simulation of coupled process effects . Priority = 3 _ . , m. ,, ,e
- - - - - - - - - - - - - - - - - - - - J
o o o Effects of Heat and Excavation on Flow (Drift-Scale) , i Significance of uncertainty to TSPA = Moderate o Estimated peak dose rate uncertainty < 30X o Current model assumes that THC effects on flow properties will be similar to observed spatial heterogeneity , Current confidence assessment = 2 i o . Field thermal test data are for the middle nonlithophysal ! unit o THC effects are not included in current models Confidence goal = 4 o Planned field testing will evaluate the lower lithophysal unit o Effects of evaporative processes on flow properties are being simulated, and evaluated in lab experiments Priority = 2 e eo,~ ,2,eee.eer se
Chemistry of Water on the Waste Packages i Significance of uncertainty to TSPA = Moderate o Studies show low sensitivity, but corrosion processes l not in TSPA could be important i Current confidence assessment = 3 o introduced material effects may be important
- o Coupled processes may strongly affect inflow
- composition
- Confidence goal = 5 o Model development and supporting lab tests, will support prediction of water composition at the WP o WP material sensitivity to the bulk chemical l
environment will be reevaluated i Priority = 2 _ _,, e
o o o i Flow and Transport in the Saturated Zone i Significance of uncertainty to TSPA = Moderate o Peak dose rate uncertainty < 50X for the current model (dilution at pumping well held constant) i Current confidence assessment = 2 o Flow-tube model is probably very conservative
- o Characterization data gap (5 to 20 km) i Confidence goal = 3 o Ongoing collaboration with Nye County o Model refinement and review of hydrochemical data o Goal is constrained by schedule for SR/LA Priority = 1 HARDIN-121698. PPT 21
O O O Summary o Technical work needed is identified using prioritization of principal factors o Considers postclosure performance and other factors such as cost and schedule e Prioritization of technical work is essential for focusing efforts to achieve goals for SR/LA o Technical work will continue in key areas for performance confirmation HARDIN-121698. PPT 22
-----_-----------------------------------------_-------_------_-I
c~7 O O O , c YUCCA MOUNTAIN PROJECT Connarative Ana yses to Determine Sensitivity of Uncertainty for Princi;pa.. Factors Presented to: Advisory Committee on Nuclear Waste 105th Meeting Presented by: Dr. Holly A. Dockery Deputy Operations Manager, Performance Assessment " CRWMS Management & Operating Contractor U.S. Department of Energy Office of Civilian Radioactive December 16,1998 waste Management
o,. o o Purpose of Comparative Analyses o Primary goal is to determine the sensitivity of the TSPA results to the uncertainty in the parameters and the models o Two types of sensitivity studies were performed on the TSPA results
- Regression-based sensitivity analyses "One-off" sensitivity analyses DOCKERY-121698 PPT 2
n p% ~ Sensitivity Analyses for TSPA Attributes of Principal Factors Heterogeneity / Uncertainty Uncertainty the Repository Variability in Addressed in Addressed in Safety Strategy Base Case Esse Case Comparative Analyses Precipitation and infiltration of water into the mountain / / Limited water / contacting Percolation to depth waste packages
/ / /
Seepage into drifts / / / Effects of heat and excavation on flow / Dripping onto waste package / Humidity and temperature at waste package / Long waste Chemistry on waste package / package lifetime integrity of outer waste package barrier / / Integrity of inner waste package barrier / / / Low rate of release Seepage into waste package / / of radionuclides integrityof spentfuelcladding / from breached
/
waste packages Dissolution of UO2 and glass waste-form / / Solubility of Np-237 / / Formation of radionuclide-bearing colloids / / Transport within and out of waste package / Radionuclide Transport through unsaturated zone / / concentration Transport in saturated zone reduction during / / transport from the Dilution from pumping / waste packages Biosphere transport uptak9 / / DOCKERY-121698. PPT 3
o o o". Regression-Based Sensitivity Analyses o Purpose is to determine which parameters contribute the most uncertainty to the TSPA results o Performed on results of probabilistic base case run, sampling from all uncertain parameters simultaneously
- All parameters retain their assigned range of uncertainty - Interactions among the various parameters are maintained o Suite of results examined using: - Scatter plots - Regression analyses - Contributors to variance DOCKERY-121698. PPT 4
TSPA Regression-Based Analyses for 10,000 year Performance Period Most important Uncertain Variables for the Base Case 10,000-year Period 10,000-yr Dose-Rate History , ! t All Pathways,20 km . [] l ! 104 N' /Nd O ! O Number of pverule fadures p
~f .
j 10 I
] .g: } M Mor22 corrosen rate
-r
]
E ] _. a seepage fract,on S T ! P ') # i ! l E 10 WP s I V A
- -r"
_r [Mmm ,- LF g 10-2 -r- . T T _ g o* 10-4 -r a " . . - .. - D '. :.' o
-r
~
n ' - . O 10-s _ j q ; ~.f.' - - ;1 ._ ... _.,.1.. T~ \ i j y;3 -
? p, Q.p.W.;, fjiyqhq g:-g;_ r vy i;-yyg ;; [L?
; .g}: b;;;;w: .
10'g , , , ; . o 2000 4000 6000 8000 10000 .' 4.'gIMsjk?gi?:gy lt./;f?: ;Qgi;M..;'.?O... f. . /.. . M.~" Time (years) -.+AxNraths.AMhdaa
, . . ; i hnC> M c]'M ;
i 28 realizations have no waste -
/
// ' "-
- _i _ _ ' .
package failures and no doses These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings. DOCKERY-121698. PPT 5
o o o
~
TSPA Regression-Based Analyses for 100,000 year Performance Period Most important Uncertain Variables for the Base Case 100,0Cayear Period 100,000-yr Dose-Rate History , All Pathways. 20 km O Number of juvenile failures O Moy-22 corroson-rate vanabety 104 , , , . . a moy-22 corrosa> rate y : _r - ' ' '- ' -*--i--'7 i
- i a se.p ge tracnon j 102 -r ' - - - -- -- r - -
_ _ a e -r" ~ - - '- " l E 10 -r
$ 10-2 .
C T
~
, , , _ = - - <
c 10-6 -r
~
--4 -t---' 7
_r
'-b-'-- h~ *~~
'l ;
10-8 , l ; O 20000 40000 60000 80000 100000 1 - Tirne (years) /
. . ;. . , . l. .'. . ;. .'.
20 realizations have no waste l package failures and no doses t These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings- DOCKERY-121698. PPT 6 i
O O O TSPA Regression-Based Analyses for 1,000,000 year Performance Period i Most important Uncertain Variables for the Base Case i 1,000,000-year Period j 1,000,000-yr Dose-Rate History Alt Pathways,20 km -
<=
'l ] ] l 0 A%22 conoson ra!e I O Dose converson factor for N;>237 10 4 , , . . . i ;
a saturated zone asuten tactor e _[_ ._4 . t a seapeo tract:on f102 e
.-rr i- .
~
~
@ 10 -
L-t
~
j 10 2_ _ _ g _ -v q e 10 , ; j l g . _ . . 3 ._ __ .._p. , q Q 10 - I 4- - i _ ---. . . -- .. . . - . . . - . - . - . . . - 4 10-6 . ! ! L 0 200000 400000 600000 800000 1000000
. _ _ _ _ . i___ _
t wnp.a n p.m.oo venanc. These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These resu!ts are model-specific and may be insufficient for future licensing proceedings. DOCKERY-121698. PPT 7
o o o TSPA Regression-Based Analyses for Modified Parameter Case - 1,000,000 Years Most important Uncertain Variables in 1,000,000-yr Period (Removing Uncertainty in WP Degradation & Seepage) 1,000,000-yr Dose-Rate History
^ ~ " ~ ~ ^ ' - ~ ~ ~ ' ~ ~ ~ ~ ~ ~ ~ ~ - ~ ~
(Removing Uncertainty in the WP Degradation & Seepage) AH Pathways 20 km 104 , , , ,
"'"*"*"""""'""*""d"
, i r _[ . . _ ._ _._. .. ; a ..___.._4... _ .
