ML20151J585
| ML20151J585 | |
| Person / Time | |
|---|---|
| Issue date: | 07/23/1997 |
| From: | NRC ADVISORY COMMITTEE ON NUCLEAR WASTE (ACNW) |
| To: | |
| References | |
| NACNUCLE-T-0115, NACNUCLE-T-115, NUDOCS 9708050184 | |
| Download: ML20151J585 (260) | |
Text
~
~
~
Official Transcript ef Proceedings AcWiv7-alic D LNUCIEAR REGULATORY: COMMISSION Title- Advisory Committee on Nuclear Waste .
93rd Meeting j TROB (ACNW)
RETURN ORIGINAL TO BJWHITE M/S T-2E26
. Docket Number:- (not applicable) us-n3o THANKS1 .
1 Location:- San Antonio, Texas P;
42 ;
Date: . Wednesday, Jnly 23,1997 i 1
I l !
Work Order No.: NRC-1178 Pages 1-258 4 R DV M N N LE kkkkk i.
T-0115 PDR NEAL R. GROSS AND CO., INC. 4
. Court Reporters and Transcribers -
1323 Rhode Island Avenue, N.W. i yt ] f
.Y . Washington, D.C. 20005 j 0,i j "f,./ 5,.
O#nn5n i ,
o u
- e >
(202) 234 4433 V1 ll ACNWOFFICE COPY- RETAIN FOR QTflEUFEOFTHE00leiDTEE
Official Transcript cf Procasdings ACHWT-0//c' '
l Q NUCLEAR REGULATORY COMMISSION
Title:
Advisory Committee on Nuclear Waste 93rd Meeting TRO8 (ACNW) t RETURN ORIGINAL TO BJWHITE M/S T-2E26 ,
Docket Number: (not applicable) us-n30 THANKS!
Location: San Antonio, Texas O !
Date: Wednesday, July 23,1997 l l,
H i Work Order No.: NRC-1178 Pages 1-258 D DV M NA LE kk,k k,k .!* .
- l T-0115 PDR l NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribe. 2 -
I 1323 Rhode Island Avenue, N.W. (
Q O v ,so' t) n Washington, D.C. 20005 h
'~""})n3 ! 't L
C -) (202) 234-4433 y!(\; G ACNW OFFICE COPY- RETAIN FOR QTHEUFEOFTHECOMMITTEE
k
(/
DISCLAIMER PUBLIC NOTICE BY THE UNITED STATES NUCLEAR REGULATORY COMMISSION'S ADVISORY COMMITTEE ON NUCLEAR WASTE JULY 23, 1997 i
l The contents of this transcript of the proceedings of the United States Nuclear Regulatory Commission's Advisory Committee on Nuclear Waste on JULY 23, O
) 1997, as reported herein, is a record of the discussions recorded at the meeting held on the above date.
This transcript has not been reviewed, corrected I and edited and it muy contain inaccuracies.
l l
i i-i i
U #
NEAL R. GROSS Col'RT REIORTERS AND TRANSCRillERS l 1323 RilODE ISLAND AVENUE, NW (202)234-443t WASi!INGTON,DC. 2000$ (202)234 4433
1 l
1 UNITED STATES OF AMERICA g 2 NUCLEAR REGULATORY COMMISSION
, i s
%-) 3 + ++++
l l 4 ADVISORY COMMITTEE ON NUCLEAR WASTE (ACNW) 5 +++++
6 93RD MEETING 7 +++++
8 WEDNESDAY, 9 JULY 23, 1997 10 +++++
11 SAN ANTONIO, TEXAS 12 ++ +++
13 The Committee met at the Southwest Research
/~N b) 14 Institute, Conference Room, 6220 Culebra Road, San l
15 Antonio, Texas, at 8:30 a.m., B. J. GARRICK, Chairman, 16 presiding.
17 COMMITTEE MEMBERS: i 18 B. JOHN GARRICK, Chairman i
19 GEORGE HORNBERGER, Vice Chairman
]
20 PAUL W. POMEROY ,
21 ACNW STAFF PRESENT:
22 ANDREW C. CAMPBELL 23 VIRGINIA COLTEN-BRADLEY l
, 24 LYNN DEERING l
i
("..i 25 T.LY M. GARKINS V
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
l 2
1 ACNW STAFF PRESENT: (CONT.)
j ,f- 2 HOWARD J. LARSON l\
] 3 RICHARD K. MAJOR 4 RICHARD P. SAVIO 5 ALSO PRESENT:
1 6 RALPH ANDERSON 7 AMIT ARMSTRONG 8 KEN ASHE 9 JACK BAILEY 10 ROBERT G. BACA 11 MICHAEL J. BELL 12 DANIEL BULLEN 13 ASAD CHOWDHURG
('S V J 14 RICHARD CODELL 15 STAN ECHOLS 16 NORMAN A. EISENBERG 17 MARGARET FEDERLINE 18 EAN FEHRINGER 19 STEVE FRISHMAN 20 APRIL GILL 21 ALI HAGHI 22 CAROL HAULON 23 WILLIAM J. HINZE l
24 ROBERT JOHNSON f3 25 ROBERT JOLLY V
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
3 1 ALSO PRESENT: ' CONT.)
,S 2 JAMES LIN
( )
3 CHRISTIANA LUI 4 PAT MACKIN 5 TIM McCARTIN 6 KEITH I. McCONNELL 7 R.K. McGUIRE 8 SITAKANTA MOHANTY 9 PARVIZ MONTAZER 10 PRISCILLA NELSON 11 CLAUDIA NEWBURY 12 WES PATRICK 13 ENGLISH PEARCY g~.
\ t
'v' 14 FRED RODGERS 15 JOHN ROSENTHAL 16 BUDHI SAGAR 17 MARTIN STEINDLER 18 JUDY TREICHEL 19 RAY WALLACE 20 CHRIS WHIPPLE 21 JEFF WONG 22 RAY WYMER l
23 24 C) 25 l NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS
! 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
41 .
i 1 A-G-E-N-D-A 2 Acenda Item Pace 3 Opening Remarks by the ACNW Chairman 5 4 Overview of PA Activities within the Division of i 5 Waste Management 11 l I
6 Overview of NRC's Integrated Performance i l
i 7 Assessment Activities 11 8 Overview of High-Level Waste Performance 9 Assessment Activities, Priorities & Schedules 32 i
10 Allocated Staff & Expertise 60 l
11 Hardware Capabilities and Utilizations 63 l I
12 Overview of TPA 3.1 Code 76 13 Abstractions in TPA 3.1 Code 96 rs
(
()' 14 Sensitivity and Uncertainty Analyses of the TPA 3.1 161 l
15 Plans for the DOE TSPA-VA and Implications for 16 Further Developments of NRC's Review Capability, 17 Including the TPA Code 192 ;
I 4
18 Summary 220 )
l 19 1
20 21
- 22 l
l l 23 24 A
/ 25 C/ )
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
5 1 P-R-O-C-E-E-D-I-N-G-S 7_ 2 CHAIRMAN GARRICK: Good morning. Our meeting l
/
\
3 will now come to order. This is the first day of the 93rd 4 Meeting of the Advisory Committee on Nuclear Waste.
5 My name is John Garrick, chairman of the ACNW.
6 Other members of the committee include George Hornberger 7 and Paul Pomeroy.
8 We have two consultants with us today, well 9 known to the committee: Drs. William Hinze and Martin 10 Steindler, and four invited experts: Stan Kaplan, Jim Lin, 11 Chris Whipple, and Ray Wymer.
12 During today's meeting, we will review the 13 performance assessment capability in the NRC high-level fN
\_ l 14 waste program and prepare ACNW reports on igneous 15 activity, and we will discuss an upcoming letter on 16 defense of DEF.
17 Dr. Andrew Campbell is the designated federal l
18 official for today's performance assessment session. This 1 1
1 19 meeting is being conducted in accordance with the l 20 provisions of the Federal Advisory Act. We have received 21 no written statements from members of the public regarding l 22 today's sessions. l l
I i
23 Should anyone wish to address the committee, 24 please make your wishes known to one of the committee (s)
, 25 staff. It is requested that each speaker use one of the l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
6 j l
1 microphones, identify themselves, and speak with 7~ s 2 sufficient clarity and volume so that they can be readily
/ T
's /
'~'
3 heard.
4 Before proceeding with the first agenda item, ;
5 I'd like to cover some brief items of current interest.
6 First, Dr. David Morrison, the former director of the ,
1 7 Office of Nuclear Regulatory Research, has left the NRC.
l 8 Ashok Thadani has been selected to succeed David. l l
9 Robert Jolly is serving as a summer intern on 10 the ACNW staff. Rob is majoring in environmental science 11 at Virginia Polytechnic Institute. He will graduate in l l
12 December. !
l 1
13 On May 21, the NRC approved a final I
(~h >
l (s/ 14 decommissioning rule which will revise 10 CFR Part 20 to 15 provide specific radiological criteria for the 16 decommissioning of lands and structures at NRC and 17 agreement-state-licensed facilities and establish a 18 consistent regul& tory basis for determining the extent to 19 which lands and structures must be remediated before 20 decommissioning of a site can be considered complete and 1
21 the NRC or agreement-state license terminated.
22 The Commission adopted the 25 millirem per l
! 23 year all-pathways dose limit, coupled with a requirement 24 to reduce radiological doses resulting from residual
/m
( ) 25 activity to levels that are as low as reasonably NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
I 7
1 achievable, ALARA.
c 2 Rust Federal Services of Hanford has processed
/
3 almost one million gallons of high-level nuclear waste 4 during the first 1997 run of Hanford's 2.42 (a) evaporator.
5 As a result, an additional 360,000 gallons of double-shell 6 tank space have been made available for future use.
7 In May, an International Nuclear Regalators 8 Association, INRA, was established to influence and 9 enhance nuclear safety from the regulatory perspective 10 among its members as well as worldwide.
11 The association which plans to meet at least 12 annually has eight founding members: Canada, France, 13 Germany, Japan, Spain, Sweden, United Kingdom, and the U 14 United States. Shirley Ann Jackson, chairman of the U.S.
15 Nuclear Regulatory Commission, was elected to serve a two-16 year term as the INRA's first chairman.
17 There's some recent information regarding the 18 low-level waste contact states. First, California: On 19 June 26, Senator Murkowski introduced legislation to 20 convey land in Ward Valley to the state. The state must 21 pay $500,100 to the U.S. Treasury and commit in writing to 22 carry out environmental monitoring and protection measures 23 recommended by the National Academy of Sciences and 24 subject to NRC oversight.
,m 25 Texas: On June 25, the House Commerce NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W. l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
8 1 Committee approved an interstate low-level waste disposal 7y 2 compact among Texas, Maine, and Vermont. The Senate 3 Judiciary Committee had previously passed an identical 4 bill on March 20.
5 Ohio, Midwest Compact: On June 26, the 6 Midwest Interstate Low-Level Radioactive Waste Commission 7 voted to cease development of a regional disposal facility ;
8 in Ohio and to revoke host state designations. The 9 principal reasons: Believe there's sufficient access to 10 existing low-level waste disposal facilities and the 11 Commission is at a critical point immediately prior to the I
12 commitment of considerable funds to a site selection
_. 13 process in Ohio.
t i E/ 14 Illinois: Governor Edgar approved legislation 15 that delays all work while a two-year study of state and )
16 national siting issues is conducted. The target date for 17 opening the facility is extended from 2003 to 2012.
18 And, finally, a coalition of utilities called 19 the Private Fuel Storage Limited Liability Company filed a 20 license application for a temporary storage site for spent 21 nuclear fuel on the Goshute Indian Reservation near Salt
- 22 Lake City. Needless to say, there is considerable 1
23 opposition to this proposal in the state of Utah.
24 One additional thing that I want to say as the q_) 25 new chairman of this committee to reassure my colleagues NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
9 1 is that I want to acknowledge that just because I happen s 2 to be strongly identified with the risk-assessment
['~') 3 community and business and that this first meeting is 4 heavily oriented towards performance assessment which has 5 a strong risk overtone, is not to suggest that I'm going 6 to be obsessed with that subject and make every agenda 7 performance assessment or risk assessment.
8 VICE CHAIRMAN HORNBERGER: You mean you're 9 going to change?
10 CHAIRMAN GARRICK: As a matter of fact, our 11 committee -- and this was coincident -- our committee 12 planned this particular meeting, and it was coincidental 13 with the change in the chairmanship. I wanted to make
/-
$mv) 14 that point.
15 What we are going to continue to emphasize, 16 however, with great vigor is what goes into the 17 performance assessment, what is the information base that 18 constitutes the underlying knowledge that is being 19 processed into performance measures that can be a 20 reasonable basis for making a decision. That's what the 21 emphasis will be.
22 So with that, the purpose of the meeting today 23 is to evaluate NRC's PA staff expertise and modeling 24 capability that forms the basis for reviewing DOE's pre-C\
( ) 25 licensing activities and documents such as the total
( NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344 433 l
10 1 system performance assessment. i
,_s 2 Now, this is not a new subject to the
, j
\"
3 committee. We have written letters, had workshops, and '
4 have a considerable history of interest in this subject 5 and having made a number of recommendations.
6 For example, the ACNW reviewed the NRC's staff 7 high-level waste performance assessment program in October j 8 1991 and made a series of recommendations at that time, 9 published them in a letter to the chairman dated December 10 2, 1991.
11 To further evaluate the progress in developing 12 the modeling capability, ACNW held a working group on 13 high-level waste performance assessment program in May
,~
k.,-) 14 1994 and wrote a letter in May 27 to the chairman, the 15 then-Chairman Selin, review of the high-level radioactive 16 waste performance assessment capability of the NRC staff.
17 And there have been a variety of other working group 18 activities, addressing specific issues.
19 So with that, I think we will move forward.
20 Of course, tomorrow we will be having a working group 21 session on performance assessment in general whereas today 22 we're focusing pretty much on examining what the staff is 23 doing and their capability.
24 So let me turn the microphone over to the
~
'(s) 25 first speaker this afternoon, Keith McConnell, and, Keith, l NEAL R. GROSS i COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(
WASHINGTON, D.C. 20005-3701 (202) 234-4433 (202) 234-4433
11 l 1 I trust you'll introduce the subsequent speakers.
)
,,_ 2 MR. McCONNELL: Okay. Thank you, Dr. Garrick.
I 1 l
l N' '/ 3 My name is Keith McConnell. I'm the section 4 chief for the total system performance assessment and 5 integration section within the division of waste I
6 management, and I'm here this morning, along with Mike 7 Bell, my branch chief, to give you an overview of the PA j l
8 activities within the division of waste management and l l
9 also describe the framework and the schedule for those I 10 activities.
11 Before I start, though, I guess what I'd like 12 to do is to take care of a couple of housekeeping items. I 13 First of all, I know a lot of you were at the technical i
\)m 14 exchange that we had Monday and Tuesday. A lot of what 15 you'll see in this presentation and succeeding 16 presentations was presented at that technical exchange, j l
17 and I'll rely on the committee to tell us and the I l
1 18 presenters that, Hey, you're going too far into detail we 1 19 already know. So don't feel hesitant to tell us to move 20 forward. l l
21 Second, since we discussed with your staff the 22 scope of the agenda and the approach we'd take, we've 23 thought a little more about it and basically in this first i
24 presentation, we're going to run through the first three
(
n
( 25 agenda items right up to the break within this one !
t
\_-)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
12 1 presentation. I'll start off, and then Mike Bell will 73 2 follow me directly. There won't be a break in
( i x
3 presentations.
4 So with that, what I'd like to do is just lay 5 out the three objectives of this part of the presentation.
6 First of all, we want to convince the committee or 7 demonstrate that we have an integrated high-level waste 8 performance assessment program, integrated both up and 9 across the KTIs.
l 10 Second, we want to demonstrate that we have 11 the capability within the NRC staff and complemented by )
I 12 the center staff to independently evaluate DOE's TSPAs, I 13 the TSPA-VA and succeeding TSPAs that would be used in l
/~T l sY 14 documenting a license application.
15 And finally we're going to try to appraise you l 16 of all of our high-level waste performance assessment 17 activities, the schedules, and give you some inkling of i
18 the relative priorities.
19 The outline of the presentation is basically 20 again to give you an overview of the activities, talk 21 about priorities and schedules, discuss in some detail 22 allocated staff and the expertise that's assigned to the 23 various activities, and then Mike Bell will discuss the 24 hardware capabilities, both NRC's advanced computer system A
() 25 and here at the center.
1 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2:4-4433
13 1 There are three components to NRC's high-level fcx 2 waste performance assessment activities. This is a
( i us 3 diagram I showed Monday, but I've changed the sequence of 4 the various components, just to keep you on your toes.