_q = % wa=*== ~ i.*a D 102
'"~~ ~
E _r- ose *=. - m 100 + 5
"**~ ~ ~
- _f :
7_._. ] . _. h .. .__. _ j _ _ - q $ 10-2 _ _ _ _.{. _ _ _ j j . _ .. _ _ q g _
...,. g _ __ . . . . . . . _
q g 10-4 --;
-r-- -
~ ~
; -- -- -i O '
' t' ~
~~~t --~
'~l '
l o 10 - -
!- - - - - - - - 1 - - - - + - -
i ! i _ _ _ j. _ l._. _
.y - _ _q 10-8 ! ! ! !
0 200000 400000 600000 800000 1000000 Time (years) ,/ o ai o'2 o'.s o'4 o'.s imp.a n -+ - These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future Ucensing proceedings. DOCKERY-121698. PPT 8
. . _ . _ _ - _ - . . _ _ - _ . _ _ - - _ - _ . - _ . - _ - .-. - --_0
e O O . Benefits and Limitation of Regression- , Based Approach o Benefits of Method
- Indicates which parameters contribute the most uncertainty to dose rate '
- Maintains important synergistic effects among parameters
- Helps identify those parameters for which work to decrease uncertainty will be most useful .
t o Limitations of Method j
- Computationally intensive, thus limiting the number of sensitivity studies possible
- Can not be used to assess the relevance of "certain" '
parameters or of conceptual models
- Appropriate levels of uncertainty must be assigned carefully ,
to ensure that the true importance of the parameter is illustrated occKERY-12tGea PPT 9
O O O . t "One-Off" Sensitivity Analyses o Purpose is to assess how the uncertainty not easily captured in a probability distribution function affects i the results o Performed using deterministic base case results
- All parameters fixed at mean value, except for parameter / factor being investigated
- Factor is varied in different ways, depending on the .
assumptions used to conceptualize the factor l DOCKERY-121698. PPT 10
,l .!: l(;;! :II- il ll; l!!!
hJl. ; t!u! lli _o 1 1 0 _ 0 T _ 0 P 0 P. _
~
~
;H 0
0 1 8 9 6 1
;h1Ta "*
0 2 0 s 1 - 0 - Y 0 _ 0 R e - p E 0
- - njg y K c) _ *" C _
. 0 _
0 O _ s -
- 0. r)s 0
D - ge
. ' 0 .a 6
y (y n *v { r~~:E 4 q_*"l-e 0 l 0 t a -= _ 0m 0. 0 i T
.Reso - 0
_ a ,:lo1gq:g, 4a D ,J - up
~
4 0 n
,1 0
T 0
,j,c.&m__
##-rT'-f r 0 At_pa__wy y 0
. 2 A e . . go 3 .
i o T _ LMe 6 o [ 0 Ra
- o yLi' :
yt A e' l. l7J l 2 0, gi;;
= n_ 4is a
- ,ht_j r
t v c i - 7
/
vR .i =.t . . 4 - - o t. -
%r
- 0 3-i i i-O i 0, 0, ~ P
- R . t =- Cg s n E o i so
,; [:
.J4j,0 0,
0, 0 0
.e.N=*e:I1:a I
is gi nde uce sr or n _ ni =ls i ;o ep e a t t eg min _ lo i n s Sr t ke 0l ic 2 e _ f l t r au eu t f i dr d r tf a r u roes e r f o _ On l n s ders e ef st on I o e a o a ege t a de laf i ic e i c h c er du uf _ ta de is r er sata ivn gve r n sge t s cakw di ne i b sy eis a ir ecay l ya O a t ha xo sr n re .n c pb i w m _ a f hd t n _ r sb , 0e a a pa A n e n i c nn 1 s l ic oye laf _ P oma i i h deitn .o e i i, fode ~ d ne ap t s n-ic S t t a t ss agec l rt s e s0 0 ele sd e r o pm _ T r3r e aeoaba es s se 0, e r re s ea urcpmt t r ss o l i e~w c e n ul 0 yl la s t u f I n blo n cpsnc aeeuo a r0 Co1 n are ee ss I ee hh TT
~
O O O . TSPA One-Off Sensitivity Analyses - . . Seepage o Seepage fraction and e4 g,,- i==-vr =- a= ; ' seepage amount are ve , a 2 -+, H . uncertain functions of g' ,0 4! a:x ;; ; percolation ,? /W* C tjV*~ o Causes several orders of . n i=.a+vr=-a H magnitude effect on dose I,'7 Il; du t+ ? for 95th percentile. No j ,,,.?.,g - ^ ki . seepage yields very slow g kou ox ' __1 1
/-
corrosion of Alloy-22 ,,. 7 ; ! i ! t o' 7 ,6 M g I -j-- gse 33 jr!+J Qyyy]
/
gir e-f 3 +--{ j1'y /- 1,000,000-yr Dose Rate bse- N su.
- ; Expeded l ]
3g a. ' - 0 200.000 400.000 600.000 800,000 1.000.000 h (yea 3) , These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings. DOCKERY-121698. PPT 12
o o o TSPA One-Off Sensitivity Analyses - Alloy-22 Degradation Rate
- Both mean Alloy-22 ., , , , ,
degradation rate and 3::)
'; " ". *7"* R variab.lity i are uncerta.in i.r, = = !_ _ =<,
jw
- 1
-,- W, i
4 l ; Wide range of doses '"*" =
% -d
/several orders of - b' 1 p. m."*vr a- aadi magnitude) due to j" number of waste 1:q gsg:p; p ;pq l' packages failed and fraction of waste " y: jp5-
=
4.o..y,s-n W package surface I :: O f M y . - /- t- t-
,er----
_ __ g,,c H 1 degraded 3;; h p - ;- J i l l ! , ! i
$ = = = = == uxuoco These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. -
These results are model-specific and may be insufficient for future licensing proceedings. DOCKERY-121698. PPT 13 j
O O O TSPA One-Off Sensitivity Analyses - Cladding Degradation
. Cladding degradation ':p 10,000.y, oo.e ma,e l
is uncertain _1,'::Lg[!@ggy j j
, , ,7 j 'o' r 1 '
H - 1a; - Bounding
. ,0 ' -
N -, , e.; N - assumpt. ion of all '"~ 100,000-yr Dose Rate 7 cladding failed
. f I,
1 m
- r i ;
_ g; i ; increases early s'- r +@,r5 , doses by > 50 x '{INEI$@s75; oN20.000 40.000 e0.000 eg;xed.0m
\'"! 17 ;
j j
--w, i j
}
19 - -
- su A - n,.
3 . h '" ~ I
}~ 'n $W , 300 'l 2 '" - ~ ~i' 'r_ L*Sg",',g 1 j io - --p q ;
'o'- -1,000,000-yr Dose Rate 3 10s O 200 000 400.000 000.000 000,000 1.000.000 m= (yaws) i These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. ;
These results are model-specific and may be insufficient for future licensing proceedings. DOCKERY-121698. PPT 14 , f I
! ! .l ti ;' l! (i, :rl !l!;l,! -
4 5 o 1 0 T 0 0 P
. ,i 3i 0 0
P.
- 0 8
- 9
- H; 1 0
0 6 1 2 s 1 i . 0 s. o 0 Y c
- o. " 0 8 R e - E 1
1-1 _ i-,J~- .* {. E : c o K s ,
=
L,e ' D a_M,~,il::M' o y ~
- I-
"r o
,i a l
- oo C
_ r o o o. d l Q-o
" 0 a ,
=
ibi r v, a> ,,;pl-0 - '
- 00 2
0 0
'_t.,fp-0 n a
=.
0" 0 4
,,F l/
" 9, +5+.4 W ;.__
0 _ A , - i lim o o 0 o , a 1 s 1
; o y3g1 y y
,r v i
=. wr
.+,;',7\m r
_ itt i.! o s ,
,o ,
i
- = _9; ]
_ vi 0 ii l
,+#t '::F1
'0 1 ','
3-P R . t _ ib f E~ j Cg s I n o su gi . ie nd se uc sr or _ nl ep e S ol a c t eg min r T lo s _ S e v s
, r o
e v i n ke 0l ic 2 e f pe r r id t r au eu t _ f N s e a e eg selaa n t t f a r r o O-ef tain st - sd nu y su de on omd i e u r e lafic . e at pn i 0 0e r gd dus du is uf ivn d i - n sg 0, a t xe sr pco in be sy ya O ya 0n t Ny px a w m _ 0ar ae e bxefu hd t 5i l t a pa n i m m 1 A <o l l ic . i bf o ~ ey md ye laf n ape ic _ P u sd e so0 s i t l l eroiv t s n-ele ors l t sd ed 0 eot e S ms r o pm _ T se pdr ie t s s us 0, 0 t r n l atn e c e v e re se ar ss o No t o Ad as 0 Cle 1 t o n Acia v d yl la s n are ee t u ss ee hh a * *
- TT
o o o TSPA One-Off Sensitivity Analyses - - Saturated Zone Dilution o Saturated zone '"
" i
. i .
dilution varies from [ ':{J-~~ 10,000-yr i i , i j Dose Rate - 1-20 3 ZEl+! d-! M o " Expected" value ~ 2 . N. , 'r~ - ,,,- ,
- 100,000-yr Dose Rate - l 1
i _{in73 7 ; q 1* r :.Ts l l j ~x ge 7 7 - y , , 3. -m
~
- 10. .
l- -
,5 }-
'"I
,e
/
- i. i .
1 "X TJ. W " i, -- b ~ t* y ,, 3 j" T/t
,7 4,e p-9%
I i 1
. 1,000,000-yr se Rate 3
I; ! l i l te - These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings.
"" #s i eE "" ""
DOCKERY-121698. PPT 16 3
! o o o Benefits and Limitation of "One-Off" Approach 'o Benefits of Method
- Computationally simpler, allowing more sensitivity analyses ,
- Can be used to assess the influence of conceptual models as well as parameters on the result
- Can be used to assess the influence of "certain" parameters o Limitations of Method
- - Can not capture how the relative importance of a factor changes as other factors or parameters are varied
- As a result, may not show important insight into interactions i
of factors DOCKERY-121698. PPT 17
l o o o TSPA Design Option '
- Sensitivity Analyses ;
- o Uncertainty in performance assessments can be addressed by conducting additional scientific l investigations or adding design features to enhance EBS performance i
1
- o Three design options are being investigated
- Backfill
- - Drip shield placed over waste package and backfill 1
- Ceramic coating on waste package and backfill 1
l DOCKERY-121698. PPT 18
~
O O O 1 TSPA Design Option Analyses - Alloy-22 Drip Shield with Backfill i i i i 2-cm Alloy-22 drip shield y,: E 'a a
- vr oa - a a ] i No D6se .
degrades analogously to j'; j l Alloy-22 inner waste package jH l i i i
,q , , i :
o Mild steel degrades under drip '"' o - ..- ..- ..- , shield by humid air corrosion ,, [~ ,
- ' * ~ r-- 100,000-yr Dose Rate Alloy-22 drip shield must 1
o t --i i i i a'-r -N6-Dosed develop holes before seepage i 3, , : ; 4 ; can encounter waste package I;;; .__j i f !
- i 1
!. i i o Alloy-22 waste package will "K vp"-
" fail" at same location as _ ,';
Alloy-22 drip shield j;;;4_,,,,.y,oo,,,,,,_; i 1 r_ ;;; M' i l 4==4 i i i o No waste package failures for g t _ l _e c! . t
~100,000 years I,dtt o
r7m 200.000 4co.ooo coo s_._, i soo.ooo 1.000.000 These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings- DOCKERY-121698 PPT 19
o o o TSPA Design Option Analyses - Ceramic Coating with Backfill o Ceramic coating protects , .
. i i underlying mild steel from d: U ' *" " """ ;
- humid air or aqueous }" ; l-NpDose" corrosion 1 1 , , .
i '" ,_ : i 1 - o Ceramic coating protected by e - eg; - - backfill , , _ io' -r- 100,000-yr Dose Rate , o Ceramic coating reduces i- , : . . i s- , LNo-Do:se4 , degradat. ion of m.ld i steel 3 ,_m-t-H l 3 ,_m-4 ! o Ceramic coating may fail by '? i t i ! i
" blisters" at several hundred "r_g; > " -
thousand years "]4_..
,,000.y, oose ; n.te ,
i +; - n -u a t ,o. ! -- o Carbon steel and Alloy-22 - .
; l-F u ; under the ceramic coating i ,o.:E.
a ,_ l= _ _o}eAy=l - continue to provide protect. ion ,,
' ! el '
o 200.000 000 s00.000 ,.000.000 These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. ! DOCKERY-121698. PPT 20 These results are model-specific and may be insufficient for future licensing proceedings.
l O O O
- Summary and Conclusions o
" Expected" dose rate of Yucca Mountain repository system using a reference design, measured at 20 kilometers l
- ~ 0.1 mrem /yr to 10,000 years
- ~ 30 mrem /yr to 100,000 years 1
- ~ 200 mrem /yr to 1,000,000 years I
- o Range in probabilistic dose rates
- ~ 0 to 10 mrem /yr to 10,000 years j
- ~ 0 to 1,000 mrem /yr to 1G0,000 years
- ~ 0.01 to 3,000 mrem /yr to 1,000,000 years DOCKERY-121698. PPT 21
----..j O O O i Summary and Conclusions (Continued) i o Significant factors affecting performance include i
- Degradation rate of waste package
- Seepage into drifts o Uncertainty can be addressed and performance improved by
- Adding additional design enhancements ,
i
- Conducting additional site- and design-related scientific
! investigations t [ i DOCKERY-121698. PPT 22 i
O O O . YUCCA MOUNTAIN PROJECT n - _n m. ..- Postc.osure Defense in De:pt: 1 Presented to: Advisory Committee on Nuclear Waste 105th Meeting Presented by-Mr. Jack N. Bailey Director, Regulatory and Licensing CRWMS Management & Operating Contractor k U.S. Department of Energy Office of Civilian Radioactive December 16,1998 Waste Management
~
O O O . Defense in Depth (General) o Defense in depth needed to assure safety when quantitative assessment includes significant uncertainties o is essential element of Postclosure Safety Case o Safety philosophy that employs multiple protective measures to ensure that failure in any one of them does not imply failure of the entire system o in general, defense in depth means
- Multiple physical barriers to assure safety in design basis events - Conservatism, redundancy, and diversity in system design - Other measures (e.g., QA, emergency plans)
BAILEY 2-12/16/98. PPT 2
~
O O O .! Postclosure Defense in Depth o Postclosure safety of geologic repository places special constraints on defense in depth:
- Passive system in which principal safety issues are exposure of waste to water and movement of mobilized radionuclides to accessible environment
- Postclosure safety is demonstrated through analytical predictions based on models containing uncertainty o The proposed postclosure defense in depth therefore relies heavily on multiple barriers
- Redundant and diverse barriers to the movement of water and radionuclides
- - Multiple barriers to address model uncertainty and limitations BAILEY 2-12/16/9fiPPT 3
O O O Regulatory Context i o Draft Part 63 states that NRC relies on defense in depth to assure safety when quantitative assessments include significant uncertainty o NRC advocates multiple barriers that act together in a way to enhance performance of repository system o DOE expected to
- Identify the principal barriers of the system
- Provide transparent assessment of contribution of each barrier
- Indicate the system performance enhancement provided by multiple barriers and whether system performance hinges on any single barrier performing as anticipated BAILEY 2-12/16/98. PPT 4
O O O j Evaluating Defense in Depth i i o Evaluate expected performance and identify principal barriers o Determine threats to performance--select . neutralizations o Neutralize barriers reflecting threats--determine contribution of each barrier to total system performance o Evaluate overall postclosure defense in depth provided by system BAILEY 2-12/16/98 PPT 5
lI! Illll [i! ll !! !iii:' l() !!:fi f
)I!!