5 Basically the first component is perhaps one 6 of the most important, and that's to integrate our own 7 program, and we do that by integrating and evaluating i
8 information across and within the KTIs. We work with our l 9 other KTI leads, and the KTI is like igneous activity, I
l 10 structural deformation, and isothermal flow and others to ;
11 determine basically which issues are important, which 12 subissues are important, and how we want to approach the 13 prioritization of various issues. And I'll discuss that in (3
i ;
k' 14 a little bit more detail in subsequent slides. I 15 We also within our evaluation of this 16 information want to understand the uncertainties and how 17 they might complicate or, I guess, drive the compliance 18 calculation.
19 The second component of our PA capabilities 20 within the division of waste management are to provide 21 feedback to DOE, and that was the purpose of the last two l 22 days, our discussions, detailed discussions with DOE on l
t l 23 their capabilities and their approach to TPA, TSPA, and 24 our own IPA efforts.
( ,)
25 So we believe we have the role of l NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
14 1 independently evaluating DOE's TSPA, and we're in a 7- 2 position to do that at this time. Also we will use our PA l )
3 capabilities in the future to evaluate the sufficiency of 4 data necessary for a license application.
5 The third component is to support the 6 regulatory framework development. I think most of you i
7 know that we've working with EPA over the last year or so i
8 to evaluate the implementability of a site-specific 9 standard for Yucca Mountain. j 10 We're also using our PA capabilities to gain i l
11 from the IPA work as well as the work leading up to our 1
1 12 TPA 3.1 code, to give us insight into developing a risk- i 13 informed performance-based, site-specific standard or (D
(__) 14 implementing rule.
I 15 And also we use our PA capabilities to develop i l
i 16 the acceptance criteria and eventually a standard review l 17 plan that would accompany an implementing rule.
18 Now, what I'd like to do is just describe in a 19 little bit more detail these three components of our PA 20 program. Basically, the integration role is very key.
21 What we're trying to do is change the perspective of PA 22 from the Great Satan to the Great Integrator within the 23 division; in other words, bring all the KTIs together, 24 have them reach out to the KTI leads, and develop a common n
() 25 buy-in to the PA and our own evaluation, not only l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
15 1 internally but also of our reviews of DOEs TSPA.
,~ 2 So the approach we're taking is to upgrade
( ,
\~' 3 NRC's TPA code to make it user-friendly. We intend, as 4 subsequent speakers will talk about, to involve the KTIs 5 in the sensitivity studies using the code, manipulating 6 the code, and doing our own process-level sensitivity 7 studies.
l 8 It's work-station-based now. We hope it will 9 eventually be PC-based. We're also upgrading the code to 10 include the new repository design and some more 11 mechanistic approaches to the various modules.
12 We're also working with the KTIs leads to 13 identify those areas that are key to the attraction L
O)
(m/ 14 process. What are those important issues and subissues 15 that need to be identified in order to demonstrate ;
16 compliance with the total system performance objective?
17 As I mentioned before, we intend to work with j l
18 the KTI leads to conduct process-level sensitivity ,
19 studies. Virginia Colten-Bradley will talk about that 20 this afternoon.
21 We also, following the process-level 22 sensitivity studies, we intend to do in the fall, late 23 fall, a total system sensitivity study to help us re-24 evaluate the subissues and issues that we have on the ts
(,w./) 25 table right now within the KTIs.
l NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
l 16 l
1 All of this information or all this work will l ,s 2 be reported in the issue resolution status reports on
( )
l
3 abstraction and relative importance, and also the process-4 level sensitivity study results will come out in the 5 individual KTI issue resolution status reports.
6 Providing early feedback to DOE on system l
7 performance, particularly the TSPA-VA: We've in essence 8 started our preparation for the review of TSPA-VA by j l
9 having the technical exchange the last two days and also i 1
10 in preparation for that, reviewing the TSPA-VA plan, the 11 waste containment and isolation strategy version that we 1
12 have, and also attending the various abstraction and l i
13 expert elicitation meetings that DOE has had on the p) t
\_- 14 various topics. l l
1 15 We intend to use the upgraded TPA 3.1 code to '
16 review and independently evaluate DOE's results. We'll 17 take an approach much like we did in TSPA-95, which is to 18 use a common set of input parameters. This is to the 19 extent practicable.
20 Because of the differences in the codes and 21 how the approach is taken in the codes, we can't use the 22 entire set of input parameters that DOE would use, but we 23 would certainly use a majority of their input parameters 24 and compare the results.
rx
( ) 25 We also intend to use our own independent set NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE. N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
17 1 of input parameters that have been derived through working
,s- 2 with the KTI leads and the process-level sensitivity
(
^'
)
3 studies that we intend to do, and, again, look at the 4 results and compare them with the TSPA-VA, and then when 5 you compare the results -- in order to build confidence in 6 our results, we would compare the results of our TPA work 7 with past IPA efforts and with past TSPA products that DOE 8 has produced, TSPA-93 and TSPA-95.
9 And, again, the results of this will be 10 reported primarily in the IRSRs, but I think if you were 11 here yesterday, you probably heard that we will be 12 preparing a commission paper on our views on the TSPA-VA.
13 The final component of our PA activities or (3
k_M 14 the third component of our PA activities is to develop a 15 risk-informed, performance-based rule. Right now we're 16 using the total system performance assessment to help us 17 in our development of a strategy for this risk-informed, 18 performance-based, site-specific rule, and we intended to 19 use PA to simplify, focus, and explain the regulatory 20 criteria that will be associated with that rule.
21 Basically what we intend to come up with is a 22 very clear, logical, and easy to understand and implement 23 implementing rule, using PA as the basis. We also intend 24 to establish a clear relationship between NRC acceptance (3
( ,) 25 criteria and the compliance calculation.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
18 1 In other words, acceptance criteria will be r~s 2 tied to their significance to performance, and I'll i \
Qi 3 illustrate that in subsequent slides. But, basically, 4 acceptance criteria will be derived for elements that are 5 tied to the performance assessment calculation. What is 6 needed to demonstrate compliance with the total system 7 performance assessment?
8 And finally we're going to ensure that an 9 implementing --
10 MEMBER POMEROY: Excuse me. In terms of 11 developing acceptance criteria, I didn't quite understand 12 what I heard yesterday. How are the acceptance criteria 13 going to be conveyed to DOE in a timely manner, ,
/~T l l t i N/ 14 considering that you need a certain amount of information l 1
15 from them as well as giving them some return guidance?
16 MR. McCONNELL: Well, we see it as an l 17 iterative or evolutionary process. The acceptance 18 criteria are coming out now in the IRSRs, issue resolution 19 status reports, from the individual KTIs. But there will 20 also be a roll-up issue resolution status report which is i l
21 referred to in my slides as the abstraction issue 22 resolution status report. l 23 What that IRSR attempts to do is to roll the 24 acceptance criteria into one package that will eventually
/~
(_,T) 25 evolve into a standard review plan. The timeliness issue NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
19 1 as we're proceeding now along, based on what we hear DOE 2 doing and our own knowledge of IPA work to develop these 7_ ;
t
+
3 acceptance criteria -- and they'll be coming out 4 periodically over the next year and year and a half --
5 IRSR on abstraction is scheduled to come out initially in, 6 I think, October. I have it on the schedule.
7 So that will provide the framework, and it'll 8 provide some of the details based on what the other KTI 9 leads have developed and what TSPA section has developed 10 at that time. But it will be sequential over the next 11 year, building up to an eventual license application 12 review plan.
13 MEMBER POMEROY: Since I've interrupted you,
.O m- 14 can I ask you the other question I wanted?
15 MR. McCONNELL: Sure.
16 MEMBER POMEROY: And that is the current 17 status of the TPA 3 code. Are you going to get into that 18 at some later point?
19 MR. McCONNELL: Yes. We'll talk about that.
20 Basically, I can get you a little insight now, I guess, 1 21 and Bob Baca will also be describing the code in a later 1
22 presentation. l 23 Right now we have an early version in, and 1
24 we're testing and verifying it. We hope to be able to,
() 25 starting on September 8, have a tested and verified code NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
20 1 ready to initiate the sensitivity studies. We're 7- 2 basically in the testing and verification stage of the
( )
3 code work at this point.
4 MEMBER POMEROY: And when do you intend to 5 have -- the sensitivity studies are perhaps one of the 6 things I'm more interested in -- the sensitivity studies 7 done for the igneous activity? Is that on a schedule?
8 MR. McCONNELL: Well, in essence, we've done 9 some scoping calculations to date, using the existing 10 code, but we wouldn't to -- we want to make sure that we 11 have a fully tested and verified code, and so we know what 12 the results mean and that we can have confidence in them.
13 My expectation is in the October time frame, f}
(_/ 14 the process-level sensitivity studies, which would include 15 volcanism and igneous activity, will be completed.
16 MEMBER POMEROY: Thank you.
I 17 MR. McCONNELL: I think that's -- basically !
18 the third bullet just indicates that we intend to make it 19 a -- use our awareness of the site characteristics of I 20 Yucca Mountain as well as our knowledge and our, I guess, 21 past work in assessing performance, to make it a 22 knowledgeable implementing rule; in other words, an 23 awareness of what we've done in the past, an awareness of 24 the site-specific data.
m.
25 We used this diagram in the technical exchange (w -)
NEAL R. GROSS i COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
c 21 1 also to basically illustrate our conceptual models of all l
f ,.-~s 2 of our PA activities, including integration activity, our
, i
( /
3 development of the code, our -- in essence, our review of 4 the TSPA work that DOE does, and our development of the 5 regulatory framework.
6 It has three levels basically. The focus is 7 at the top which is total system performance, which we 8 believe, based on National Academy recommendations, will 9 be -- and our interactions with EPA, will be an individual 10 dose calculation.
11 That's broken down into the subsystems which 12 embody the Commission's multiple barrier defense-in-depth 13 philosophy, including an engineered barrier, the
- Q
() _ 14 geosphere, and the biosphere. We then break that down 15 into the subsystem components, primarily the components of 16 the geosphere, and then in the third tier is where the 17 integration with the other KTIs and our own work occurs, 18 and that's the identification of the key elements of the 19 subsystem abstraction. l l
20 Again, what is necessary? What are the ]
21 principal elements that are necessary at the Yucca 1
22 Mountain site to demonstrate compliance with an overall l 23 system performance objective? And it's in this area where i i l 24 we will work with the KTIs to do the process-level l (v) 25 sensitivity studies, where we will look at the relative N!EAL R. GIUDSS l COURT REPORTERS AND TRANSCRIBERS I 1323 RHODE IS!.AND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
r l
22 1
1 priority and importance of various issues and subissues, 1
g3 2 and it's also where we will initiate the development of
(
l
'~~') 3 the acceptance criteria.
4 What we're doing in this third tier is using 5 our own IPA knowledge, as well as our knowledge of the DOE 6 program, to come up with a uniform set of key elements of 7 abstractions that apply across the board and develop 8 acceptance criteria that will go with them.
9 Conceptually, we would not expect the upper 10 two tiers to change. The lower tier will change, based on 11 the results of the sensitivity studies, our working with 12 DOE and their approaches and their results and their data.
13 Basically before you leave, I just wanted to t% I
(
x/ 14 emphasize that what this diagram attempts to do is to 15 basically depict the functionality of our code, how we're 16 approaching the 3.1 code, the depiction of a potential 17 outline for an implementing rule that is risk-based -- )
l 18 risk-informed, performance-based. And, third, it's a i 19 depiction of how we're going to generate our acceptance 20 criteria and what they're going to be applied to.
21 This is an appropriate breaking point. What 22 I'm going to do following this slide is to go into the 1
23 various activities, so I guess if there are any I l
24 activities -- or any questions about how we're approaching
(],
l 25 things in the integration aspect, maybe they'd be NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
23 1 appropriately addressed now.
,w
- 2 CHAIRMAN GARRICK: You're probably going to
(',,)
3 get into this later, but we have in the past asked 4 questions about the KTIs and how they came about, and one 5 of the things, of course, we learned is that the KTIs 6 evolved from a variety of processes but not necessarily as 7 a direct result of performance assessment.
8 Now, I'm curious, as you now advance your PA 9 model and try to address the KTIs, what the process will 10 be -- and I don't want a long discussion about it -- in 11 the event, for example, that the performance assessment 12 indicates that other issues are more important than the 13 ones you have or the issues that you have are not
(
V) 14 appropriately ranked or what have you.
15 In other words, are you going to -- do you 16 have a specific method of evaluating KTIs and allowing the 17 performance assessment to be an important tool in )
18 establishing their importance and also reaching closure on l
19 that and possibly seeking others that are not at this time 20 considered KTIs? Or are you just focusing on the KTIs and 21 taking them as a given? I guess that's my question.
22 MR. McCONNELL: I don't think we're taking 23 them as a given. I think to kind of go back to the first 24 part of your question, I think the KTIs were developed in j A
( ,) 25 a performance-informed atmosphere. In other words, there NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
1
24 l 1 was the background of the IPA Phase 2 work, our knowledge
, _s 2 of what DOE had done in TSPA, as well as our own informed i
\
)
/
3 judgment about what was important to performance, but it 4 was basically qualitative, a qualitative judgment.
5 What we're attempting to do now with the TPA 6 3.1 code is sensitivity studies at the process level and 7 then the overall -- the system level sensitivity studies 8 is to give that a quantitative basis, to look at the 9 various issues and subissues within a particular KTI, 10 determine their relative importance and their overall 11 importance, and go from there.
12 So we're using the PA tools and capabilities 13 that we have to basically do what you asked, which is to
'n \
(_/ 14 go back and look at the KTIs, go back and look at the l 15 issues within a particular KTI, and make sure we're 16 focused on the correct or on the most important aspects of ,
1 17 the repository system.
18 Now, one important aspect of that is to have a 19 consensus view of that. It can't just be a performance 20 assessment view, and what we're attempting to do is, 21 again, to integrate the work, bring the KTIs into the 22 process, and develop the buy-in that's necessary. It's a 23 key aspect of our program, to build the support for the l
l 24 TSPA and the PA efforts that we're doing.
m
( ) 25 So, in essence, it's going to be a uniform NEAL R. GROSS COURT REPORTERS AND TRANSCRlBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
25 1 judgment or a combined judgment that's recommended to 2 management, to look at the various priorities within a KTI
[s) 3 and across KTIs.
4 CHAIRMAN GARRICK: Thank you.
5 DR. HINZE: Keith, one of the things that I 6 find extremely interesting about your diagram on the total 7 system is the fact that the biosphere is included in That's a new concept, at least to me, ;
8 defense-in-depth. 1 9 and I see the geosphere and engineered system as something 10 you would evaluate and re-engineer.
11 Can you expand a bit on how the biosphere 12 enters into the defense-in-depth concept and how you see 1 13 this developing?
V 14 MR. McCONNELL: Well, with the change from a 15 release-based standard to a dose-based standard, the 16 biosphere becomes critically important. Uptake factors 17 from water to plants and then the consumption and 18 lifestyle factors that are involved with what would 19 eventually be a critical group, the definition of a 20 referenced biosphere, all become very critical factors 21 within the compliance calculation.
22 And that's why biosphere is given the 23 prominence it is. It's a key component now in the overall 24 calculation of compliance with a dose-based or risk-based
) 25 performance standard.
v NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
26 1 OR. HINZE: There certainly is going to have 7s 2 to be much more in the way of policy decisions related to k
')
3 the biosphere proposition than in these other areas which 4 are based upon -- more on, if you will, science and 5 technology. And, therefore, this seems to be not in the 6 same -- handled in quite the same way as -- biosphere not 7 handled in quite the same way.
8 MR. McCONNELL: Well, there may be policy 9 issues across the board, but -- or how we approach all of 10 the subsystems. But I think you're correct, that there 11 are a lot of policy issues, and I think that in defining 12 our strategy for an implementing rule, those will be 13 outlined, and we certainly would probably end up -- we
,O\,
's/ 14 will be going to the Commission for a determination on 15 those policy issues as we move forward with developing the 16 strategy for the implementing rule and the eventual rule.
17 DR. HINZE: Thank you.
18 CHAIRMAN GARRICK: One more question.
19 MEMBER POMEROY: With regard to your 20 acceptance criteria, I presume you're ensuring an 21 integration, at least a careful cross consideration 22 between all the elements, between the different KTIs in 23 the process of developing these acceptance criteria within 24 each of the KITS. Is that correct? Or could you clarify l3
() 25 that for me?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005 3701 (202) 234-4433
27 1 How does that happen? How are you ensuring 2 that there's not some conflict between acceptance criteria i
g) 3 in one area versus the others?
4 MR. McCONNELL: Again, that's the integrating 5 role. What we are attempting to do is to, in identifying 6 key elements of the abstraction, to use those as the focus 7 of the acceptance criteria development.
i 8 What that does is it drives the various KTIs 9 to focus on a specific element of performance, not on 10 specific KTI and a specific subissue within a KTI. So, in 11 essence, they're focusing on a particular element in the 12 compliance calculation.