6 T P P. 8 9
/
6 1
/
2 1 - 2 Y E L I A B - h t i r ey rl s e s - p BA r a an - D e tc n uio dt a n oi zl Ca r _ n 3 ue
.t N e g s i en n n ina o o r
e r m hei rCiar r z t o ns f gi r te mal i ia asp s s ae t t a euBa r ee Dmnt
.iiu D Dmo n r c hn Ta 4a x M
N e C f e em nr io mfr r e e e t eP Dto D 2 - g _ i n d t c e e e pcf r yie s r t xnir a EanB t eme rdl aoia p t a - l u l E ufrdi ae vPannc i r - a 1 P v E
; Step 1. Evaluate Expected Performance and identify Principal Barriers o Barriers identified in TSPA analyses o Barriers are features that inhibit movement of water or '
radioactive materials, e.g.,
- Overlying rock units (barrier to water)
- Waste package (barrier to water)
- - Cladding (barrier to water) I l - Waste form (barrier to radionuclide mobilization)
- Drift invert (barrier to radionuclide transport) i
- Unsaturated zone (UZ) radionuclide transport barriers
- Saturated zone (SZ) radionuclide transport barriers BAILEY 2-12/16/98 PPT 7
! ;!i r 'l !:8i il;if}Iilti! :! ii tlj!!ll-. ;! ; l :!'
_ 8 _ o T P P. 8 9
/
6 1
/
2 1 2 Y E L I A B _ h s rse e t iy _ p rl r a an BA _ e tc n uio dt a D n oi Ca r l z
.t 3u e i
n N en e g s na r no n
.o e
i hei t r mCia r r r z fo gins r s emai asia ep s t l t _ t a e uBa r a ee r c Dmnt u D Dmo
.i e
n i hn 4xa N c Ta M _ofe em nr i o mf r r ee e t e P Dt o D 2 _ n g d e t s i c r e pcf r e yie t xnir t _ a Ean t emeB aoia rdl a _ l u l ufrdip ae nc vPan a i E r P 1 _E v _o "
~
O O O ., Step 2. Determine Threats to Performance l 1
- o Evaluate threats to individual barriers o Evaluate " common-mode" threats, those involving more than one barrier at a time
- Common conceptual model
- Other common sources of uncertainty
- Failure of one barrier affects another BAILEY 2-12/16/98. PPT 9
,I lji ;!t !. ti!, !li;t I! i ; i !i f;[!ll, r !( ! !
- 0 - o 1 T P P. 8 9
/
6 1
/
2 1 2 Y E L I A _ h rs ey s e B _ t i p l r ra an BA _ e tc n ui o dt a D nz oi l Ca r
.t 3u e i
n N en e g s na r on n _ e s ihei m rCiarr z e mai t r fo gir tasia ns o s t a ep l
'a euBra ee r c Dmnt .iiu D Dm 3 n hn Ta 4xa N e C e em M nr mfo
_ e i of t r r e Pe e Dt o _ D 2 g _ n d t e c s r i e e yie pcf r t xnirt _ a Ean a t emeB rdl aoia l u l E ufrdip ae vPan r nc i a 1 P v _E o g _ 4 ,
o o O Step 3. Conduct Barrier Neutralization Analyses o Approach: Conduct analyses where, one by one, barriers involved in threats and common-mode threats are " neutralized" (do not inhibit movement of water or radionuclides) o Analyses can be done with full TSPA methodology or simpler models--not strongly sensitive to model limitations BAILEY 2-12/16/98. PPT 11
o o o Example: Neutralize Overlying Rock Units I Overlying Rock Units Neutralized C E I T ! h O ! t : g j Base Case E r r Time (year after closure) BAILEY 2-12/16/98. PPT 12
! l .Ie !>tlt!,; !:'; rii ! it i( ! ,-!- iit'! t! }tiI:I! lt !!I; 3 -o 1
T P P. 8 9 - i
/
6 1
/
. 2
. 1 2
Y t E r . IL A e , B v I n , , - f t ' - i d r e z e D i l a s r a > t u C . , e N t e e s a B
)
e r u - i z r e v n l s o l i c oar t _ f i r - D r t e f a _ t u r a e e y _ ( e - N 1 i T m l e , - p , - m a , x , E -o
,5 : - - rE : - E:-
Cee i:- ,E i . i : . ewe uJOuF s- gscc4 l: . r:: -
! i
O O O Evaluating Defense in Depth
- 1. Evaluate Expected Performance 2. DetermineThreats 3. Conduct Barrier "E
and identify to Performance Neutralization Analyses Principal Barriers I
- 4. Determine Maximum Change in Barrier Neutralizations lI Data for Design comparison BAILEY 2-12/16/98. PPT 14
- _ _ _ _ _ _ - - _ _ - _ - _ _ _ _._ - _ - - , - - - - = - - = _ _ - - o
Sep 4. Determine $aximum Change in Barrier Neutralizations : o Use simple measure
- Maximum change in any individual neutralization analysis !
- Root-mean-square average of all neutralizations o Measure indicates whether any principal barrier is not compensated by others Compare measure to regulatory performance 1
o objective, e.g.,25 mrem / year in 10,000 years ' t o Use measure for comparison of alternative designs BAILEY 2-12/16/98. PPT 15
O
~
O O Summary o Approach leads to identification of principal barriers and contribution of each to system performance o Approach illustrates redundancy of barriers and resilience of system to potential threats to performance o Evaluates postclosure defense in depth of system within context of total system performance assessment o Results can be used along with expected performance, cost, and other factors to compare designs o Traceable and repeatable BAILEY 2-12/16/98. PPT 16
O O O YUCCA MOUNTAIN PROJECT _K Total System Performance Assessment Overview Presented to: Advisory Committee on Nuclear Waste 105th Meeting Presented by: Dr. Abraham Van Luik Senior Technical Advisor, Performance Assessment Yucca Mountain Site Characterization Office - 7 U.S. Department of Energy Office of Civilian Radioactive December 16,1998 waste Management
O O O Purpose , I o Show correlation of the attributes of the first element i of the Postclosure Safety Case with the TSPA model components o Briefly update the Committee on the conceptual models used in the TSPA o Discuss the TSPA results : t
! l VANLUlK-121698. PPT 2
O O O Correlation of Repository Safety Strategy Attributes and Factors, TSPA Model Components, and NRC KTis Attributes of the ' Reposhory Safety Strategy NRC Key Principal Factors TSPA Model Components Technicalissue Limited water Precipitation and infiltration of water into the Unsaturated and contacting waste mountain packages Unsaturated Zone Flow Precipitation to depth unk Mermal Conditions Seepage into drifts Repository Design Effects of heat and excavation on flow Seepage and Thermomechanical Effects Dripping onto waste package Thermal Hydrology - Mountain Scale Thermal Effects on Humidity and temperature at waste package Thermal Hydrology - Drift Scale Flow Long waste Chemistry on waste package Evolution of the package lifetime Near-Field Geochemical Environment Near-Field Environment integrity of waste package outer barrier . Container Life and Waste Package Degradation Integrity of waste package inner barrier Source Term Low rate of Seepage into waste package u Ides fr Integrity of spent fuelcladding breached waste Dissolution of UO2 and glass waste form Waste Form Degradation packages Solubility of neptunium-237 Radionuclide Mobilization and Formation of radionuclide-bearing colloids Engineered Barrier System Transport Transport within and out of waste package Radionuclide Transport through unsaturated zone Unsaturated Zone Transport Unsaturated and concentration Transport in saturated zone Saturated Flow reduction during Dilution from pumping Saturated Zone Flow and Transoort under Isothermal transport from Conditions and t e waste Biosphere transport Biosphere Transport and Update Tran VANLUlK-121698. PPT 3
O O O I i Discussion o TSPA model components were developed to represent various YMP process models produced by the site
- characterization & design organizations o Principal factors represent various aspects or i parameters used to model the process o Disru otive events, included in the TSPA, are correlated to a different element of the Postolosure Safety Case VANLUfK-121698. PPT 4
a -u- 2 me a m--msn*ama-A u e i a E2 cc II g li elga l Im g F o e gh, 3: O. C A kx? 1' EO O'1' *3i g$El4N- ' 1 ,[ , ! ] 4
,8- 9 9 ,, /
<r
]$ I 22 6f / % l , p' s= *t1 fbtff 4BZ1 j p ;,
gj i V ' Q /; $ /fijQ gh a b-O
i: i i !: ii: !lj jl1! t !!!l! . I:$ ; - 6 o T P P. 8 9 - 6 1 2 1 K - l U L N _ A V n i a gm - ot e t s A dy e s - P se _ S uht _ T e rf s s _ f e oesy e _ oos wr si ol al sy a _ t dvn a t r - l o a e a - u h h e s e s s - s t eb a c a e e mee c s y _ R t h a s el s a a nt b b an en c r oice e t i v f f ionl di t nisi i st i i t a r - l mb p ar a a a r mteb m er er o o - voDPC o Sin ef
O O O Use of TSPA Deterministic Base Case o Also called the " expected value case", the deterministic base case was a single realization o Uncertain input parameters were sampled at the mean of their range o The usefulness of the base case is to illustrate the relative influence of various components or subcomponents on individual dose values o A deterministic base case would not be used to assess system performance against a regulatory l requirement vANLUlK-121698. PPT 7
- - - - - - - - - - - - - - - - - - - - _ - - _ - - - - - - - - - - _ - - - - _ - - _ - J
o o u e o Deterministic Ana ysis Results
! I r i r
1 y T~ 10,000-yr Dose Rate !
'I l !
! 10' - I l
]
3 g aa . - Tow y I __. "Tc
! i !
3 10'-r -
~~ '"' ;
10 o \ 2.000 4.000 s000 s.MO. coo N Tim e (years) 10' , , ,
+
to* , - - - 100,000-yr Dose Rate , ! row ' i 7 108H :
"% {
3 Te 2 i l 5* ] Z l"4
; -t-
-d a to' y '"Pu '
)
1 qy., 10-' g ! . a
- 1T 10 s
] f; - i !
--j> >
i Inhalath>n Dose 0 20.000 40.000 60.000
' 80.009AN.000 Ingestion Dose Extemal Dose Time (years) -- few
&r -
1,000,000-yr Dose Rate ._ Q* ar,e, i o. . Yucca Mountain ' 102 q--- --y North
-A :
/ N*%
20 km ' ' 8
- 8'h*
! 10'- r- N S'^ ^ ~
Ma _ ? , o. _i __ y' ?-^~'\ L l a>%_g_.e.w.......s ..ss pm x
^l A f ,
,,. i ! i!
WM khhhhhh,hb!bN
- W N Y, , ' 2" "'
'"ra'"e c r'e"m"> ' '" "' ' '"' *"
====4 ,s ,,,
%( , $ ' ' .'. ' , ' l ' n' '.
Staess-e These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and rnay be insufficient for future licensing proceedings. VANLUtK-121698. PPT 8
i Eff+ct of Various Components on Determin'stic Results , t 10,000-yr Dose Rate 3-- ...
"'** * * " P 10* ; , e 1 i
; ui..mn. o- "
gos _ } l -. - ~ . y> ,, . ings.t.a Des.
;-.t- - g ,. ,3 ,, --. q
- l j
{~ 5E 3or -r----- "- 4 l--" i
~~
- Totas _ i - ..Q M
-h 10' -r-_ _ - - -. )
,, I
'Tc ,
t
. == .- .= .= ==
e *
. 10'-r- - - *8 '
i- ,
-~
,l '****"
j o-, . -. _ _ _ . _ . ._ A -, a 30, _ . - _ _ _ _ . - _ _ a 10-8 ~ 3 0 2.000 4,000 6,000 8,000 10,000 Time (years)
; , , ,, ,, j ,,,, , !
I r bothTc and treg 3 ir
/ \ i ;;; };" :== ~ ~ ~ ~ ~ '
=.; i no sorption high solubility i u- r -
g g i :::1 r
"' y fast transport fast EBS release j llh ~
_i e- _- :: - aws
~ . . .
*s .". '. ' gm ,
n' w.
= }j,,l ,_ ,
.I _I
- .l . . . . ..
- -i::: ' '
1 .. !. m."' ,! v : 'EEE=,=.1, 2".' r % - - f ,, . ' i I 3 <y xs p ", f j ".' i= =.:= r i m. J
- a. ,,
.1., ,.> .
um A
. = , . -, , . . . . . i These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. ,
These results are model-specific and may be insufficient for future licensing proceedings- VANLUtK-121698. PPT 9 i
e o o Performance of Unsaturated Zone with Respect to Tc-99 Transport (10,000 yrs) i "Tc Release From EBS
- 10-*
_ NY
~
~ nE 108 -r- - - Z (( ! --4
; z!" j r i a toi - --
t--- F e I * [ny, - w j e
~5 ,0 e
p4, y l1 ,
' ' "Tc Trevel time la the UZ
>'Wa e
~g ll f
- k y ~ ~.-.
5 J l / l '
~
#42V "Te n.we.. From ur 1 /d
- e 0 8,, ,,, ,,,
f 10' , Time (years) l l k 2 Re en 1 3 10' r- r:::::' -
..Tc ! x:: .e g ,,, .r_ _ ..::::
l no sorption high solubility 3.
,10 I
7 -- L--
'~~Y, .p . fast transport fast EBS release E
10' [' " 'i : 0 2.000 4.000 6.000 8.000 10,000 Time (years) I t
+===* !
These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings. VANLutK-121698 PPT 10 I
O O- O Performance of Saturated Zone with Respect to Tc-99 Transport (10,000 yrs) Yucca Mountain !
< 20 km
- Bio 8 Ph"'
._ :4% ; y .
- 4&L iwy ' 4 -
pc , _ ,_ , - _g ' me p ' " 'W Sk =s'~sy\.. - n , ,,e s
. u . ;,=*e D
- 4:.. .,.a-o. ,-,[ o.g .y d.ei~,w o ,
t l el
.,ee.w -
r t
,,,,, ',--T -ee / / I^ 'l 0 e'T , / / /
- fr
~ ~ .
/ - ~* .
Tc-te tisse Flut at 28 km to * , re ,t. a ,re. 2 Saturated Zone Breakthrough Curves f oi ~ r' _g, x;, , , , ,
* - - - = * " - - >
z "****i
~ ::,".'
t iv ,- r:::' , r.4 e ier _ :llll:1 %m
! ,0
; 7_2!5! _g / 5 ** - M -
I-
~~
! ',e -- d yp a
"* {' c i , . Af x
! '" [-g--: j //r' Si to
- e 2.000 4 600 e 600 0 000 t> 1*4 ) -/ ,,,
0 2,000 4.000 6 000 8000 t 0.000 10.00s ,I Tirne (years) Tene (yms) 0 -- -+ -- ~ "- -- 100 1000 10000 100000 1000000 Twne (years) Te-99 Concentration at 2C km 10' *
**Te Concentration at the Water Table DHution F setor at100,000 years t0' 10 ' -r $
"' _m /
-:::: g ,e p ;
i f
# $, f &
i1
$ j I
P i JJ1_, Ad 5': // een E '" 10
,/
. 02 0 2.000 4 000 6.000 8 000 to 000 0 2.000 a.496 4 000 4 000 10 000 #
Time (yea 2) 00 - - - - - 01 1 10 100 E#ective Ddution Factor m These analyses rer. resent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings. VANLUlK-121698. PPT 11
o o o I Probabilistic Base Case for TSPA . 1 i i o The probabilistic approach uses linked deterministic models with their related parameter uncertainty , propagated using a Monte Carlo technique o Multiple realizations are used to define the range in dose rate j i i VANLUlK-121698. PPT 12 , 4 i
l , l. , ' ' !! ;l[ 6tf ik! lii!;fi';! ! : ill t!ll
. !j j!
- 3 1 O T P P. 8 ds 9 6 1 2 1 nie K U l aro L N A V - st es - t i aH - R e . 0 em 4 0 0 0 0 . 4 0 1 si - T 1
]i ; :l' .
_. \ hg , g so )r oT i m 80
= \( \
.\
\
j
/
y m 0 De
- 2 3-l j - !i ,.