13 And usually -- in many cases, these key O
'd 14 elements of subsystem abstractions are a mix of KTIs.
15 Waste package corrosion is a good example, up in the upper 16 left or what is my upper left corner. Basically that's a 17 combination of work being done in container-like and 18 source-term KTI, and the near-fiel'd environment KTI.
19 So what we're attempting to do again is use 20 these elements as the focusing point for developing the 21 acceptance criteria. And that gets to the integration l
22 aspect.
23 MEMBER POMEROY: Okay. Thank you, i 24 DR. HINZE: Could I ask one more question? i 25 CHAIRMAN GARRICK: Sure.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
28 1 DR. HINZE: Keith, of the ten KTIs, seven 73 2 received prominent attention; three were not studied as
. t )
\_)
3 intensely. Are -- has that changed? Are there certain 4 aspects here to the key elements that are included in 5 those three that were downgraded at a lower priority?
6 MR. McCONNELL: They're included. We have 7 staff at the NRC that is working on those particular KTIs 8 or issues.
9 DR. HINZE: So those three issues now are no 10 longer second priority? They are being intensely 1
11 investigated? For example, this retardation in fractures; 12 my understanding was that's one of those that was 13 downgraded. I would think that that would be in there,
,/-
's / 14 and I'm concerned about the timing. Are you completing 15 the work in time so as you can have the input to the UZ 16 flow and transport?
17 MR. McCONNELL: Well, I think you're picking 18 up on one of the factors this diagram is resulting in, 19 which is an identification; again, this shifting of 20 perhaps emphasis on various issues. We're recognizing 21 now, using our informed judgment, what -- quantitatively 1
22 what is going to be very important to the compliance 23 calculation.
24 I think when we get done with our system-level
'( ) 25 and process-level sensitivity studies, we may reorient the l
NEAL R. GROISS l COURT REPORTERS AND TRANSCRIBERS :
1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344 433 .
l
29 1 various priorities and what was changed or what was
,s 2 downgraded in the last budget reduction, and what we might i \
t 3 approach in the future.
4 I think if you want a specific answer to 5 retardation in fractures --
6 DR. HINZE: No. I was just using that as an 7 example. I'm wondering about these issue resolution 8 reports.
9 MR. McCONNELL: Right.
10 DR. HINZE: Will there be something, an issue 11 resolution report, in a timely manner to put into this l
12 program that you have?
13 MR. McCONNELL: It will -- specifically it gq i l
\~j 14 will be adopted in the roll-tp IRSR on abstraction. It's 15 a key part of the abstraction for UZ flow and transport.
16 Acceptance criteria will be identified for it. I don't l 17 have the time frame of when, but --
18 DR. HINZE: I'm just worried about getting all 19 of the priority elements --
20 MR. McCONNELL: Right.
! 21 DR. HINZE: -- in place when --
l 22 MR. McCONNELL: What we're doing in the 23 abstraction issue resolution status report is to identify,,
24 again, all the key elements of abstraction that are needed 7'%
i, ) 25 to demonstrate compliance. They won't all have acceptance s_/
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS j 1323 RHODE ISLAND AVE., N.W.
lg (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
30 l 1 criteria within this first initial document that comes l
t 2 out, but at least we'll have a placeholder for that
(/7 i
r \
i
')
3 particular element in the abstraction.
4 DR. HINZE: I assume by the end of the day, 5 we'll have a better feel on what's -- what will be the 6 sense and what will be in the issue resolution report.
7 MR. McCONNELL: Basically if you're going to 8 get that, you're going to get that right in this 9 presentation, so I need to do a better job in 10 communicating.
11 DR. HINZE: When is the first one scheduled 12 out?
13 MR. McCONNELL: There's one out right now on w, 14 climate that came out in the last two or three weeks, and 15 I think there are several -- maybe I can ask Robert l
16 Johnson to respond. There are several coming out in the 17 near future, I believe.
1E MR. JOHNSON: Yes. Robert Johnson. I l
19 September 30, we should have about, I think, 20 eleven are scheduled and then Keith had mentioned two, one 21 on abstraction and important performance, would be coming 22 out, you know, later in the calendar year, and then 23 there'll an update to all the IRSRs in the spring, around 24 April time frame, so that around April, you know, all the l 25 IRSRs with all the subsections and acceptance criteria --
(J N
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
31 I l
1 1 You know, that's our current plan, to have 1
-, 2 them out -- all of them out by April time frame, so the
( :
3 first phase is by the end of September, and then the two 4 that Keith is working on would be in the end of the 5 calendar year.
6 DR. HINZE: I would think the committee would i
7 be very interested in seeing one of these just as soon as 8 possible, because I think as we discuss the responses 9 within the committee, there's been a perception of what is 10 going to be in the IRSR for volcanic processings, and I'd 11 like to make certain that that perception is reasonably 12 correct. If we could obtain copies, it would help us.
I 13 MR. McCONNELL: Okay. Margaret, did you want l /D l \
, \_) 14 to say something?
\
15 MS. FEDERLINE: Yes. Dr. Hinze, I just --
l l
16 Margaret Federline, division of waste management. I just i
17 w nted to add to Dr. Hinze. The budget cuts have hurt us.
18 I man, I think if we were to say that we are not able to 19 do a ll the scientific work that we want to do and there 20 are some key elements of radionuclide transport that have l
l 21 had to be eliminated at the Center.
! 22 You know, we've worked very hard. The 23 Chairman supported our budget for next year, but I think 24 this process is a process to identify what are the really (n) s/
25 significant pieces of scientific work and make sure that NEAL R. GROSS COURT REPORTERS AtJD TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
32 l
t l
1 they rise to the top, and we're going to do that fw 2 reprioritization every year.
a l
\' ') 3 So, you know, I just didn't want to leave you 4 with the impression that we felt we were able to do all l
5 that we wanted to do.
6 DR. HINZE: That's helpful, and that's 7 understood. But it is also a concern that you have to do 8 a certain amount of science in order to know what the 9 problems are and which are the critical problems.
10 CHAIRMAN GARRICK: Okay.
11 MR. McCONNELL: We'll proceed then. What I'm 12 now going to do is walk you through the various activities 13 within the high-level waste performance assessment area, 7
! \
s/ 14 and I should say that these are only a part of the 15 activities that are carried out in the section. We also 16 do performance assessment work in low-level waste.
17 I'm sure you're familiar with the low-level 18 waste BTP that's just recently gone out for public 19 comment. We work quite intensely in the performance 20 assessment activities related to site decommissioning 21 management, the site decommissioning management plan, and 22 the various sites that are on that list. And thirdly, 23 we're also involved in some uranium recovery work where we 24 support some of their analyses, particularly in the area
! r~'s l(~- ) 25 of occupational safety, I believe, or health physics.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344 433 WASHINGTON, D C. 20005-3701 (202) 234-4433
33 1 The first topic I'll discuss or the first
,-~ 2 activity is the TPA 3.1 development. It actually should i /
'~'
3 be 3.1. This has been a top priority of our efforts over 4 the last six to eight months. Basically it's a joint 5 NRC/ Center work effort. The Center is doing primarily the 6 code work. We're working with them to develop the 7 conceptual basis for the code and the outline for the 8 code, how the code should be approached.
9 As I indicated, the work is ongoing. We got 10 version 3.0 in March. That's gone back for subsequent 11 update. We now expect a revised version of that code, 12 what we call 3.1, to be available on September 8 to 13 initiate the process-level sensitivity studies at that
/ i D' 14 time.
15 At that -- we won't be able to go out with a 16 fully documented users manual at that time. We don't l
17 expect that until November, but -- because we're going to I
18 continue to use the sensitivity studies as a testing and l
1 19 verification aspect of this code development. I 20 So we're expecting in the November time frame 1
1 21 to have a completed users manual and a complete 3.1 code l 22 available for public distribution.
23 Tim McCartin, who you'll hear from this 24 afternoon, is the lead. He works with Bob Baca here at
/
(3,) 25 the Center. Just to give you some insight into the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
34 1 expertise, Tim was involved in IPA phase 2 and goes a long
,_ 2 way, even though he looks rather young, in the performance
(\') 3 assessment activities, all the way back into his research 4 days. And he also -- he won't tell you this, but I 5 will -- was just awarded a Meritorious Award for Science 6 by the NRC.
7 Part of that process -- and I've mentioned it 8 several times before -- is the sensitivity studies, both 9 the process-level and the system-level sensitivity 10 studies. Virginia Colten-Bradley will talk to you this 11 afternoon about that. We expect to begin on the process-12 level in October -- or actually to complete them in l I
13 October, start on them in September. I l
(~%
l (_-) 14 We will then step back, relook at our base 15 case definition, based on the results of that -- those l
16 process-level sensitivity studies, redefine our base case, l
17 and then go forward with the system-level sensitivity i
18 studies. But, again, we'll look at the relative 19 importance of the subissues and the issues.
20 Virginia, again, is the lead. She's our l
21 interface with the KTIs. She, you know, walks the earth l
22 science walk and talks the PA talk. So she's our basic 23 interface, and we rely on her to be inclusive in our work, 24 and she's supported by Christiana Lui who has brings with c.,
25 her from research and other activities some PRA insight (v)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
35 1 into the process.
,_s 2 DR. HINZE: Keith, before you leave that, one
( )
3 of the things in looking at the discussions in '94 4 regarding PA and the NRC, one of the major topics was 5 confidence-building; one of the major efforts in the l
6 future as discussed at that time regarded confidence- l 7 building.
l l
8 Where is that in this?
9 MR. McCONNELL: Confidence-building in what 10 respect?
11 DR. HINZE: Confidence-building in the models 12 and in the programs; it dubbed this terminology that the 13 NRC was using. I don't hear that here, and I think it's n
(m l 14 really a great term. I'm wondering where it is?
15 MR. McCONNELL: It's perhaps more implicit 16 than explicit. Basically confidence-building is two-l 17 phased. It's the human part of confidence-building, which 18 is to bring everybody into the process, have them run the
- 19 code, have them develop confidence in what this effort is I
l 20 producing. And therefore they buy into the results.
1 21 There's confidence developed in the results. !
22 There's also the computational confidence-23 building, and that's where we are doing the testing and 24 verification of this code; that we understand what's in
- v; 25 the module, that we know, looking at the intermediate NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
i l
- 36 l
l 1 results, what's coming out of the modules, the various l
l 2 modules within the code, like volcano; evaluating the 7-
- l' ') 3 results; then stepping back and taking a look and building 4 confidence then that what we've produced in an overall 1
5 system evaluation is also understandable, traceable, 6 transparent, to use the word, and also we have confidence 7 in what it's telling us.
1 8 So I guess maybe it's my newness to the 9 program, but it's more implicit rather than explicit.
l 10 Does that answer your question?
11 DR. HINZE: Yes. I think it gets at it. I'm 12 wondering: Is there any attempt made at history 13 benchmarking of the process codes, that type of thing; i /~h
' k-) 14 any -- you're moving very rapidly to the TSPA-VA. In your 15 independent evaluation of that, as we've heard of cutbacks 16 in the program, I'm wondering just the impact upon the -- i 17 this is something that I think the committee and I will 18 try to come back to as we go through in the next couple of 19 days.
20 Are we -- have we lost anything in this 21 process of cutback in research, cutback in the KTIs that 22 we're interested in, et cetera? Where is the confidence?
I
! 23 Are we being assured that we have some confidence, not 24 just that the programs will run and that we've got n
() 25 parameters put into it, but that these really do a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS
, 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433 l
l
37 1 reasonable approximation of what the processes are going 7s 2 on and the engineered barriers --
('~' )
3 MR. McCONNELL: Yes. I think -- I'll let Wes 4 Patrick -- I think Wes had a comment.
5 MR. PATRICK: One of those things --
6 CHAIRMAN GARRICK: Do you want to give your 7 name --
8 MR. PATRICK: Wes Patrick from the Center.
9 One of the things that was lost, Bill --
10 you're stating things in absolute terms which beg a yes 11 answer that there has been an impact. You used the term 12 confidence-building. Another term that we used when the 13 research program was active was a confirmatory type of g))
(s 14 research that was being conducted. You probably well 15 remember that term as well.
16 We are not able to do many of those thinge.
17 Beginning in the end of '95, we transferred the -- what we 18 viewed at that time to be the most essential core parts of 19 that into this KTI structure, so we have been able to do 20 some confirmation.
21 I think you're aware of some of the work that 22 we've brought out in appeal and done with regard to some 23 of the key structures that exist out there in the economic 24 vicinity, some of the potential buried volcanic sites, O
L j 25 some of the key elements with regard to vegetative NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 1
38 l
l 1 covering which could have a significant influence on fm 2 infiltration.
( )
x~/
3 But certainly there has been a reduction in 4 those sorts of confirmatory and confidence-building 5 activities. We've tried, though, to identify and grasp 6 the most important of those and to continue those to the 7 extent we can.
8 We fully anticipate that as these sensitivity 9 studies are conducted, that there will be a shifting in 10 where we place those scarce resources, so that we confirm 11 the most important things and that we're able to build 12 confidence in the most important things. But certainly we 13 cannot do all in those areas.
f p\
's l 14 MR. McCONNELL: I think that you'll hear when l 15 Tim discusses the various modules the process we took to 16 not only identify what needed to go into the modules but 17 what we expected out and what the weaknesses or 18 uncertainties in the approach were. l 19 And so I think you're correct that the budget 20 reductions have had an effect, as Margaret has indicated, 21 but what we're doing now is identifying those areas and 22 when we get done with the sensitivity studies, then we go 23 back to management with our recommendation about where we 24 need to proceed, where these weaknesses in confidence-n
) 25 building are.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
39 1 CHAIRMAN GARRICK: Marty?
j ,.~ 2 DR. STEINDLER: You indicated that sensitivity
' ('~'! 3 studies are going to serve as the testing and verification 4 of the code, and you indicate you're going to do that at 5 the process level, the system level. What sort of utility l
6 do you expect to get out of that when you consider that 7 the design of the repository is in flux?
8 MR. McCONNELL: Well, we'll get an 9 intermediate result. What we're trying to do is build 10 flexibility into the code, where we can make changes in 11 things like design and testing those changes. We hope to l
12 be able to accomplish those fairly easily and quickly, so 1
i 13 we hope to be quick on our tcce. l
' /~'i l (s/ 14 To address this specific design that we had, 15 the reference design, basically that's what DOE's going to j 16 go forward with in their TSPA-VA, once they freeze it.
17 And that's what we will evaluate, realizing that it's a 18 snapshot in time, but we hope we have the flexibility to, 19 I think, evaluate any design that they would decide they 20 want to proceed with or design factor, like trip shields 21 or backfill or any other factor.
22 DR. STEINDLER: So you believe then that the 1
l 23 verification process that you're going through by these 24 sensitivity studies are going to be design independent.
em (t- ) 25 MR. McCONNELL: No. I believe that they will NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
40 i
1 be design dependent, but should the design change, we will
- 2 go back and re-evaluate the sensitivities of that i i
)
l 3 particular design issue. This is -- these sensitivity 4 studies are going to be ongoing.
r l
l 5 We kind of -- I guess I kind of left the 6 impression we're going to do them once and that's going to 7 be it. But if DOE comes in with a significant change in 8 design, we would go back and do our own sensitivity 9 analysis, both at the process level and the system level.
10 That's what we're hoping to -- you know, in 11 the past, running the TPA code has been a very significant 12 effort. It's been on a CRAY at INEL, I believe, and only 13 one or two people could run it. Making changes in the C\
(-) 14 code was very difficult and cumbersome.
15 What we're hoping to do is build a system 16 where it's much more flexible. We can make those changes 17 in relatively short order and evaluate these things in a 18 more tactical time frame rather than a strategic time 19 frame.
20 MEMBER POMEROY: Keith, I have a quick 21 question, I think. I'm concerned a little bit about the 22 overall -- the question of the overall code and how you're 23 going to determine it's calibrated, if you will, in some 24 way.
! (w j 25 That is, are there any external problems that NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l 41 1 have been run by other people, EPRI or overseae, whatever, f
7_3 2 that might give you the opportunity to contrast the output
, 2 i
~'
3 of your new code directly with somebody else's codes, just 4 to see whether the same answers came out or whether --
5 MR. McCONNELL: Well, it's difficult, I think, 6 to compare the results from this code with the results 7 from an overseas code. Maybe I'll let Tim talk more 8 directly to this. But in that sense, the way we're going 9 about calibrating it is to compare -- we have a baseline 10 in the IPA Phase 2 results. We're going to try to run the 11 same data set that we ran for the Phase 2 work and compare 12 the results.