0 e P _
- 10ti:
0 3 1 r m R .S t 1 f _ _- .)
].
. M* n=g
.Anae
,' ;l j
0t ( e Cg I n k j gi i 1 ar a beM nd ._Oa - 6 e y Q
~ M 0
e s ie suce sr or R ( fi ,l d , 0o o 1 e 0 e ,s d f
- 0 0i m
' 2 0
d io ui ep r e l a t ep eg I_f. u min 01 r 0T epr - Pe ,
- 4 j '
'0 I:
pr a , 10i d _ _. r a ye ilo ns n ae- v
- e y0 id ke c s ba e y-00 2
- n s
0i
~ 0 0
0 j M eM + _ 000 000
- 0. 0 0
, 0i 1 k a
2 t r l e f au ooD 000 . 0
-- 2 s 11 1 3 e eu t
- l
+
- -- 0- P t f 5 a
- - 1 r r
-- o
~
ef - n st 4
,.- 7i-rp' r r O r
- 0 on de 2 1 i 4 0 - 4 4 0 2 ic 0 0 0 0 0 laf oo it 1 1
[pF E$gQ 1 1 1 BED H 0 1 mO e i 0 1 5o daEShaE8 c 0 1 uf du is vn id i ib ne t sy un hd ya a w m bo n t a pa l ic ii lai f c rt n ape t sl a t s n-e le sd er o pm ie e re - D R s ra e ss yl t O la u n are ee s ss ee hh TT
O O O
- Dose Rate Time Histories for 10,000 and 1-Million Year Periods 1 04
- 102 - _a___m___4___-.___
10 - i a _ _. . h5 10-2 _
-r '
-- --- t
$ 104 - r~ - k g .{--
j o 104 - p-- - -- - - - - - - -
\
104 , , ! , 0 2000 4000 6000 8000 10000 Time (years) 10' g ; ; ; - i
-r
[10 e - g - e 10-2 _ 10-4 - y d. _ __ _.1 _ ___ __ o 104 -- -
- - - + -l - - --t--i j_ , _ _ _..__ q 10-8 . . ! : :
O 200000 400000 600000 800000 1000000 Time (years) These analyses represent an all pathways ind:vidual dose rate at 20 kilometers using ICRP-M. These results are model-specific and may be insufficient for future licensing proceedings. VANLUlK-121698. PPT 14
o o o i Time Variation of Statistical Descriptors of the Calculated Dose-Rate Distribution ' i A y ,0. .H rn r,% i_y l Curves generated from mean, median, ; 2 . _F __ F and 5th and 95th percentiles at each
} joi
,* -H_ _ ,.
_+ -, 4=qi time step (100 years), consistent with -
,,0, g,..?.__J_F=6_y, g-j methodology proposed in draft 10 CFR so..
_ g,_ _ _ M n Part 63 !; O 2,000 4,000 6,000 8,000 10,000 ! l me < veers > ,0
,'w-i , ,
. For 10,000 and 100,000 1 1-
_ , . - r, , T1 m year graphs,5th , g ,o' , - percentile dose rate is
# r 1
! 10"-
g20' r- tf 7"y---- 10.. .
- r---e- i--i b--- 1ud zero 10-' -
r r M--~;J- " - f [~l ! 33, r/ ."E~~ --t--- i -- ~+1 "-~ ~ 1 10,000 20,000 40,000 60,000 80,000 100,000 Time (years) 10 , , , , i ! __ L__ _l__ i 3 03 _ 10' 3 - wh"-Y ^_ _-^-- g 10'-r- T t--f i 1
; 30e .j ._.y rp- :- j-ep- _ q gio,q ._u_.u p_q l-gie. y,ey,L rn- __ q,y 108-104 ; -
rf ---
- -)
0 200,000 400,000 600,000 800,000 1,000,000 These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. * #*' j These results are model-specific and may be insufficient for future licensing proceedings. VANLUlK-121698. PPT 15
O O o , Percentage of Realizations in Which Specific Radionuclides Dominate 10,000-Year Period for Three Periods i, Other C-14 4% Zero dose 1-129 2P. 31*A Tc-99 100,000-Year Period 38 % Pu-239 Other 2% 1% l-129 Zero dose 16% 20% Np-237 Tc-99 1,000,000-Year Period 31 % 30s Pu O mer 8Y. #A 1-129 9% Tc-99 2% Np-237 79% These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings. VANLUtK-121698. PPT 16
O O O 1 Disruptive Events '
- o Third element of Postclosure Safety Case
- o Disruptive analyses in TSPA include the effects of:
- Igneous activity (direct release, enhanced source term, !
indirect effects) 4
- Seismic activity (rockfall, indirect seismic effects) ,
! - Nuclear criticality (in package and out of package)
- Human intrusion (stylized release to saturated zone) i !
i VANLUlK-121698. PPT 17 !
o o O Igneous Activity - Direct Release Dose Rates , i i o Only about 5% of all igneous events in the repository block result in direct ie1oo : releases 1 : . I 10-1 =-+ d-I - 'l - o Doses were assessed at a i sese case peak dose
=; : -- Volcanic direct release !
point 20 kilometers south of 510-2 o
+ -
1- J . the repository (location of ; EL l : l
" -" =
critical group for dose j'~' 7m tN ' " 7+
~
effects) 1
"" ' f** PP " ~
j 104 "i "i "i i "; "i "i "1 "i o About 10% of all realizations ' o ' ' o* 1o" 10- 10~ ' o~' 10 o' ' o' o 'o' with releases had a wind direction from north to south and resulted in a dose effect These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings. VANLUlK-121698 PPT 18 l I
O O O Igneous Activity - Dose Rate Time Histories for Enhanced Source Term Scenario -
- Consequence Analysis for Two Realizations i
- o Magma in repository drifts increases waste ,0 , ,
- package degradation _ , os .;_ ,
rate and alters chemistry { 202_; _ EL _ i e % of waste form - s ,0, ; . .. . _ , . - [
= S '
o Radionuclides become I a 10 -- r - Enhanced source term 1 ennanceo
-- Base case source rem > 2 both more accessible to i 10-' , -
+
groundwater and more 1 o~' T-
* ' - 2 '
easily dissolved
'~ : : : :
0 200,000 400,000 600,000 800,000 1,000,000 j o Altered source term * * ** 'W i used in nominal j groundwater calculation F I t These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. [ These results are model-specific and may be insufficient for future licensing proceedings. VANLUlK-121698. PPT 19 [ [ t ! i
o o O igneous Activity - Peak Dose Rate CCDFs for Enhanced Source Term D s S 1 00 .=- , , , ., ,
$ , i 100,000-year results E I !
g igi_7_.f___ j_ ____j _ j ! __ l ,
$ i i ! '
3 - Base case I h102- r- - - Igneous enhanced source . - - _- _ ~ _ , o (rnean probatxhty) - 6 l i i : 108-p= = . _ _ _- - n
! l g104 ,
o 104 10-8 10-2 10' 100 10' 1 02 108 1 04 U Peak Dose Rate (mrem /yr) S 1 00 ., , , , , , o , 1.000.000-year resu ; ! e 102 -r--? - [ 7 [- - L t ,
% ! i
' I 3 ) l
__,_q_y_ i i i h102 , _ ..j _ _ , o 1 1 ; 1
- Base case 3(103- --
-- - Igneous enhanced source --r-.--,
e (meen probatukty) E l i (. e i : 3 o 104 ' " ,' , , , i b 104 108 102 10' 1 00 10' 1 02 108 1 04 U Peak Dose Rate (mrem /yr) These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings. VANLulK-121698. PPT 20
o o O Seismic Activity - Number of Waste Package Failures from Rockfall with Time o Waste package failure is a function of robustness of u. 10 , package and size of fallen i a -- . - ,
+'
1 rock e ! j 6----- - - --- - -- pg o Rockfall only damages i l
!c . n o g!id '
packages as they degrade ] ,2--_ .m j .i i m significantly i 0 ; ; g. n,;g i , 3 Z 0 200,000 400,000 600,000 800,000 1,000,000 Time (years) These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings. VANLUlK-121698. PPT 21 1
i! ! ! : :i 7! :l:!! ,vi?!i i! !!ill ! I!.[ill; 0 0 - y ' 0, 0 2 2 - Oi t l l
' 0 0,
T P P. i l bf a _
' 1 9
6 2 3 1 - ak ./~ 1
/ / =- -' K l
U L
/- N bc o
- A 0 V
- rR 0 o !