13 We also have the opportunity to compare the
,r'T
\s/I 14 results with DOE's TSPA-VA, again, an effort to calibrate 15 what's coming out of this code. And maybe, Tim, you could 16 provide more information.
17 MR. McCARTIN: I guess there's a couple things 18 that we're doing or will do in the next six weeks or so, 19 and certainly in terms of the accuracy of the code, if you 20 will, we will be doing some fairly diligent hand 21 calculations to make sure that things are being sent 22 through the code and processed correctly or the way we 23 intended them.
24 Additionally, we will -- the code as it's now l
,3 25 envisioned -- Bob Baca will talk to this to a certain (J l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
! l l
42 1 extent, so I don't want to say too much, because he will 1
l l f~s 2 get into some of it. -- is that we have a lot of lk'"
l 3 intermediate outputs at a lot of the subsystem levels.
i l 4 We will be looking at those results, those 5 intermediate results, to see, Do I believe those numbers; i
6 do they make sense. And so it will be a fairly careful 7 look at, why do I believe the numbers coming out the end, j i
8 because clearly the end product is a dose. i 9 It's very difficult to say, Gee, is that l l
10 reasonable or not. There's almost no comparison that I'm 11 aware of that you could look at and get comfort with that, ;
1 1
12 but looking at the intermediate results, with overall 13 infiltration, the reflux, the source term, as you go i/ 14 through that system, you should be able to at least look 15 at that.
16 And potentially other codes from around the 17 world, seeing what kinds of nuclides they're getting out, l
l 18 et cetera, might be some support as to why this seems i
19 reasonable.
20 CHAIRMAN GARRICK: Just one comment I want to 21 make, Keith, while you're up there, because I think that 22 we have to remember that the purpose of our meeting today l l
! l 23 is to consider the performance assessment capabilities and i i
24 the staff, and while we as a committee can hear the (h
/ 25 presentations and form our own judgments about that, it NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
43 1 would certainly help us if you w5bld telegraph or be as 7 s 2 explicit as you can on what you believe and your staff to
( ,
3 be the soft spots or the capability issues.
4 And our antennae is out on this, on the basis 5 of Margaret Federline's comments about budgets and some of 6 the frustrations that seem to be in place as a result of 7 that. So as we go along and talk about these different 8 problems, if there's any signals that you can send us --
9 I know you're proud of the capability, and you 10 want to accentuate the positive, but we can help you if 11 you let us know what your problems are in terms of 12 capability, because you're in the throes, for example, of 13 a major software development activity, and having been in p 4
,/
\~ 14 that position myself many times, I can appreciate the 15 absolute importance of having certain capabilities on 16 hand, such as programmers, the right kind of programmers 17 or what have you, to end up with a product that will do 18 the things you're talking about, and that could be made 19 available to other people, et cetera, et cetera, and, you 20 know, so that we can get to the PC version as quickly as 21 possible.
22 So I guess I'm just pausing here and saying to 23 you that what we want to do is be constructive in our 24 advice to the Commission about these capabilities, and we in
/ \
( ,/ 25 especially want to be accurate and sensitive to the issue l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
44 1 of soft spots or capability problems that are really
, ,_ 2 handicapping what you're trying to do. It's extremely 3 impcrtant.
4 MR. McCONNELL: Okay. I'll try to be as 5 candid as I can, and I think Norm -- a comment?
6 MR. EISENBERG: Norm Eisenberg. May I say 7 something?
8 MR. McCONNELL: Yes.
9 MR. EISENBERG: I'm Norman Eisenberg from the 10 division of waste management. With regard to this 1
11 question of verification of the code, let me just make two 12 points. It's very hard to verify the code overall, 13 because few people, other than, I would say, DOE have a rs
/ I
'N_ / 14 system-level code that's modeling Yucca Mountain. EPRI 15 has one also, but there are not that many places to 16 benchmark our code.
17 However, pieces of the code have been a 18 benchmark, so, for example, we're using the NEFTRAN code 19 imbedded in our system code, and this has participated in 20 international benchmarking efforts going back 15 years, so l
21 we have that experience.
I 22 And, for example, the code use to estimate the l l
l 23 effects of volcanic eruption have been matched to the ash 24 deposition from a number of volcanos around the world, so
(T
( ) 25 there is some history matching on a number of the x_/
l NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE , N.W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
I 45 1 components of the code. There's some parts that may --
i
,_s 2 don't have the history matching, most notably, I'd say,
,, )
! /
" But to the extent that it can be done, 3 the source code.
4 we have done some exercise into that regard.
5 MEMBER POMEROY: Can I ask one other question?
6 CHAIRMAN GARRICK: Sure.
7 MEMBER POMEROY: Norm, I tend to agree with 8 you. On the other hand, in those two places where there 9 is some other code, I have some problems, for example, 10 with comparison -- comparing directly witn DOE, because l
11 some of the same codes are being imbedded in both codes, 1 12 as I understand it.
1 13 You mentioned NEFTRAN, but I understand 1
, ,fg j b 14 it's -- as a result of its recent rewrite of TPA 3, that l 15 there've been significant changes that have had to have 16 been made in NEFTRAN. Are those changes being evaluated 17 in a systematic manner, for example, or are they being 18 cottpared with the previous NEFTRAN results, to see whether i i
1 19 something can -- at least the logic order is safely in 20 there?
21 MR. McCONNELL: I think Tim McCartin might be 22 able to respond to that.
23 MR, McCARTIN: We certainly evaluate all the I
l 24 changes we've made to any of our codes. NEFTRAN -- I
/~N
( ) 25 don't know if I'd call them significant, but we certainly I
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
46 1 have made changes to NEFTRAN. We are comparing it to the 7_s 2 benchmark problems that were published with the code, to f I
' 3 check those results. We're also comparing the results to 4 other codes, to make sure that we haven't introduced 5 something that we didn't want basically.
6 But, yes. I mean, we certainly do a fair 7 amount of effort to do that. Not too surprising, any --
8 we have found some minor problems with NEFTRAN, both in --
9 when we were using it in low-level waste and currently in 10 the TPA code, all related to the last modification made to 11 NEFTRAN.
12 If you -- the history of NEFTRAN is a storied 13 one that started out and moved quite along and
,,\
x- 14 improvements made to it. The last modification made to it 15 was time-varying velocity. And that aspect -- as the last 16 modification to NEFTRAN, one would expect that there might 17 be some minor errors to it.
18 But I don't know if I'd say there were major 19 modifications, but certainly we have found some minor 20 bugs, depending on how the code is exercised, and clearly 21 the best way to find little bugs with codes is do Monte 22 Carlo sampling, and you get to parts of the code that just 23 weren't exercised, and we have found some -- I would l
24 classify -- I don't know if I'd call them major, but
,c
( ,) 25 certainly there were some errors that we've picked up in
, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
47 1 the sampling part of the exercising the code.
,x
,- 2 MR. McCONNELL: I think we're talking NEFTRAN
\ ' ^ ^ '
)
3 and some of the other codes to areas where they haven't 4 been exercised that much, and so in that sense, we are 5 identifying some minor changes that need to be made.
6 MEMBER POMEROY: Okay. Thank you. l l
7 MR. McCONNELL: One thing I would say is that 8 the code effort, as Dr. Jarrick identified it, has been a ,
l l
9 massive effort over the last six months, both on our part 10 and the Center's part, and I think we've been fortunate in 11 that we're -- have our timing down correctly. In other 12 words, we're able to do things sequentially in this list.
1 13 If everything had come and hit us at one time, I
! \
'% / 14 I think we would have had a great deal of difficulty 15 accomplishing everything at the same time.
16 Just to move on, I wanted to talk about 1c, 17 which is the importance analysis. This is a more PRA-type 18 of importance analysis. This is under development now, 19 and Norman Eisenberg and Dr. Sagar here at the Center are 20 working on the conceptual framework of that approach to 21 analysis.
22 We hope to have it in place, both imbedded in 23 the TPA 3.1 code and also the conceptual basis done prior l
l 24 to the VA review, so we can implement that in our
!f) l( j 25 evaluation of the TSPA-VA. I think Norm will be talking a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
i 48 1 little bit about that tomorrow.
7- 2 Two is the EPA standard review. As you know, N3 ]
3 we have worked with EPA on testing the implementability of 4 their standard as they were developing their draft 5 standard. We haven't heard much of anything in the last, 6 I'd say, six months or so, and so basically it's a TBD 7 activity. We're waiting for EPA to do -- to issue their 8 standard.
9 Tim and Janet Kotra would be the leads on 10 that. Maybe you know Janet Kotra from Commissioner 11 Curtiss's staff.
12 4a and 4b are our issue resolution status 13 reports, one on abstraction. Again this is the roll-up of f3
\ )
'/ 14 that bottom tier of the diagram, the key elements of 15 abstraction, and we're going to start in October and 16 define the acceptance criteria that would be the basis for 17 determining compliance with those key elements of 18 abstraction, so it will be at the abstraction level.
19 And Christiana Lui is working on that, as well 20 as the next one which is the importance, and I think I've 21 talked enough about that, so we'll move on to the next.
22 Developing the strategy for the site-specific 23 high-level 'eraste implementing rule, actually it should l
7 24 say. We were on hold for a while. We restarted, and l r~'N k_) 25 we're now in the process of developing an outline of that NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
f l 49 1 strategy. We've briefed our own division of waste j ,, 2 management on that strategy and are moving forward now I' ') 3 with a schedule that has us briefing Carl Paperiello on 4 the outline of the strategy in August of this year.
5 We hope then, assuming we get his ap7roval to 6 move forward, to have a draft done in the September time 7 frame, and then go to the Commission in late September 8 with the strategy for approaching an implementing rule for 9 a site-specific high-level waste site.
10 And Janet Kotra is back at the -- if she's not 11 there at the meeting, is back working on the strategy now.
12 Assuming we get Commission approval to move 13 forward with the strategy we recommend or approach in the
, s,
' \
's ,/ 14 Commission paper, we would expect that in the June time 15 frame, June of '98 time frame, we would have a proposed --
16 draft proposed implementing rule ready to go out for 17 public comment.
18 The only other thing I'll identify on this 19 page is number 8, which is the annual report chapters. )
20 They're a major effort here and at the Center. Each of l
21 the KTIs, probably as you're familiar, does a chapter l 22 within the Center's annual report, and it's a joint effort 23 between the Center and the staff, and it documents all our l
l 24 activities for the year, the past year. And we've got to i
(Q) 25 the TSPA and IKTI, the tota] system performance assessment v
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
50 1 and integration KTI, and what we call the rule KTI 7s 2 chapters.
( \
x~/ 3 The only thing I'll point out on this slide is 4 11 and 12. This is basically our international 5 activities, where we keep in touch with what's going on in 6 the international community within performance assessment. )
I 7 We're heavily involved in the revision to ICRP )
i I
8 46, the high-level waste disposal effort, and also we're i'
9 involved more in an observer status in the BIOMASS efforc 10 which is the definition of -- international effort to l
11 define criteria for reference biosphere and, I think, 12 critical group. l 13 And Chris McKinney is working on that aspect .
f~h l k_) 14 of the program.
i 15 Some major upcoming tasks that we see on the !
I 16 horizon for the performance assessment group and also the I 17 division as a whole: The TSPA-VA review, everybody is 18 aware of. We were told, I think, yesterday that we could 19 expect the reference data set from DOE sometime in the 20 January time frame, so we could continue our review, and 21 then we might get the actual TSPA-VA results and, I guess, 22 package in the June time frame. I assume we'll have 23 several months to look at it before we go forth with a 24 Commission paper, probably in the September -- August, r
(y) v 25 September, October time frame.
i NEAL R. GROSS l COURT REPORTERS AND TRANSCR:BERS l 1323 RHODE ISLAND AVE., N.W.
l (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
{
51 1 We also, assuming we get -- move forward with
- f. 2 our strategy for the rule and our strategy and development
( )
3 in the code, we will go forward with the development of a 4 standard review plan; again, taking that bottom tier, 5 develop risk-informed performance-based acceptance 6 criteria and rolling those up into our abstraction issue 7 resolution status report which would then eventually roll 8 up into a standard review plan.
9 So we're approaching this basically from the 10 bottom up, but with a risk-informed constraint, if I could 11 say.
12 We also intend to comment on the -- use our PA 13 tools and capability to comment on the sufficiency review
(,-)\
x_ 14 for -- that DOE will do for the license application, and 15 also in fiscal year '99, we intend to initiate a program 16 where we look at preclosure performance assessment and 17 what's needed to demonstrate compliance for the 18 preclosure.
19 Again, we're at a breaking point, and I don't 20 want to take all the time, but if there are any questions 21 on activities, now is the time, I think.
22 MEMBER POMEROY: Keith, I thought John, as 23 always, made a very important -- his last point was very 24 important, in asking you to point out where the weak
/-
- (x) 25 points were. I guess one of the concerns that I have is HEAL R. GROSS 1 COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W. l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
52 1 the distribution of activities between NRC and the Center.
.s
- 2 I know you've worked very, very diligently and hard to l \
~
3 fill those in an integrative fashion.
4 What happens if some external force, for 5 example, caused removal of a significant portion or all of 6 that Center support? What is the -- is the NRC, per se, 7 team capable of carrying out most of these operations at 8 this point in time, without Center support?
9 MR. McCONNELL: I believe we have the 10 capability, but what it would do, it would significantly 11 extend the time that it would take. The Center provides 12 very high quality, strong support to our efforts, and we 13 rely on them quite heavily. We have the capability in-7 k
\> 14 house; we're expanding our capability in-house. But we 15 have limited people, and they're quite talented, but you 16 can only make them work 12 to 14 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br /> a day. You know, 17 beyond that, they get really testy.
18 MEMBER POMEROY: I know you do that.
19 MR. McCONNELL: So I think that -- to be, I 20 guess, less flippant, we're aware of that possibility. We 21 believe we have the capability to proceed in-house, but it 22 would be a -- have a great impact on our program.
( 23 MEMBER POMEROY: As we go through this, can 24 you place where it's obvious -- can you talk about the n
k ,) 25 appropriateness of the distribution of activities between NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 2344433
53 1 the two?
,m 2 MR. McCONNELL: I think the best -- maybe we t
(s)
~
3 should go back to the first viewgraph in that sequence. I 4 had a column of -- although it's not very definitive about 5 what's being done where, but I can give you -- the primary 6 area where the Center is giving the total system l l
7 performance assessment section a great deal of help right 8 now is the code development. ,
I 9 They've taken on the burden of the actual )
10 coding work and doing a lot of the testing and l 11 verification themselves. We are working quite -- Tim will 12 get into great detail, I think, about conceptualizing the l
13 code, what needs to go into the modules, what needs to 1
,n ,
('s ')
14 come out of the modules, what the testing program should 15 encompass so that we have confidence in the results, so ,
1 16 there's that split in activities. And I think Tim and Bob 17 Baca will probably get into more detail.
18 And that's -- they also are helping us in the 19 issue resolution status reports, too. So I would say 20 right now the distribution is 50-50 in those principal 21 activities we're working with the Center.
22 CHAIRMAN GARRICK: Marty?
l 23 DR. STEINDLER: You keep mentioning acceptance 24 criteria. What do you mean by that?
(3 25 MR. McCONNELL: Well, what we mean is
(/
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(;.02) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
54 1 basically for a particular issue or subissue, there are 2 certain levels of information that are necessary in order i
)
3 to, I guesa, prove the point, demonstrate compliance, with 4 a particular -- resolve the issue, depending on where you 5 are in the hierarchy.
6 At the KTI level, it's more, What is necessary 7 to resolve the issue. And there are acceptance 8 criteria -- when you get copies of the climate issue 9 resolution status report that has acceptance criteria in 10 it, you will see that there are specific milestones that 11 we would expect DOE to reach in order to demonstrate that 12 they have addressed the issue; we have no further 13 questions at this time on this issue; in essence, the
,e I
\__,, 14 issue's resolved; and we move forward.
15 DR. STEINDLER: Are those criteria 16 sufficiently well defined so that -- for example, you've 17 identified the uncertainties that you're requiring or 18 willing to live with, and does it go down to the detail of 19 how those data are obtained, especially experimental?
20 MR. McCONNELL: The first part of your 21 question: It -- I forget what the first part of your 22 question --
23 DR. STEINDLER: Well, they deal with l 24 uncertainties. To what extent do you also build n
I. ) 25 uncertainties into --
\s' NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
55 1 MR. McCONNELL: We are considering -- in 2 development of the acceptance criteria, we are considering 7_
b 3 that the uncertainties have to be addressed somehow.
4 Either they have to be -- the uncertainty has to be 5 bounded or whatevar approach is, you have to consider what 6 the uncertainty is. So it has to be considered. The 7 acceptance criteria will recognize that.