- - l ~ '/
' ' ,l 0,
0 0 )s 1 r P o a e - f t -
- (
y o e n - 0T i e m - nu o enni o i' 0 0, 0 oDi s e s unr" so "^ o" r 1 - ar r ro cC" . i 0
-- - 3-
-- P R .
ur -- ' Cg s I n pl o . gi E nd 0 e i -
- 0 isue c Om a C
- 0, so 0 * ~ 6 4, 2 0.1 r r ep 1 0 o 0 0 t eg - oF s D i o2ao>g3EoO - min lo i n ke s - Ce - gs u 0l ic 2 t r au eu t e ar e tf a r o r - icv t yk e g ef st on de ic are laf i va uf du is _ i mtf ivn d i ne c P a a i b - t sy ya a - dl s w m Aet l l a n i t r a hd t a n pa c c s e l if lai c _ ia f k u y ne ap t s n-ele mW cs0 o r e sd e r o pm s r u0 r e se ar - f o c 0, ss o i yl t la u c0 nare s - Oe S No3 ee ss ee hh TT
' i i
o e O Nuclear Criticality - Time History of Dose Rate at 20 km for 10,000-year Criticality Compared to Base Case Biggest contributors o 1000 i t are 1-129, Tc-99, ! ,o,ooo-yeer criticeiity . g. !' e 100 r- --
'w;/g^>-
Np-237, and U-234. p !
- Base gesestartin at 15,000 years
/e g 10 r~
rv i E - 7 e 1 = lii ! .. ..v../ T - a o 0.1 r-- y -- - o ! o 0.01 r- :! .
~
0.001 ' ' ' ' 'i ' ' ' ' 10,000 100,000 1,000,000 Time (years) j These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-spec;fic and may be insufficient for future licensing proceedings. VANLUlK-121698. PPT 23
e O O Trend of Nuclear Reactivity and Waste Package Degradation with Time o Probabilities for various scenarios are expected to be low because
- several unlikely events must occur .
WP degradation i i for the scenarios to j occur ;
- eff
=
.5s c .
O c-4 k i Tirtie j VANLUlK-121698 PPT 24
o O O Human Intrusion - Comparison of Dose Rates from a Human Intrusion at 10,000 years Postclosure with Base Case i i o Uses assumption that waste is i directly moved into 300 i the saturated zone { i ._ sese case at 10,000 years ]' [ 'f , O.1 l p .- 2700xg, High 2 - o Shows range of a ' '
~
l
- Dissolution Rete y,
0.01 ' behavior from h.i h a$ : f - 550 kg/ Low 0.1 i oissoiotion
! - i Rate o.001 mass /h.ig h
,0000 ,5000 20000 dissolution to low o
' ;oo,a ' ;oo,a ' ;oo,a ' ;oo,e
',,000,000 mass / low *****
dissolution These analyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings. VANLUlK-121698. PPT 25
O O O YUCCA MOUNTAIN PROJECT Repository Safety Strategy Presented to: Advisory Committee on Nuclear Waste 105th Meeting Presented by: Mr. Jack N. Bailey Director, Regulatory and Licensing CRWMS Management & Operating Contractor H i . U.S. Department of linergy December 16,1998 'S[t$$*"((2'
O O O Overview o Present the evolution of the Repository Safety
- Strategy (RSS) and its relationship to the Postclosure j Safety Case
! o Describe the.Postclosure Safety Case o Review the role of the Repository Safety Strategy for the Site Recommendation and the License Application 1 BAILEY-121698. PPT 2
- - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ .- i
O O O i Repository Safety Strategy o The Repository Safety Strategy provides the framework for integrating site information, repository design information, and the results of performance assessment to focus on the information needed to complete the Postclosure Safety Case o Organized by key attributes of the repository system. l - Limited water contacting waste packages
- Long waste package lifetime
- Low rate of release of radionuclides from breached waste packages
- Radionuclide concentration reduction during transport from waste packages BA! LEY-121698. PPT 3
4 O n o e T P P. 8 9 o t s 6 1 2 1 Y sd a E tL A B e e C sdy u et _ c ef e _ FoNaS yn _ et u giroe t a s Otamo r _ rl Soo f t yn s o t _ fe eP I ah e St h y gt . r n e *S t oiyt e if l p w sit wo on pe m p edo O RI C
c ; Evolution of the Repos" tory Safety Strategy i - t L F Reposkory Safety Strategy * (Rev 2) i r, ,. ...m..._,. j * ' 19 PrinsIpel Festers . steupedesserdngtothe . Reposkory Safety Strategy (Rev 1) [' her eastbusse w use mysean , t - ' ' ~ * - - --- * "**** * 'h ** ! y - *- Principal attrthutes of synessa ) fj : * *estore s t priertileadto des'ne - ! t- . - ummmeusewesnessene== a. ' I **rk P8'"* r - .
~ . .; ' [E ,
f:
- ' n==mme==
- tame ==meeresamesame=" J r1 - "w , !
- 21. tourseenWmamase .. t' , -- - l' Reposkory Safety Strategy (Rev 0) * '
{. ; , c. e ,;,n,,,,, J * e,,s.,, , y m. - , , ~u , - t e ,,, ,1 ; dg p
- PreneipalettethWeectoyateen y ; j ,,,,,,,,,,, ,, ,gg,,,, P j ;
t.- 1-s J y- ' ., l Br w these earthuses e s ; ! y' ~~""*"'""**""*"""' ~;
' c 8' W W develop Sto ' ,~ I She Characterhation Plan s 1 '" possetesuresesseycase R ; . i
-^'""" '
; -. m ~"i T- ; Y '.,.c 7.. pie,,,ap,a .-m , ,,n,g 4 .
- . . ceases - ;
, .w ( a, s.w; .
-)
.,3 i .u. .
; / .+ -
t - -- . p;4 .unnessamme auman.n. e. f*..ciahyPeeheeseseevehemsethese 'k 1 ! L jy smassamesosanasse mor ' I _ martbutes 1 New inf.,rmation
@it p. mammessenesonst r- t , 4J ^ps i
, , i l'+
j
@.'hRmusemasemanneur ei ameM! 5l < t 1-I ~ ; .- - TSPA sensitivities using improved I g en., Gee.,es, m ,,,qs,. .,,,,e, 3
^
M ( - %c >'
. 1 process models
.M and pressesse.eWQ.P. .
s4 J~y. u. , ? jL . In situ test results Q,sg P
- m..n ,
- Materials testing data
- 7. - .e cp/m : 9 er xShnewm. 9:. , t:- !
pg.4.y .. New information 1 6-di@ E6.Dnt.M ' - Site data t I L
-TSPA sensitivities !