8 VICE CHAIRMAN HORNBERGER: If I could 9 interpret, what is being accepted is the way that DOE has 10 approached, let's say, the climate issue. Is that the 11 sense of acceptance?
12 MR. McCONNELL: Basically, yes. And I think 13 in the case of climate, we would say, This is an
(- 14
(. / acceptable method of considering climate, if you do these 15 things. To get the --
16 DR. HINZE: Carry that one step further. Your 17 resolve is not to solve the problem, but that you have no 18 further questions, although you may be in disagreement.
19 MR. McCONNELL: That could be -- well, I don't 20 know. I think the intent is to resolve the issue and to 21 have no questions, that basically we are in agreement on 22 approach --
23 DR. HINZE: Resolve it dtesn't mean that you 24 have no questions.
fN t,JI 25 MR. McCONNELL: I think we haven't resolved it NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
56 1 if we still have questions about either the approach or 2 the result.
,7 ~s 1 i
/
~
3 DR. HINZE: So you will not reach a point 4 where you agree to disagree?
5 MR. McCONNELL: Well, I think we've done that 6 in the past on specific issues, but I don't think we've 7 called that resolution. We have recognized that there are 8 differences in approaches that perhaps need further study 9 and further analysis, but --
10 DR. HINZE: You know, I'm thinking about 11 probability of volcanic activity.
12 MR. McCONNELL: Right.
13 DR. HINZE: There still are concerns about it, o
l !
(_/ 14 and I believe you have a deadline of the issue resolution 15 this fall. Does that mean you're going to -- that 16 somebody's going to give or be forced to give, to reach a 17 solution on that, or will you agree to disagree?
18 MR. McCONNELL: Well, I think that we're 19 coming together on the issue of how to approach 20 probability of volcanism, and I think there was a recent l l
l 21 letter sent to DOE, basica13y saying that we have reached l 22 agreement in the approach to considering probability of l
23 volcanism. I'm not familiar with the details.
I 24 DR. HINZE: That's again different than the --
/m (s-s ) 25 that's a methodology, not the results. l 1
NEAL R. GIUDSS !
COURT REDORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433
57 1 MR. McCONNELL: Well, I think we were looking n 2 at -- well, I think we already have agreed with PVHA in i )
w/
3 general approach, so the methodology to arrive at using 4 expert elicitation -- and correct me if I'm wrong, anybody 5 in the audience, Margaret, Norm -- but I think we agreed 6 on the apprcach to using expert elicitation in the 7 process.
8 So we agreed to the process, and we were, I 9 think, differing on how to use the results, and I think we 10 basically have coue to closure on how to use the results.
11 And I think there's a letter that just recently went out 12 from King Stablein to Steve Brocoum, documenting that.
13 Robert?
,n I
C') 14 MR. JOHNSON: Robert Johnson, NRC. I just 15 wanted to add to what Keith said. In the issue resolution 16 status reports, you not only have the acceptance criteria, 17 like Keith said. They lay out, you know, what we feel is i
18 needed to, you know, resolve a particular issue.
l 19 But there's also a rection of the status 20 resolution, and in that section, you know, relative to, 21 you know, what we've reviewed in DOE's program, you know, 22 that's where you would give a conclusion about are there 23 any further questions at this time. Either we're in 24 agreement with what, you know, DOE is doing right now, and
,O
) 25 therefore there are no questions at this time, or we may NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-370 t (202) 234-4433
58 1 still have some outstanding steas that we're working on
,r g 2 and those we have further questions on.
Q) 3 So there's two parts to an IRSR.
4 MR. McCONNELL: Margaret?
5 MS. FEDERLINE: If I could just add --
6 Margaret Federline -- what we're working towards is to 7 have acceptance criteria in the standard review plan that 8 would be available and would provide guidance to the staff 9 to conduct a review at license application.
10 What we're doing with the issue resolution 11 status reports is that we're working to put down on paper 12 what we know specifically about the resolution of an issue 13 at this point in time. At some point, we're going to have
-] 14
~
to have acceptance criteria; in other words, somebody's 15 going to have to sit at their desk and say, Either this 16 approach that DOE's taking is acceptable or it's not 17 acceptable.
18 We may not have achieved agreement on all 19 these things by the time of licensing, but the issue 20 resolution process is an attempt to put down on paper, 21 make it clear to all the parties where we stand on a t
22 particular methodology or a particular set of facts, and 23 iteratively move that forward to the point when we get to 24 the license application, we will have our sort of bottom
/x
/ \
( / 25 line of what we're going to use to conduct a review, and NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344 433
59 1 it will have been thoroughly vetted. People -- everyone
,s 2 will have an opportunity to know where we stand on the N'l 3 issues.
4 But definitely speaking, we may have an issue 5 resolution status report that comes out in '97. If some 6 new information in '98 or there's new information that DOE 7 has, it may get revised. And that's why Robert was 8 addressing the status report. He's -- that's the overall 9 concept.
10 CHAIRMAN GARRICK: That's all the more reason 11 for us seeing these IRSRs as they come out.
12 DR. HINZE: Could I --
13 CHAIRMAN GARRICK: Yes, sure.
Ch
(/ 14 DR. HINZE: I would like to ask the question:
15 Where in this schedule is the movement towards putting 16 your TSPA on PC and what sort of priority is being given 17 to that and why?
18 MR. McCONNELL: We're planning -- in talking 19 with the Center, getting it PC-based was put off until 20 next fiscal year, because we thought it was more important 21 to do the testing and validation effort this year, so we 22 could move on to the sensitivity studies. But we hope --
23 and maybe Bob can talk about this a little more. We hope 24 to have it on PC next year.
(x-~)\
( 25 Am I right, Bob?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. i (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l l
l
60 1 MR. BACA: Yes. That's correct.
,,_s 2 MR. McCONNELL: That's Bob Baca back there 5~'h 3 from the Center.
4 CHAIRMAN GARRICK: Keith, we're not as 5 disciplined as your technical exchange meeting where you 6 defer the questions until the end of the presentation.
7 But I would hate to be too far off schedule, so we'll put 8 some of the burden on you for getting us back on schedule.
9 MR. McCONNELL: I think I'll be done in one 10 more viewgraph. I just wanted to go very briefly, talk 11 about the allocated staff. We have a core group within 12 the section, the total system performance assessment ,
13 section, and also here at the Center that's very
('
(_)) 14 experienced. They were -- most of them were involved in 15 the IPA Phase 2 effort.
16 That core group is supplemented by now a 17 broader-based group of people, people who had worked on l 18 auxiliary analysis for the Phase 2 effort, and also this
( 19 broader group of KTI leads and other people that we're i ,
i :
20 bringing into the system.
21 And we also have Norm to advise us on the IPA i
22 efforts and keep us along the correct path to PA work.
23 The progress in the future or where we're 24 going is, again, to broaden the capability within the NRC
()
s 25 staff, and that's to make the code user friendly, make it HEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
i 61 l 1 available to the KTIs, have this inclusive and broadening l ~s 2 experience for the scientists and engineers at the NRC.
)
3 We also have been fortunate to have a couple 4 of new hires brought on board, and in general, they're 5 experience base is broader. They have more computational 6 skills in most cases, and when they do come on board, we 7 team them with an experienced staff member to bring them 8 up to speed fairly quickly.
9 And basically that's the end of my 10 presentation, and if there are no questions, I'll turn it 11 to Mike to talk about the computer resources.
12 CHAIRMAN GARRICK: One more question.
13 MEMBER POMEROY: One more question, if you'll
(- t i_/ 14 allow me. Can you give me some ideas of FTEs associated 15 with this? The third bullet is easy, because it's one.
I 16 The other two, give me some broad numbers if you can. I 17 have numbers of people I know in my head, but I'm not sure l
l 18 I know everybody that's involved.
19 MR. McCONNELL: Right. I would say in the 20 core group, that within the performance assessment 21 section, there are five to six people that are directly 22 involved in the total system performance assesement 23 effort.
24 We also fortunately are able to reach out and
(\
( ,) 25 rab people like Janet Kotra who's in a different section NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
62 l
1 to work on the rule-making. So, in essence, five, six, l
f3 2 seven range of FTE working on the broad issue of TSPA, and I] ~' l 3 that involves not only the high-level waste activities but 4 the low-level waste, BTP when it was under development, 5 and some of the STNP work and the uranium-recovery work, 6 so it's -- those five, six, seven FTE involve all that 7 work.
8 High-level is more like five --
9 MEMBER POMEROY: That sounds like what's in 10 NRC.
11 MR. McCONNELL: Yes. Within NRC.
12 MEMBER POMEROY: Are there any FTEs here?
13 MR. McCONNELL: I believe we -- and Bob maybe, I
t )
' \/ 14 I think, can help me out in this. I think it may be 15 approximately four or five people here devoted to total 16 system performance assessment, within his group. )
1 l
17 Is that correct, Bob?
l 18 MR. BACA: There are five people in the PA 1 l
l 19 group here.
20 MR. McCONNNLL: Okay.
I i
l 21 MEMBEE ;OMEROY: Thank you.
22 MR. McCONNELL: I don't know whether I was 23 clear enough. Was that clear?
24 MEMBER POMEROY: That's fine. ;
1 i,x -) 25 MR. McCONNELL: I'll turn it over to Mike Bell I
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
63 1 now.
,w 2 MEMBER POMEROY: Maybe while he's going up 3 there, you might add -- I see new hires mentioned here.
4 It's a little unusual these days to have -- at the Center, 5 are you expanding the program beyond this level that we're 6 talking about now, five or six here, five or six at the 7 Center?
8 MR. McCONNELL: Well, not new hires here. New 9 hires at the NRC staff.
10 MEMBER POMEROY: So you are expanding the NRC 11 staff.
12 MR. McCONNELL: Well, not totally. As you 13 know, we lost Rex Wescott, a big loss to the program, and I i k/ '
14 we were able tc backfill behind him. We also were able to 15 hire Christiana Lui from -- we stole her from a research 16 group and brought her in. But -- unfortunately, we just 17 lost one of our new hires. He went elsewhere. So it's 18 not that we're greatly expanding the staff. It's more a 19 replacement for people who leave.
20 MR. BELL: Good morning. I'm Michael Bell.
21 I'm acting chief of the high-level waste performance 22 assessment and integration branch. I've been acting in 23 this position for about two months now while John Austin 24 has been out working on the external regulation of DOE O
l 25 project.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 64 1 What I'd like to cover today is sort of the
, ,_s 2 third leg of the stool of our performance assessment
' (' ' ')
3 capability. Basically performance assessment requires the 4 people, the codes and the software, and then the hardware 5 to run it, and just very briefly, I'd like to describe to 6 you what we've been doing in that area, because some of 1
7 the questions are already touching on that.
8 Our performance assessment capability in the 9 high-level waste program at NRC originally was developed 10 for us within the office of research. Sandia National 11 Laboratory in the early '80s, with a contract through our 12 office of research, developed for us the SWIFT [ phonetic]
13 code which was the beginning of the NRC's capability.
,f )
(_)
2 14 The SWIFT code included module subroutines l 15 that are still in use today. The first versions of 16 NEFTRAN of the Latin hypercube sampling routines were part 17 of SWIFT, and basically as Keith mentioned, it was a code 18 that only a few people could run; it had to be run on the 19 mainframe.
20 Basically when it reached the point where it 1
21 was up and running and we wanted to get results, we would 22 have to tell Sandia, you know, Set up this way; run it A l
i 23 couple days later, some results would come back. We'd 24 look at them and say, Well, that wasn't exactly what we em (x- ) 25 expected, and, you know, it was a very long, slow, l NEAL R. GROSS
[
COURT REPORTERS AND TRANSCRIBERS i 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
l 65 1 laborious iterative process.
- 7. 2 Over time, we had been moving to bring this
> \ }
'~'
l 3 capability in house, but I think maybe it wasn't entirely 4 clear why Keith was emphasizing getting buy-in so much.
5 Basically we would still have this sort of few people who 6 were identified as performance assessment who worked with 7 the models and codes and ran the codes on mainframes, and, 8 you know, the people working in the sciences would not 9 always agree with how, you know, things were being 10 modeled, what the assumptions were that would go into the 11 codes.
12 And so one of the things we were trying to i
13 address is to get, you know, the earth sciences and the l
(~N
\
l
\_s) 14 material science staff all involved in participating and, l
15 in fact, in a position to use the codes that we're now 16 developing as we're bringing them in house.
17 Even as recently as IPA 2 -- and for those of 18 you who are new, IPA is iterative performance assessment. l 19 Basically IPA was the second round of our development of a 20 particular capability here at the Center, and we're 21 essentially now on IPA 3, the third iteration. The TPA 3 22 code is the name for the code that is this third round of 23 the iterative performance assessment.
24 And basically each round, we try to improve
(
l (,o) s.-
25 the models, add more capabilities. Computers have been NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
66 1 improving, so we can do things now that we couldn't have 73 2 imagined modeling 15 years ago. So we're moving in the e' i s t 3 direction of having an iterative performance assessment 4 methodology that could be run on workstations in-house at 5 NRC, essentially by all the technical staff involved in 6 the program.
7 The geologists can vary parameters and look at 8 the sensitivities and things they're interested in; the 9 material scientists, parts of the problem they're 10 interested in, working with the PA staff and then 11 basically the PA staff would look at the sensitivity of 12 various parameters and models to the total system 13 performance.
,/ y f \
(. / 14 The situation right now is that we rely 1
15 heavily on Sun workstations for the current generation of 1
l 16 the TAP code, but I think I'll probably address this on a I 17 later slide. The high-end PCs are now reaching the point i 18 where we expect to be migrating from Sun stations to PC-i j 19 based systems which has some advantages as far as l 20 integrating what we do with the rest of the agency,
! 21 because NRC agency-wide systems are essentially PC-based.
22 Basically the hardware that we're relying on 23 at the NRC is contained in what we call our Advanced 24 Computer System. This system was originally put together (3 25 for the high-level waste program.
t w )
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AYE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
67 i 1 We developed a number of capabilities: high-gs 2 end scientific computing, full-text imaging and document i ;
)
3 retrieval systems, 3-D geologic information systems that 4 basically we could model the hydrology, the stratigraphy 5 at Yucca Mountain, and show it in 3-D representation. We 1
6 could load into a data base the technical documents our l
7 staff would rely on in doing their review and do the 8 scientific calculation.
9 But this became -- once we had this hardware, 10 it became a resource for the whole division and 11 essentially the whole office NMSS. And we basically now i
12 have a large number of applications of software, 13 engineering codes, criticality heat transfer that's used i r~N c
\/ )
1 14 by the spent-fuel storage office, criticality codes that j l
1 15 are used by the fuel cycle division. l l
16 Even within our own division, the uranium-l 17 recovery branch is now starting to put their sites onto i 18 the 3-D data base or utilizing information system, so that 19 it can help them in their reviews of groundwater transport 20 at mill tailing sites. They're also loading their 21 technical documents into the CDOCS, full-text database 22 retrieval system.
23 So we have, we think, developed, you know, 24 quite a powerful tool that basically is in a room up on (n) us 25 the seventh floor of 2 White Flint that's half the size of I
l NEAL R. GROSS
( COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N.W.
l (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) .T4-4433
68 1 this. It has a couple of servers, plotters, color
,-w
, 2 printers, scanners for being able to scan in maps and V 3 cross-sections of technical data, and, you know, it is 4 bringing in-house and, you know, allowing us to be much 5 more efficient in the way we do our work.
6 I mean, essentially the staff using the TPA 7 code can set up a problem at night before they go home.
8 It runs overnight. They come in the next morning, and 9 they have results. It's not waiting for three days for a 10 contractor to run it on the mainframe somewhere.
11 Just to emphasize the point I was making about 12 all the use this gets, you'll see we have some 50 users in 13 the division of waste management, substantial number of
, ,7,x i !
%/ 14 users in fuel cycle, in the spent-fuel project office, and 15 even users in other parts of NRC, because we have the only 16 capability in the Agency, for example, for this 3-D l
I l 17 geologic information system.
1 18 Just to close, where we'd like to see it go:
19 We moved in this direction because we were filling a void. 1 1
20 The NRC just didn't have in-house the high-end scientific 21 computing capability we needed, the full-tech search and l 22 retrieval that we felt we needed, the geologic information 23 system capability that we needed. So essentially we 24 developed this for our program.
/'~'s
( ,) 25 The Agency is now beginning to catch up.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344 433
69 l 1 There will be what's called an electronic hearing docket
! 2 where essentially in the future, the entire docket for a l (7_
' ' ' ') 3 licensing proceeding, including, we anticipate, the high-4 level waste repository application, will be put into an 5 agency-wide full-tech search and retrieval system and be l
6 made accessible to all the participants in the proceeding.
7 So what we see, for example, is when the 8 agency is doing things like this, we would eventually plan 9 to phase out our system and transfer to the agency-wide 10 systems. The -- as I mentioned, we rely very heavily on 11 the Sun workstations, but they're incompatible with the 12 rest of the agency. That means when we have a problem, l
13 it's often very difficult to get IRM to help us solve it,
! )
' \ _. / 14 and we're essentially left on our own work it out. l l
1 15 And, you know, to move towards PC-based system l l
16 would make us compatible with the rest of the agency, and 17 we see an advantage in doing that. And it also appears 18 that it will be cheaper, which is a big consideration, 19 because I hope I have been able to impress on you how 20 useful and how valuable this system we've developed is, 21 not only the high-level waste program, but our whole 22 division and our whole office.
l j 23 Yet for some reason, you know, the budget for 24 computing and software development is something that the (n)
%./
25 budget-cutters always seem to be -- one of the first NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
I (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
70 1 things they look at to cut. They don't feel they can cut
- 2 license reviewers or things that would be -- or even
/ ,)
3 contracts to assist in the review of a license, but for 4 some reason, they hardware and the software that you might 5 be using to do that review is a target.
6 So if there is one area C'd like to mention in i
7 response to Dr. Garrick's question earlier, this is an 8 area where -- that we could use your support.
9 DR. HINZE: I'm surprised, Mike, that you 10 didn't mention putting the -- in terms of moving to the PC 11 base in the system, putting this in the hands of the 12 engineers and scientists.
13 MR. BELL: Well, I thought I did.
l (m )
L/ 14 DR. HINZE: I'm carry. I didn't --
15 MR. BELL: And not only that; the other 16 advantage we see to see PC-based system is it will make it 17 easier, for example, make it accessible to the state, 18 local --
l l
19 DR. HINZE: More accessible. Sure.
l 20 MR. BELL: -- governments, other people who 21 are involved in the program.
! 22 CHAIRMAN GARRICK: Paul?
23 MEMBER POMEROY: Mike, I wanted to return to 24 your last comments, because I think they were very n
(w- ) 25 important. I presume this ACS network, as you've NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
71 1 diagrammed it here, more or less exists now. Is that 2 correct?
'n)
/
3 MR. BELL: That's right.
4 MEMBER POMEROY: And my questions really go to 5 the fact of whether you have the resources available, both 6 for software development capability, for instance, faster 7 networking, for example, along -- of this activity.
8 But more specifically, do you have the 9 resources available or what do you need to have in terms 10 of hardware and software in order to meet the time 11 schedule that you have to meet: namely the time frame of 12 doing the appropriate reviews at the appropriate points in 13 time, as we understand them at this point?
l
(/ 14 MR. BELL: Well, I mean, if the snapshot in 15 time you're looking at is our review of TSPA-VA and --
16 MEMBER POMEROY: That's certainly the first 17 point. l l
l 18 MR. BELL: -- or, you know, what we'll be ;
19 doing in the next year or so, certainly we're going to l
20 rely on this system. I 1
l 21 The life cycle of this hardware is about three 22 years. I mean, already the server that we're using for 23 the geologic information system is obsolete. It's no 24 longer supported by the manufacturer. We would like to
,a
() 25 replace that server and, for example, haven't been able to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISL AND AVE., N W.
(202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 2344433
72 1 get the funds to do it.
7- 2 MEMBER POMEROY: The RS6000 is relatively.new.
k' ~ '
)
3 MR. BELL: Yes, I guess there is one last 4 slide in this that deals with the Center capability, and 5 some of the same systems that we have are duplicated down 6 here at the Center. In fact, I think on your tour on 7 Friday, the plan is to show you, for example, the 3-D 8 geologic information system and how it's used.
9 I don't know if, you know, you want to see a 10 CDOCS demonstration. You can search for all the ACRS or 11 ACNW letters that have been written on a particular topic 12 in the last several years or --
j ,
13 CHAIRMAN GARRICK: By the time we get through l l 1
\_ / 14 with them, we never want to see them again.
i l 15 MEMBER POMEROY: Mike, I heard some mention of l 16 parallel virtual machines as an interim step in this 17 process. Is that something that's being considered at 18 all?
l 19 MR. BELL: Yes. As a matter -- in fact, I l 20 think that is --
l i
21 MEMBER POMEROY: There's some parallel l
22 computing software.
l 23 MR. BELL: Yes. One of the things that we 24 have been looking at in this transition from mainframe to tr%
(,x ,) 25 PC is, you know, using a number of workstations in NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
f 73 1 parallel to essentially speed up the calculations. And 2 Dick Codell was the staff member who's been most involved (7
%j w) 3 in that. He looks eager to speak to it.
4 MR. CODELL: I'm Richard Codell. We've used 5 parallel virtual machine twice, once on a low-level waste 6 performance assessment and more recently in some )
l 7 calculations we did in-house in support of the NAS l I
8 recommendations for the rule.
9 We'd like to use it here for the current i
10 ground. The structure of the code is being mainly aimed i 11 toward a single workstation, and use by many people does )
l 12 not work well in this kind of environment. However, we're '
13 talking about making the modification somewhere after the
/ ,s (x/ i 14 current round of work we're doing to set it up so we could 15 run it on our RAY workstations, mainly because right now 16 we're facing some very long run times on the --
17 To get through several hundred cases, it's 18 likely to take more than overnight, and in order to get a 19 large number of runs which we'll need for our sensitivity 20 studies, we're seeing that we'll have to find some more 21 computer power, and this is one way we're hoping to 22 address it.
23 MEMBER POMEROY: Great. Thank you.
l 24 CHAIRMAN GARRICK: Any other questions?
l /'N.
25 (No response.)
HEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
74 1 MR. McCONNELL: I think, Dr. Garrick, we're at
, 2 your break point. This is basically the discussion of the
(' ' '
)
3 front end part, the organizational structure, the overview 4 of our activities, and we will intend to go to the TPA 5 code work following your break.
6 CHAIRMAN GARRICK: All right. In that case, )
l 7 we'll take a 15-minute break. l 8 (Whereupon, a short recess was taken.) )
9 CHAIRMAN GARRICK: All right. I think we want i
10 to come to order. I think before we continue, Keith, I I 11 want to make one simple announcement and that is that the 12 tour that was planned for Friday, we have rescheduled that 13 for Thursday, between five and six o' clock, which gives us s s I
1
(
(,,/ 14 great incentive to finish on time. And we did that 15 because we were losing so many people that we felt that i
16 would be much more accommodating. j l
17 So people that are interested in the tour and 18 that are involved, with that information, will maybe make 19 some other travel arrangements. So that's Thursday 20 evening at five o' clock.
21 Okay.
l
[ 22 MR. McCONNELL: If I could, I'd like to 23 clarify one point that I brought up in talking about 24 volcanism. King Stablein -- is King back at headquarters (n) s.-
25 yet?
I NEAL R. GROSS
! COURT REPORTERS AND TRANSCP:BERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 l
l 75 i l )
l 1 MR. STABLEIN: I'm here.
l ,s 2 MR. McCONNELL: Okay, King. Would you like to
'( )
\"# 3 go ahead and clarify about the letter to DOE?
4 MR. STABLEIN: Right. Thanks, Keith.
5 Good morning, everybody. This is King 6 Stablein at the NRC. There was a reference made to a ,
1 1
7 letter that I hope to be sending to Steve Brocoum of DOE I I
8 about the agency's views on volcanism and the probability l l
i 9 of volcanism. 1 10 I wanted to clarify that that letter hasn't 11 gone out yet, and hence it might be a little premature to j 1
12 predict all the details of what will be in the letter, and i I
13 as Keith said, the agencies are beginning to move closer n
) i
\_/ 14 together on resolution of various issues in determining j i
1 15 the probability of volcanism, and the letter will reflect i'
16 that develcpment.
17 Are there any questions on that?
18 DR. HINZE: The anticipated date?
19 MR. STABLEIN: Well, I had anticipated a week 20 ago, as Keith's remarks indicated. I would say imminent.
21 MR. McCONNELL: Okay. With that, Dr. Garrick, 22 we'd like to proceed into the description of the TPA 3.1 23 code, the general approaches, and with that, Bob Baca, l 24 who's the element manager here at the Center for total l (r~T
! ) 25 system performance will make the presentation.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
76 1 MR. BACA: Thank you. In this presentation
,s 2 I'm going to try and give you an overview of the TPA codes t )
\'# 3 and of the essential elements of this code. We've been 4 working on the codes since November, I think about five l 5 months of development activity and about four months of 5 testing activity.
l 7 This has been a very intense effort by a i
8 number of people, both in my group and also support from 9 the NRC staff who contributed to parts of the code, and 10 we've also drawn on the resources of the Institute to help l 11 us.
i l 12 These are some of the ideas that went into the l
- 13 design of the software, some basic principles. We wanted 1 (~s\
i
(
! (_,/ 14 to build on the methodology, the calculational methodology 15 of IPA Phase 2, which basically outlined a six-step 16 process for doing performance assessment. We've 17 implemented that and the calculational approach in TPA 18 3.1.
19 We have retained in this code calculation of 20 outputs, performance outputs, from the Phase 2, and we did i 21 that primarily so we could have a comparative base between l l
22 this code and the outputs from the previous. We have 23 added peak dose which replaces the calculation of the 24 population dose which was don in the Phase 2 code.
j%
(/)
w 25 One of the major changes to this code was to l NEAL R. GROSS l COURT REPCRTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
l l 77 1 accommodate the DOE repository design and in terms of
~
2 their emplacement mode, the characteristics of the waste I ')
3 package and thermal-loading options.
4 The code has benefitted from a lot of products 5 that have come out of the various KTIs. They have 6 provided enormous amount of the knowledge base that's 7 coded into the program, a lot of data, and some of the 8 submodels, in fact, are incorporated into this code.
9 Finally, I think as has been mentioned by some 10 of the other speakers, we try to come up with a code that 11 a lot of people could use, unlike the Phase 2 code which 12 was really a code for specialists. And that code was -- !
13 the Phase 2 code was developed with a rather narrow type e'x 1
( )
x_/ 14 of set of objectives, and when we started this development 15 effort, we realized that a lot of changes would have to be j 16 made in order to address some of these things, like i
17 changes in design, the expectation of a new standard, et l l
18 cetera, et cetera.
19 These are -- this is a partial list of some of 20 the major improvements and changes to the code, relative 21 to what was in Phase 2. Phase 2 code really was strictly 22 designed to look at 104 years in accordance with the old 23 standard. The new code allows you to look at an arbitrary l
- 24 tim period of interest, so you can look at 10,000 years,
/~N
() 25 50,000 years, 100,000 years. So everything from the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
78 1 source term to the calculational approach is now versatile
,3 2 and can accommodate just about any time frame.
~
3 In Phase 2, we treated climate as a scenario, 4 as a binary state. Either you had current conditions or 5 impluvial. In the new code, we're using an approach where 6 we prescribe the precipitation history, and so we treat it 7 more as a known type of quantity with some random 8 perturbations. And I think probably Tim in his part of 9 the talk may give you more details.
10 We have new and updated abstractions for 11 volcanism. The igneous activity KTI has provided a lot of 12 input to one of the modules, a couple of the modules 13 actually, that's incorporated in the current code, t
's /
14 We have a new approach for calculating plume 15 dilution, which was not in the Phase 2 code, because the 16 standard was based on total release and dilution was not a 17 consideration then. j 18 We have versatility in our ability to describe l
19 uncertain parameters. We have a larger library of J 20 statistical distributions that you can assign to input i
l 21 parameters.
22 The outputs from the code are quite diverse. ;
I l l 23 They allow you to not only look at the bottom line but 24 allow you to work back to what led to that bottom line, so
( ,/ 25 that you can interrogate what's happening in this sequence
( NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433 l
l
i 79 1 of calculations that the code is doing. And I think that 73 2 has probably the greatest value, because it's not just the
( )
!\# 3 numbers the code generates; it's the understanding as to 4 what led to those and our ability to articulate the 5 reasoning and basis for those outputs.
6 The last point, I think, has been stressed by 7 the earlier speakers: user friendly. In order for a 8 diverse group of scientists and engineers to use a code, 9 it has to be accommodate their nature and also has to 10 be -- to help them out in the sense of checking the inputs 11 and giving appropriate feedback in the event that they 12 have inadvertently made some incorrect input.
13 This is a simplified flow chart of the 79
\
\_/ 14 computer code, the version 3.1. It's actually quite 15 similar to the Phase 2 with some exceptions, some of which l 16 I've mentioned. We don't treat climate as a scenario. !
17 These are the scenario classes that we currently consider j i
18 which are faulting, the impact of that on waste package 19 damage; seismic effects, creating rock falls, and again 20 impacts on waste package, and then volcanism treated in 21 terms of what's ejected into the air and the damage on 22 waste packages as an excursive would -- as a magma would 23 pass through the repository.
24 Another new part of the computer code is the
()
,m 25 treatment of that air dispersal of the contaminated ash.
HEAL R. GROSS l
COURT REPORTERS AND TRANSCRIBERS
' 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
l 80 1 I'd like to mention that those modules were originally
! 2
,, -. developed as part of Che PA research project, so here is a
( )
3 direct benefit of one of our research projects, as is the 4 volcanism module which is -- I would attribute a lot of 5 credit to the igneous activity KTI and to one of the 6 earlier research programs.
7 The container lifetime source term KTI 8 contributed the EBSPAC code which is represented here by 9 the EBSFAIL/EBS Release modules. We have used the NEFTRAN 10 Q code to model the UNSAT zone and also the SAT zone, and 11 one of the things that we have now is dose modules to 12 consider both the groundwater pathway and the other 13 pathways.
,y l i
(_/ 14 This table gives you a breakdown the various 1
15 modules, the major modules, and what I'm trying to depict 16 here is more the nature of the abstraction. You have two 1
17 options to abstract -- to do abstractions. l l
18 One, you can look at the mathematics and 19 simplify them to -- from partial differential equations to 20 ordinary differential equations, so that you have an 21 ordinary DE where you're solving a lump parameter kind of 22 model.
23 The other is to do very detailed modeling and 24 abstract the results and capture those in a table look-up
, ^s.
( ,) 25 or a response surface. And what this table tells you is NEAL R. GROSS COURT REPORTERS AND TRANGCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234.4433
El 1 that we have gone -- we have emphasized more the
,g 2 abstraction of models as opposed to the abstraction I
'~'
3 results.
4 And the reference here to external stand-alone 5 programs, really all that's saying is that those modules 6 originally developed as independent programs and then were 7 woven into the TPA code.
8 In the minds-eye of the computer, what it sees 9 of the repository and hydrogeology is a depiction as 10 illustrated here. We take the repository area and 11 discretize it or break it up into a series of subareas, 12 and we model what happens in each one of those in terms of 13 the waste package performance, in terms of near-field
(
r"x <
'w_/ 14 conditions, in terms of water moving through that portion 15 of the repository.
16 Then we trace the movement of that material I 17 downward through the unsaturated zone, and we're using a I 18 one-dimensional characterization of that transport, flow 19 and transport phenomenon. It feeds a flow tube in the 20 saturated zone which then carries the material out to a 1
I 21 receptor point.
22 In our basically 1-D calculational approach, 23 we have woven in some two-dimensional detail, and we've 24 done that by modeling the groundwater flow, the planer l
(f--)x 25 flow, using a 2-D groundwater flow model, and then we've NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
82 1 traced out the particle trajectories or the streamlines,
,- 3 2 and then have mapped out the dimensions of our flow tube i )k
\
~-
3 in accordance with these 2-D calculations, 4 And so what we do is we're able to represent 5 the movement from selected -- the movement of contaminants 6 from selected subareas downward into the water table and 7 out to the receptor point, so if there is one subarea 8 that's dominant, that's controlling the performance, we're 9 able to sort that out.
10 This is a listing of some of the TPA code 11 outputs. Of course, the primary one is to get a CCDF or 12 CDF for peak dose. The code has great flexibility to 13 disaggregate ar to tear apart a lot of intermediate U 14 calculations. You can look at the calculations in terms 15 of a scenario class.
16 For example, you might want to know what's the 17 specific impact of seismicity in the rock falls and damage 18 of containers. What does that do to the CCDF? Or you l
19 might want to look at what happens in terms of performance 20 of the SAT zone components or of the near field, et 21 cetera, et cetera.
22 So it gives you greater flexibility to tear 23 apart the problem, to understand the outputs or the 24 results in terms of everything that came together to cause
/~N 25 that output.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
83 1 In computer modeling, you're always struggling
,cs 2 to find a balance between simplicity, which makes it easy
( )
3 to implement and fast in the computer, and complexity 4 because you want to capture a lot of detail to make it 5 realistic and more representative of the physical setting.