- Design maturation New site information 1 t
Dose-based standard Modified design 5 BAILEY-121698. PPT 5 l' i (
o o O
- Example of the Evolution of the ,
! Repository Safety Strategy l Rev 0: Hypotheses for Rev 1: Hypotheses for Rev 2: Principal factors seepage attribute limited water for limited water Low percolation flux at contacting waste contacting waste repository depth Package attribute package attribute Percolation flux at Precipitation and Limited fracture flow at repository depth can be infiltration into the repository depth bounded mountain i Capillary retention i Seepage into drifts will Percolation to depth reduces seepage into be a fraction of Seepage into drifts depth percolation flux ! Thermally induced flux Dripping onto waste ! can be bounded Thermally.induced packages seepage can be Effects of climate can be bounded Humidity and temperature , bounded at waste package i Seepage that contacts waste packages can be limited i BAILEY-121698. PPT 6
o o o Basis for the RSS: Comprehensive Understanding of the Natural and Engineered Features of the System t o Majority of radionuclides in the repository not i mobile in Yucca Mountain Environment
- Insoluble / sorb strongly o Remainder could be transported by water movement o Natural features are favorable for limiting water movement o Site provides predictable and stable environment for engineered features to further limit water movement BAILEY-121698. PPT 7
o o i o s Revision 2 of the RSS o Performance assessment sensitivity studies provided basis for testing hypotheses and identifying factors l most important to understanding total-system performance
- Hypotheses have been replaced by principal factors
- Potential to reduce uncertainties for those principal factors important to performance was used to prioritize technical work for the LA
- o Elements of Postclosure Safety Case summarized
- Establishes regulatory bases for showing the repository system will protect public health and safety i
BA! LEY-121698. PPT 8
o o o RSS Revision 2 and Principal Factors o Example: Hypotilesis 4, Seepage into drifts will be a fraction of percolation flux
- The seepage model fits preliminary data from liquid release 1
tests in niches in the ESF
- Preliminary estimates of seepage appear consistent with this
- hypothesis 4 ,
- Analyses continue to indicate that seepage is a principal factor in the evaluation of system performance
- Principal source of liquid water contacting waste packages BAfLEY-121698. PPT 9 I
i
- . - _ _ . 9-- . _ _ _-- - . - - -
9 RSS Systematic Evaluation of i Principal Factors q k i o Understand the required performance of the system ! o identify principal factors that affect the performance of i the system l o Perform sensitivity studies and uncertainty analyses to determine importance of the principal factors with j
; regard to overall system performance l
i BAILEY-121698 PPT 10 l
o o o RSS Systematic Evaluation of the 4 Principal Factors Limited Water Contacting Waste Packages Long Waste Package Lifetime
- Precipitation and infiltration into the 4 - Chemistry of water on waste package (
mountain -Integrity of outer carbon steel waste
- Percolation to depth package barrier
-Seepage into drifts - Integrity of inner corrosion-resistant
- Effects of heat and excavation on flow waste package barrier (mountain and drift scale)
- Dripping onto waste package
- Humidity and temperature at waste y
package a Low Rate of Radionuclide Release Radionuclide Concentration Reduction
-Seepage into waste package During Transport t
i
-Integrity of spent fuel cladding
- Dissolution of spent nuclear fuel and glass 5 -Transport through the unsaturated zone waste forms - Flow and transport in the saturated
- Neptunium solubility zone
- Formation of radionuclide-bearing colloids - Dilution from pumping
-Transport through and out of waste .
Biosphere transport and uptake package BAILEY-121698. PPT 11 1 _______._.____._______________.__m.______ _ _ _ _ _ _ _ . _ _ _ . . _ _ _ . _ _ _ . _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ . _ _ _ . - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ . _ _ _ _ _ _ _ _ ,
0 0 O RSS Systematic Evaluation of the 4 Principal Factors (Continued) o Significance of uncertainty in principal factor or its absence in TSPA
- High: results in a factor over 50 increase / decrease in peak dose rate from expected value - Medium: results in a factor of 5 to 50 increase / decrease in peak dose rate from expected value
- - Low: results in a factor of less than 5 increase / decrease in peak dose from the expected value l
l BAILEY-121698. PPT 12
Evaluation of Potential Design Options t l i ' o Design options and alternatives that could improve j system performance and reduce remaining uncertainties are being evaluated to support initial ! design selection for site recommendation and license application :
- o Preliminary analyses of backfill, drip shields, and t
ceramic coating showed the potential for improved
- system performance i
BAILEY-121698. PPT 13 i
. 7
- 9 0 Preliminary Analysis of Sensitivity of Postclosure Performance (Dose Rate) to Design Options 1,000.000-yr Expected-Value Dose-Rate History Average Individual, All Pathways, at 2G km 10- , . . ., ...,...,... . . . ,
1 03 ; -
--w------------ - - - - - - - - - -
-- -- - - - -- - 4 5 i * !
i E 102 4- _ _ . . ._ _ . . . _ . . . 1 : E 10 - r-
+- 7 2+-r " ------ ---
-j h 100 - h $ -- - - - --
vVC=~~2k m ; / ,. - / g 10-1 -r l~ - ---i- -- - - o . Base case Q i i)-2 _[ e f./ _.- r-- Drip
- - Ceramic coating shield + i' 10^ l1I '
-l - -
l -l - - - 0 200,000 400,000 600,000 800,000 1,000,000 Time (years) These enalyses represent an all pathways individual dose rate at 20 kilometers using ICRP-30. These results are model-specific and may be insufficient for future licensing proceedings. Ball.EY-121698. PPT 14
e o o Postclosure Safety Case i i o The Postclosure afety Case will comprise the information that DOE intends to use in the regulatory setting to provide reasonable assurance that a repository at Yucca Mountain will adequately protect the public health and safety and the environment after ] . the repository is permanently closed l a BAILEY-121698. PPT 15
~ '
O O O . Postclosure Safety Case (Continued) . o Focus on elements needed to provide reasonable assurance that public health and safety will be protected j
- Assessment of expected postclosure performance and supporting evidence .
- Incorporation of safety margin and defense in depth to mitigate uncertainties
- Explicit consideration of disruptive processes and events l - Supporting information from natural and manmade analogues
- A performance confirmation plan BA! LEY-121698. PPT 16
______-_____-_-_________________________________-___t
~ ~
l O O O. Postclosure Safety Case (Continued) , i t L o The major focus for LA is postclosure safety o A large fraction of the radionuclide inventory decays rapidly or is essentially insoluble; these radionuclides ! pose no safety issue o A small fraction of the radionuclides is mobile , f BAILEY-121698. PPT 17
~ '
O O O Postclosure Safety Case (Continued) o Some uncertainties remain regarding the characteristics of the repository system as they influence dose rate o When these uncertainties are addressed, the Postclosure Safety Case will be complete
- Testing and analyses to address uncertainties in the natural and engineered barriers - Evaluation of design options and alternatives to select a design to mitigate uncertainties l
BAILEY-121698. PPT 18 i
o o o 4 l Assessment of Expected , l Postclosure Performance l o Remaining technical work emphasizes highest-priority principal factors o Update performance assessment models usmg improved process models and the results of the design alternatives study i L BAILEY-121698. PPT 19
o o o-Design Margin and Defense in Depth Design Margin o
- Select engineered barriers to ensure sufficient design margin
- to account for variability in material properties and uncertainty in conditions to which the barriers will be subjected
~
o Defense in Depth i
- Identify barriers important to postclosure safety and clearly identify their contribution to performance i !
4 l BAILEY-121698. PPT 20
o o o ; Disruptive Processes and Events . I o Updato analyses and complete documentation of information on volcanism, seismic hazard, and human intrusion i o Further evaluate the probability and consequences of postclosure nuclear criticality i i l l BAILEY-121698. PPT 21
g g g Natural and Man-Made Analogs o Compile and review existing information on:
- solubility and colloid-facilitated transport of radionuclides
- retardation characteristics of alteration minerals in fractures
- infiltration in the shallow unsaturated zone
- behavior of man-made materials BAILEY-121698. PPT 22 -
+ _ . . _ _ _ _ . _ _ . . .
o o o t Performance Confirmation Plan t I lo Revise the existing plan taking into account the design for LA l o Specify monitoring, testing, and analyses to address uncertainties in postclosure performance , i l, i BAILEY-121698. PPT 23
o o o I L l System Evaluation For the SR/LA l 4 Repository Safety Strategy
- Evaluate Design Features
- System Attributes
- Evaluate design attematives !
- Reference Design
- Integrate design features and alternatives into natural system '
- Pn.ncipal Factors a
I Evaluate Principal Factors
- Identify set of principal factors i
appropriate to each combination of engineered and natural features
- Importance to performance
- Importance of uncertainties
- Select design 4 SR/LA System i Design l
BAfLEY-121698. PPT 24 i i
O O O Summary j 1 o The Repository Safety Strategy focuses the work effort j o issued to: i
- report the evolving approach based on technical findings
- - update the licensing case t i
- a i
4 BAILEY-121898. PPT 25 __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ - _ - - _ _ - - - _ - _ _ - -}}