6 We've made a number of assumptions in 7 stylizing our approach to total system performance, and 8 some of them are listed here. For example, we neglect y 9 lateral diversion of flow. As you, I'm sure, observed 10 from our little schematic, we model the geology from the 11 repository downward. We do not model above the 12 repository.
13 So what we have to do is take the infiltration i
f% \
V 14 distribution and project it down to that repository 15 section, and in doing that, treating it as 1-D, we neglect 16 any lateral diversion. In some sense, that's a 17 conservatism. There are some conditions where there could 18 be lateral diversion from outside our column into the 19 column, but we're not considering that.
20 DR. HINZE: Do you consider it constant 21 infiltration then?
22 MR. BACA: We do not. The infiltration varies 23 from subarea to subarea, and it'll vary as a time history.
]
24 DR. HINZE: How -- what's the similarity C,~3 25 between what you're doing and the work that fellow is NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005 3701 (202) 234-4433
84 1 doing at El Real.
,_ 2 MR. BACA: The work that Bo Bovardson is doing I
^'
)
3 with his TUFF 2 code is he has a very, very fine rain 4 resolution, so he has a very dense grid, so he's looking 5 at dimensionality of -- 3-D dimensionality of flow in 6 multiple directions, temperature changes, that sort of 7 thing.
8 Ours is a much more coarse-grained. We're 9 looking at large subareas. In other words, we partition 10 the repository into like seven sections, whereas he has 11 thousands of grid notes, so he's capturing a lot more 12 detail, but he can only do that in a deterministic sense.
13 He's not able to propagate uncertainty.
k) ms 14 So the advantage of the approach that we use 15 is that we can consider many more processes and also 16 consider uncertainties associated with either the input 17 parameters or spatial grid drawings. Okay?
l
- 18 DR. HINZE
- Thank you.
l 19 MR. BACA: Concurrent with the code i
l 20 development is the effort to prepare the database that's
- 21 going to fuel this knowledge engine, and the idea was to 22 capture in one document our data set.
23 It's one thing to get agreements on codes, but 24 in order to get agreement on results, you have to know
(\ what fed that code, so we have taken the IPA Phase 2 data
(,,v) 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
85 1 set as a starting point, and we've started to weave in all 2 of the new information, new data from DOE, and try to l
i
'/
3 upgrade to make more current that reference data set.
4 We issued a letter report that gave us a straw 5 man. We've interacted with KTI teams, and they've come 6 back to us and said, Well, you need to change the range of l
7 values for this parameter, or you need to change the l l
8 parameters in this other category. So we're drawing 9 collective knowledge to come up with a consensus on what's 10 the reference data set that we're going to use in our l 11 Phase 3 calculations. ;
12 Our plan is to, after we complete this 13 scrubbing process of our reference data set, is to issue a
( 1 k s) 14 Center report which would go into thc PDR and would also
)
15 be provided to DOE, so DOE knows what our thinking is on 16 appropriate or reasonable assumptions for a lot of these 17 model parameters. And that's going to be very helpful in 18 us ferreting out differences, differences in input 19 assumptions and also later on ferreting out differences in 20 modeling approaches and abstractions.
21 And that's what you need to put on the table 22 to work towards some type of resolution.
23 CHAIRMAN GARRICK: Bob, on the issue of 24 comparisons, there's a number of ways, of course, to do
(\
( ) 25 this. One is to compare -- make comparisons of complete NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20C05-3701 (202) 234-4433
86 1 computer programs, between different codes. And the other
,, 2 is to make comparisons of what you've with physical i \
(
'"l 3 reality.
4 Have you done any of that or much of that, to 5 increase your confidence in the code? And if you've done 6 that, do you have any sense of the extent of the 7 conservatism of the analysis? Are we talking about a real 8 upper bound analysis, 95 percent -- do a realistic risk-9 based model?
10 NR. BACA: Right.
11 CHAIRMAN GARRICK: A couple of questions 12 there.
13 MR. BACA: Yes. In terms of the comparisons
.rw
( \
C/' 14 that we have done, for selected modules, we have done 15 comparisons with the variable data. I think an excellent 16 example is the work that Britt Hill has done with the 17 ASHPLUME model, comparison with field data from a volcano 18 in Nicaragua. The name of it escapes me.
19 VOICE: Cerro Negro.
20 MR. BACA: Cerro Negro. And, you know, that's 21 very comforting, to see a very close correlation between 22 what you predict and what you can observe.
23 I don't believe we could ever check out the 24 whole code with any kind of field data because of the time
(\)
v 25 and space scales that we're talking about. But I think we NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
87 1 can select a module basis, develop the confidence or a
,-m 2 reality check to make sure that we're in the ballpark.
i \
\"'/
3 Another thing that we will done or plan to do 4 is to compare our simplified abstractions with some of the 5 detailed models, to make sure that we have captured first 6 order effects.
7 CHAIRMAN GARRICK: Are these comparisons, 8 these validation exercises, going to be in this 9 documentation, or are they in some form of documentation 10 now?
11 MR. BACA: The next document that will come 12 out will be the users guide for the TPA 3.1 code. It will 13 contain descriptions of the mathematical models. It'll l ,m
'( )
\, s'
_ 14 contain listings of major assumptions. At this point, 15 it's not our plan to have comparisons of modules with data 16 with other codes.
I l
17 We do plan, however, to do a software l
18 validation activity, where we'll write a validation plan.
19 This is software validation, not model validation, where 20 we'll lay out a series of test cases. Some of those we 21 may draw from some of the code -- international code i 22 comparison work; some of that from the open literature.
23 And then we'll run the code through those set of test 24 cases and document that in a separate report.
(p) v 25 CHAIRMAN GARRICK: Thank you.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
98 1 DR. HINZE: Bob, one of the objectives, of e- 2 course, is to have an independent performance assessment
J 3 tool. Can you give us some ideas of how -- what are the 4 physical ways in which the performance assessment system 5 that you're developing is different from that of the TSPA?
6 MR. BACA: I guess based on the last couple of 7 days, the tech exchange that we had with the Department of 8 Energy, I did get a sense for their approach that they're 9 going to use in the VA, and I think the things that stand 10 out in my mind is the DOE are going to put greater 11 reliance on very detailed, multi-dimensional models in 12 their TSPA-VA, whereas we're using a much more what I 13 would call truly abstracted type of approach.
, / \
! 's / 14 They have -- you know, that's their choice to 15 use the higher, more detailed models like their FEM 16 [ phonetic] 3-D, and their TUFF 2, and if that helps them i
1
- 17 make their case, more power to them. My general feeling i
! l 18 is that the data demands to drive dimensional codes like 19 that are much greater than they are for the approach that !
t l 20 we have taken.
l l 21 I can tell you that our data set considers or
- i 22 consists of maybe about 1,000 parameters, and it's quite 23 an effort to work with the different disciplines to make 24 sure that we have appropriate ranges and representations n
( ,) 25 for those parameters.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERE I 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
89 1 Now, imagine the data requirements for a set 1
-s 2 of 3-D, multi-component, multi-phased codes. It just l
/ 'T l
\
' 'j 3 would go up tremendously, and if they are able to come up l l
4 with the data set to justify that more detailed approach, j 5 you know, then they should be complimented. i l
6 VICE CHAIRMAN HORNBERGER: Given that l l
7 difference, should there be differences that you observe 8 between the 2-D and 3 and either TSPA-95 or TSPA-VA, how 9 will we determine how to -- where to attribute the 10 difference?
11 MR. BACA: Well, I think the best way to 12 answer that is to give you an example. As part of our 13 pre-licensing activity, we have done audit reviews,
(_ / 14 detailed reviews of past TSPAs done by Department of 15 Energy. In the audit review, for example, we did a 16 calculation; we did a CCDF for cumulative release.
i 17 And we compared that to what they published, a 1
18 particular case, and then what we did was we took their l
19 data and their assumptions and made them as comparable as l
20 possible to our code, and then made the run, and then l
21 looked at the outputs. We looked at the CCDF with their 1
22 data and with our data, and the differences in those 23 curves really could only be explained in terms of the 24 characteristics of the model.
c f ) 25 And one of the things we did in the audit w/
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS I
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 )
i 1 i
90 1 review was then to try to understand that difference, to
,_s, 2 see what caused it, and as we dug into the approach, the
(
)
3 mathematical approach that DOE had used, we discovered 4 that a lot of it was largely due to the approach that they 5 used for describing mass transport.
6 They had used a mathematical model which was 7 based on a Markovian approach which when you tore it 8 apart, basically what it did was it did not account for 9 fast paths. So these are the things that we do, to try to 10 understand the differences.
11 The differences are important, but then we 12 have to tear them apart to see whether it was a bias built 13 in to a modeling approach or whether it's due to a
\' _s'4 14 difference in input assumptions.
15 VICE CHAIRMAN HORNBERGER: Go ahead. j 16 MR. CAMPBELL: What fraction of the, I guess 17 would call them, subroutines that you're using have seen 18 the light of day in some kind of a peer review, generally l
l 19 accepted publication?
i 20 A I think of one, just offhand, that's seen 21 journal publication, and that's the air dispersal model, 22 the ASHPLUME. Mark Jarzemba wrote a paper on the 23 theoretical approach, which I think has been published or 24 will be published in Nuclear Technoloav. Some of the work
/~x 1
(
x ) 25 that Britt Hill and Chuck Connor have done, in conjunction NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
1 1
91 1 with that model, I know has led to some publications.
l '
,7 2 Most of the modules have been documented in
, ( ) I i " 3 users guides, have been documented in Center reports, but 4 a very small percent has been put into the scientific 5 literature. '
1 6 MEMBER POMEROY: Can I ask a question?
7 CHAIRMAN GARRICK: Sure. l I
8 MEMBER POMEROY: Bob, with regard to the i
9 abstracted models, you talked a lot about comparison of i l
)
10 parameters and so forth. One of the problems that's 11 always bothered me a lot is how you determine whether or 12 not ycu have preserved the full range of uncertainty 13 between process model and the abstracted model. i
,e bl 14 In the abstracted models that you've done to !
15 date, have you gone back and verified that you are 16 preserving the full range of uncertainty? And if so, 17 could you take a minute to just enlighten me a little on 18 how you do that? I'm sorry it's just a minute or so.
19 Maybe we could do this somewhat later, but a little bit 20 here.
21 MR. BACA: I guess I'll have to give some 22 thought to that.
t I
l 23 Tim, can you think of a case where we've 24 examined the preservation of the range of uncertainty from
,c3 Q) 25 the process level model to the abstracted mode.l?
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS
! 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 '202) 234-4433
92 1 MR. McCARTIN: The closest could be
~ 2 potentially what Gordon did with the 2-D saturated zone
!s)
3 that goes into the abstracted model. I mean, that 4 particular -- in the saturated zone flow portion, you've 5 got the -- a 2-D detailed model that is the velocity, the 6 variation, the velocities consistent with the heads; that 7 uncertainty is brought into the abstracted model that we 8 have in the TPA code is one. I don't know if that's --
9 it's close.
10 Is that --
11 MEMBER POMEROV: Let me phrase the question 12 differently. In all the abstracted models you've either
. 13 now done or are going to do, do you intend to run some
(~T l x_-) 14 sort of a comparison to determine whether or not the full 15 range of uncertainty as expressed in the process model is 16 carried through in the abstracted model?
17 MR. BACA: The -- one example that I can think 18 of where we will do that in a very concrete way is like in 19 the refluxing, the description of the refluxing 20 phenomenon. In the TPA 3.1 code, we have a lump parameter i 21 description of that process; that is, the build-up of 22 water above the repository and it having to come back.
23 There are plans in the TEF KTI to run the 24 multi-flow code which is very similar to what Bo Bovardson r-( ,x) 25 is using, and we'll examine whether our abstraction is in NEAL R. GROSS COURT REFORTERS AND TRANSCRIBERS 1323 fiHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
2 1
93 i
l 1 concert with what you would get with a more detailed f s, 2 model. And I think -- I think we had a comment from
~
3 Norman.
4 MR. EISENBERG: This is Norman Eisenberg 5 again. I think it's true that we don't always preserve 6 the full range of uncertainty. And this is on purpose, 7 because we want to be able to have an abstracted model 8 that's practical, so that in some cases, such as in the i
9 modeling of the waste package, we are purposely going 10 through and eliminating fixing some uncertain parameters 11 as constants, so, of course, we will not be looking at l
12 that element of uncer:ainty, but we choose parameters that 13 have minor effects on the output. And so it's a means of 7
w- 14 simplifying the problem.
15 So we think we're preserving the essential 16 uncertainty, not necessarily the total oncertainty, but 17 the important and essential uncertainty.
18 Another example might be in the dose portion 19 of the model where we use dose conversion factors that are 20 functions of all kinds of parameters, and I know one of 21 Tim's favorites is the -- how much watering of grass that 22 you do.
l 23 Well, it produces a change in dose. It's not 24 something that we want our whole analysis to hinge on, and n
() 25 it doesn't change the dose that much. So some of those NEAL R. GROSS
! COURT REPORTERS AND TRANSCRIDERS i
1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
94 1 parameters we will fix also.
,_s 2 So I think we're using a reasoned approach
/ \
' '/
3 to -- as I say, treating the essential uncertainties, but 4 I don't think you should come away with the view that we 5 have captured all the uncertainty in the abstracted model.
6 MEMBER POMEROY: Then clarify a little bit 7 more for me what you mean by essential uncertainties. Am 8 I -- first of all, perhaps you could just tell me if I'm 9 wrong with regard to the fact that the purpose -- one of 10 the stated use guides or whatever of the abstracted models 11 is that they do preserve the essential collaboratives or 12 the essential uncertainties. Do you think that's true?
13 MR. EISENBERG: Yes.
(ms/ 14 MEMBER POMEROY: And so by essential 15 uncertainty, are we talking about 98 percent of the 16 uncertainties? Or 43 percent? Or --
17 MR. EISENBERG: I would hesitate to quantify 18 it.
19 MEMBER POMEROY: I understand your hesitation, 20 and mine was facetious. But are we trying to preserve 21 everything that is essential? How do we know we're 22 preserving --
23 MR. EISENBERG: Well, one way we know is by 24 working on the modules in a sensitivity analysis mode and r
25 by developing the modules. The analysts have a feel for
(
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
95 1 what drives the model and what drives the output, and so gs 2 it's not a big black box that we're dealing with. There's
/ 3
)
3 quite a bit of knowledge of how the model works and how 4 the code works compared to the model and what are the 5 important parameters and which ones have uncertainty that 6 needs to be carried forward into the abstracted model.
7 CHAIRMAN GARRICK: Tim?
8 MR. McCARTIN: Yes. Tim McCartin, NRC staff.
9 If I give maybe one example, say, in the hydrologic 10 cystem. The fracture porosity is a value that there is 11 uncertainty with it. What that number is -- the test 12 information is going to be certainty variable.
13 However, when you have fracture flow, you have
,/ -
\/ ~ 14 relatively rapid transport, so we might fix that 15 particular fracture porosity rather than sample to a mean 16 value, and potentially the transport time now, let's say, )
17 for a particular -- transit to a particular unit might be 18 ten years. In reality, maybe if we considered the full 19 range of fracture porosity, maybe it's five to twenty 20 years.
21 Well, along the lines of what Norm was saying, 22 does it really matter whether it's five years, twenty 23 years? Ten years is good enough. Those are the kinds 24 of -- some of the things that we look to do with the code.
'/ 'N We're clearly decreasing that uncertainty, but we don't
) 25 a
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
96 1 believe we're affecting the final outcome in a large way.
2 CHAIRMAN GARRICK: I think there's a lot of l,,_x) s ,
3 questions about the uncertainty issue, and hopefully at 4 our working group session tomorrow, we'll be able to dig 5 into those a little deeper.
6 MR. McCONNELL: What we're going to do now is 7 to move on into a lower level into the code and discuss a 8 module-by-module description. Again, this follows -- if 9 you remember the diagram, it follows the components of the 10 subsystem, and Tim will be talking to each of those 11 subsystem components.
12 And I believe also that he does talk about in 13 each module some of the uncertainties that exist within 7-.
I \
E._) 14 each of the modules and how it's propagated.
15 CHAIRMAN GARRICK: Keith, one of the items 16 that we wanted to hear something -- and maybe -- it seems 17 like something was said, but it might have been yesterday 18 or the day before -- and that is the schedule with respect 19 to the development of code and the milestones. Is 20 somebody going to be able to give us some insight on that, 21 such as, for example, when to expect the PC version, when 22 to expect some other milestones desociated with TPA 3.1 or
! 23 may be other issues of TPA?
l 24 MR. McCONNELL: I touched on it in my talk,
,m.,
I 25 but maybe after we get done with the module-by-module (J
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
l 1
97 1 description, I can in the meantime develop a diagram that
, .s 2 lays everything out.
! i i
)
3 CHAIRMAN GARRICK: Okay. Thank you.
4 MR. McCARTIN: Okay. I'll try to briefly go 5 through the abctractions for the TPA 3.1 code. Some of 6 the slides are exactly the same as you might have seen in 7 the technical exchange. I have added some, so don't feel 8 cheated, and I will try to brush over some of the things 9 that I think we went over in enough detail in the 10 technical exchange, but please feel free to stop me if I 11 brush over some things that you'd rather I not.
12 As a bit of an introduction, as we said, we're 13 using site information, which includes laboratory
\~ ' 14 experiments as well as analogous environments and results 15 from detailed process models to help us in our PA 16 abstractions.
17 And along those lines, I think I would try --
18 in terms of the question of what confidence do we have in 19 the code, I will, when I go through these things, try to 20 touch upon what kinds of things we have observed to 21 suggest this is the approach we want to use, and that is 22 an important aspect.
23 And I think the suggestion, are we going to l
24 have that documented in the users guide -- there really
(
,a
( ,) 25 isn't any intention of doing that, but I think the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005 3701 (202) 234-4433
98 1 suggestion is a good one, that I think for purposes of 7S 2 explaining the code at later times, I think we should be
)
3 able to, on a module-by-module basis, why did we go this 4 particular route. And I think it's relatively easy to do.
5 We have certain lines of evidence that we're 6 using, and I think that would be useful in explaining the l 7 code. I will try to touch upon some of that. Be aware 8 that it's done sort of impromptu here, and I call on 9 certainly the NRC and Center staff to help me if they feel l 10 I've incorrectly characterized something or have something l
11 to add to the discussion. I 12 It's also important to note that DOE has the I I
13 burden of making the licensing case. We would expect that
/ \
t i
\~/ 14 their modeling and proof is much more substantial than 15 what we will do in our TPA code. We're developing this 16 code to assist our review of the DOE application and also 17 do our independent check, but that is an important 18 difference.
19 And along those lines, we're hoping that what 20 we've included in our code are areas that are certainly 21 going to be important and areas that are uncertain, and 22 we're hoping our code -- and we would appreciate your 23 comments -- that we've gotten the right parts there to 24 allow us to check those things that, at least right now,
['
( ,)\ 25 we view as being important and/or uncertain.
! NEAL R. GROSS
- COURT REPORTERS AND TRANOORIBERS l 1323 RHODE ISLAND AVE., N.W.
! (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433
99 1 And that's the desire for this code, that
, 2 we'll be able to look at a few' focused aspects of the DOE
/si
/ 3 application. And that's the desire, and it's important 4 to -- and as Bob and others have noted, we have made it 5 flexible. Clearly this is an ongoing process. As more 6 information comes in, we have to be adaptable to changes, 7 and I'll get to some of those changes that I think we will 8 be looking at with respect to some of the information 9 we've heard.
10 And that really gets to the first three tics 11 there, if you will. One of the slides that's added for 12 this particular presentation is the KTI activities. I 13 think I want to show how the KTIs are working in this I
\_-) 14 particular environment of helping us develop the code, and 15 it also gets to what kind of work the KTIs are doing that 16 support the approaches that we have taken.
17 And also I have a final slide that gets to 18 anticipated sensitivities. The code is still in 19 development. We can't show you results, but we do have 20 guesses as to the way things are headed, and as a 21 preliminary note, we will share some of those thoughts l
22 with you and with the proviso that they clearly are i
23 preliminary. Some things we're still working on, and some l.
24 things could change, but to give you some sense of where (m) 25 this code is headed and maybe assist you in determining, NEAL R. GROSS
! COURT REPORTERS AND TRANSCRIBERS
{ 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433 l
100 1 are we being overly conservative; does that make sense.
7s 2 It might help that type of thinking.
- \
\' '/ 3 And with that, there's the five areas that 4 I'll touch on, and just for time, I'm going to move quick.
l 5 Going right to the unsaturated zone flow and transport, ,
l o what are the Key elements of our abstraction? Certainly l
7 the fracture versus matrix flow is an important aspect of 8 any unsaturated fractured rock.
9 The spatial distribution of the flow -- and by l
10 that, if you'll notice the one thing, I guess the one l 1
11 thing I should point out: A lot of these bullets are the j l
4 12 same bullets that you saw in the slide that Keith showed, 13 that overall diagram that showed the rule and the TPA
/D i
( -)
~ 14 code, and that's where these bullets come from.
15 Just because of the size of the boxes, there's 16 a limitation of what you could fit in there, so it is --
17 the spatial distribution of flow in that diagram, we are 18 looking at other things like the long-term climate change; 19 near-surface influence of infiltration, 20 evapotranspiration, runoff, et cetera; thermal reflux is 21 considered in there; all that getting -- what amounts of 22 water are getting to the repository horizon?
23 DR. HINZE: Tim, in terms of [ inaudible] in 24 that bullet, I note in I believe it's called the reference (fm)
,a 25 design of the repository, that we are now seeing shafts NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433 l
101 1 for the movement of air into and out of the repository.
s 2 Where are you bringing the concern with regard to seals of
( h 3 those shafts?
4 MR. McCARTIN: Well, for now, we do not 5 account for the -- a shaft for ventilation that is then 6 later sealed. We're assuming that the sealing of that 7 shaft can be competently done by the DOE, and in a 8 fracture media, the air movement in the unsaturated zone 9 really isn't going to be significantly affected by a shaft 10 that is sealed.
11 DR. HINZE: That's an assumption.
12 MR. McCARTIN: Yes. That's --
13 DR. HINZE: At one time, there was a great kls 14 deal of effort put into seals, and this has been lost to 15 us, and I'm wondering if that is being brought back into 16 consideration.
17 MR. McCARTIN: No. I thought -- and I 18 could -- my memory is vague on this, but I thought there l
I 19 was a -- and I'll say a research information letter on 20 shaft seals that said there really weren't any great 21 problems with that, and there might have been something 22 else done, but I guess I --
l 23 DR. HINZE: It's a minor point, but it's one 24 of these things that should be certainly documented, I r"~
(g) 26 would think, the assumption.
NEAL R. GROSS l
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
l
l l 102 i 1 MR. McCARTIN: Sure.
i 2 Retardation in fractures is certainly a big 7-
!. l l 3 part of the unsaturated zone, and we heard a lot about 4 matrix diffusion, sorption on fractures and the Calico 5 Hills unit in the unsaturated zone.
6 For our approach, fracture versus matrix 7 flow -- we're assuming transport is in the vertical 8 direction. We are not accounting for lateral diversion.
9 There are a couple of reasons for that. In Phase 2, in 10 terms of the flow portion of it, we did account for 11 lateral diversion in the paint brush unit. We did a 2-D i
12 simulation that gave us some variation of infiltration at 13 the different repository subareas, based upon the 1
fs
( )
\m/ 14 diversion in that unit.
15 Since that time, however, we've seen more and 16 more data and support, suggesting that there is limited 17 lateral diversion of the paint brush, and so we have not 18 included that particular part or above the repository in 19 our analyses. It's a simplification. You know, it's an 20 aspect that we can clearly go back to the more 21 sophisticated version. We have a code. We use DCM-3-D, a 22 dual continuum, three-dimensional flow model to go back to 23 that, but it was a complication that we no longer include.
l i 24 This is a simplification.
(x_/ ) 25 From the repository horizon downward, we are l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
103 1 also assuming vertical flow. We're certainly -- if you
,x 2 look at Bo's work, that's an extensive modeling exercise I i
' '/
3 that we really do not intend to duplicate. We will look 4 at it as those results suggest that possibly there's 5 perching of water and the lateral movement of water above 6 the Calico Hills as zeolitic. I think we can readily 7 include that in our particular approach, just as a --
8 The diagram that Bob showed, we do use a pipe 9 model, if you will, NEFTRAN, going down here. These legs 10 here are showing vertical and that's what we use.
11 However, if there's a particular flow path in Bo's model 12 that suggests that we have a lateral movement, then 13 eventual downward movement to the water table, we can x -) 14 easily include that. It'd be a relatively simple change.
15 So I think our -- while we're using vertical 16 flow only, we can accommodate a kink in the pipe, if you 17 will, and if it's more extensive movement in, say, the 18 Calico Hills vitric or zeolitic, it's easily accommodated.
19 But we are not doing the detailed modeling to support that 20 particular abstraction if we were to change that, but we 21 would look to the DOE results, and as appropriate, we 22 could incorporate that.
23 There are also was some discussion that we're l
l 24 also looking at what if all the flow goes down particular
/~N
, q,)
i 25 fault zones, and it's very little, if any, flow in other l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005 3701 (202) 234-4433
104 1 areas? We're looking at some isolated spots of the 2 repository to be very wet and virtually dry in other 7s 3 locations.
4 It's sort of a what-if type of question, but 5 the view of the repository where we have some variability, 6 but pretty much water going down uniformly everywhere is 7 on aspect of the flow system. What if it's far different 8 than that, and it's focused in a few isolated spots?
9 We're looking at that particular issue to see, does it i 10 make a big difference to repository performance. It's the 11 aspect of if a lot of packages get a little wet versus a 12 few packages getting very wet, and we are -- we've always 13 questioned that. We're going to see how that affects the
,s
/ i
%_ l 14 final result. j l
15 DR. HINZE: What kind of assumptions are you 16 making regarding setbacks from the fault? Are you --
17 MR. McCARTIN: For this one, we will take it 18 into consideration in terms of the number of packages that 19 get wet. That will -- one of the parameters we have in 20 the code is the number of packages that cet wet, so we 21 will look at a -- and we're using -- the SDS KTI is i
22 helping us in terms of looking at c?uster fracture zones, 1
23 fault zones, repository layout. And let's get something 24 that gives us some idea. Is this a big problem we have to (Gw 25 be worried about?
l NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4 433
1 105 l l 1 1 In the interest of time, I'll skip to the next
,_ 2 slide which -- thermal reflux is certainly an aspect of r
~
I 3 performance that's highly uncertain, and how will that l
1 4 affect repository performance, be it increasing the flow '
5 for a period of time. We're looking at -- there's some ,
6 uncertainty, what is the most appropriate model. j 7 We have two particular approaches to the l
1 8 reflux, and I just called one the rise / fall, that it rises I
9 sharply and then gradually tapers off, and why that 10 particular conceptual model? It was based on when heater 11 tests have been done in tuff, what do we generally see?
12 The heater turns on, and we get moisture coming instantly 13 back to the heater hole.
r3 I
\_-) 14 So we said, Well, why not try something where, 15 once again, we have the repository there; we have a large 16 influx of water very early on, coming back towards the i I
17 containers. That's really just trying to see or mimic i l
18 some of the experimental results that's seen in the field.
19 And then we have a bucket approach that is l
20 looking at the storage considerations, the thickness of l
21 the reflux zone, et cetera, and it's just a different way 22 of looking at this. There's a great deal of uncertainty )
23 as to what the most appropriate model is there. We are 24 doing some detailed, non-isothermal flow modeling to try
!, ,) 25 to get some insights with respect to what is the most l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
f
}
106 1 appropriate ones.
l,- 2 Likewise, DOE is doing the same thing, but as 1e 3 you heard in the technical exchange, to date they're using 4 homogeneous properties, et cetera, and they'll gradually 5 build up. We'll be watching to see how things may or may 6 not change. We think -- with these two particular l 7 approaches, we probably can mimic the results of a more 1
8 complex code relatively easy. It is an uncertain area.
I 9 Retardation in fractures: Certainly the j 10 matrix diffusion question was discussed quite a bit, and I i
11 guess I'd like to bring up to a certain extent that the ;
l 12 research program had a natural analogs project that they 1 13 discussed before this body, at least a couple of times, I f
\ b k/ 14 believe, and going back -- I think it was four years ago; ,
i 15 I remember English Pearcy was making a presentation and i
16 looking at the movement of uranium down major fracture i 17 zones with movement into microfractures that might mimic a I 18 matrix diffusion type of an effect, but clearly it's 19 different, because it's in microfractures. It isn't 20 diffusing into the matrix.
21 But when that work was done fairly -- one 22 might argue there was great foresight that four years 23 hence, we'd be looking at matrix diffusion at Yucca i
24 Mountain, but I think that that particular result -- we'll l
l (m,
, ! ,j 25 be looking at those results closer in terms of what is it NEAL P,. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
1
! 107 l
1 really telling us about matrix diffusion, because it w 2 doesn't appear to be there is much movement at all into i ( )
l 3 the matrix over thousands and thousands of years that thia 4 ore body has been there, as well as we'll be looking at 5 the information from the DOE on the seawell, et cetera.
6 And I don't know if English wants to add 7 anything to the Pena Blanca experiment, but it's --
8 clearly that's an area that the research program was 9 providing some information very useful. As you know, 10 budget constraints being what they were, that's one of the 11 projects that was -- has been eliminated.
12 But I don't know. English, do you have 13 anything to add on Pena Blanca?
ig\
\w/ 14 MR. PEARCY: English Pearcy, from the center.
15 I think you've expressed it pretty well. I'm delighted to 16 have the opportunity to bring some of our insights from 17 Pena Blanca more fully into the performance assessment.
18 MR. McCARTIN: And I think we'll have that 19 opportunity as we go down the matrix diffusion. How much 20 credit can we take for it?
21 Certainly the uncertainties that we're looking 22 at: Matrix conductivity certainly controls the transition 23 to fracture flow. The spatial distribution of flow, we 24 have the two different models for percolation, to l ,y
() 25 different models for reflux, and the retardation in HEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 108 1 fractures, we're taking a conservative approach. As you
. ;rm 2 heard, we are not including the effects of retardation in 4
- \ i 3 fractures or matrix diffusion.
4 Now, I will say: Our models can accommodate 5 that. The reason that right now we're not including it is 6 that to date, we haven't been convinced that it's 7 warranted to include that particular process in the 8 calculation, although the models can handle it.
9 And I think that's one area that, based upon 10 our discussions with the DOE in the last few days, I think 11 we want to explore further. What is our basis for this?
12 What is their basis for this? And we'll see how things 13 proceed. On the plus side, the model can accommodate it,
/_\
\ h V 14 so it's ,ust we're not including it in the calculation.
15 In terms of the KTI activities -- and what I'm 16 trying to show here, I made this pitch, if you will. at 17 the beginning of the technical exchange. This is not a PA 18 model. This is a model of the NRC. And in that regard, 19 Center and NRC staff, through all the KTIs, have 20 contributed to this particular 3.1 code.
21 And that's what I'm trying to show here is 22 that in terms of the parameter determinations, you can see 23 we have the unsaturated zone flow KTI, thermal effects, l
24 the near-field, and the SDS all contributing to various i i r^s s V 25 aspects, be it parameter determination, abstractions --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 R,40DE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
109 1 and you can see there's a lot of abstractions there.
l 2 These abstractions were primarily done by the p _s C t s t 3 KTIs. They weren't done by PA. We certainly interact 4 with them, and there's always that give and take of how 5 much complexity to have versus how much simplicity you 6 need to have a code that runs in a reasonable amount of 7 time.
8 And then in terms of detailed modeling, where 9 we try to get some confidence in our abstractions, 10 certainly the infiltration -- there's far more detailed 11 models that Stuart Stothoff at the Center's ben doing and 12 Peter Lichtner and Ron Green have been doing two-phased 13 flow of non-isothermal flow modeling that supports some of 7
(m_,) 14 the more contentious aspects, the reflux and the 15 infiltration rates.
16 MEMBER POMEROY: Tim, can I just ask a quick 17 question?
18 MR. McCARTIN: Sure.
19 MEMBER POMEROY: Going back to my previous 1
20 question with regard to how TPA views uncertainty in the 1
21 abstraction process, you just made a statement that 22 interested me. Is it then the KTI team that decides 23 w'lether the uncertainty can be narrowed, if you will, in 24 the sense they were doing it, or is it the VA team portion
(,-,)
\-
25 of the operation or who is it, I guess, that makes that l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. (
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 I
i
110 1 decision?
2 MR. McCARTIN: It's really a joint decision.
" ) 3 The KTI teams are at the forefront of the parameters in 4 abstractions. We probably play as much the devil's 5 advocate in a lot of the things that look at what's being 6 done and question things. We're very inquisitive in the 7 PA group, so we tend to question things. And it's a 8 healthy exchange.
9 But I would say generally it's the KTI leads 10 and their groups that are taking the lead in terms of what 11 is the right parameter range for this? What is a good 12 abstraction for this particular model? And there's give 13 and take between both sides.
V 14 MEMBER POMEROY: Thank you.
15 DR. HINZE: If I may, Tim --
16 MR. McCARTIN: Sure.
17 DR. HINZE: In terms of the thermal effects, 18 we are not going to see the kind of data that many of us 19 are interested in before the PA decision will be reached.
20 What kind of problems is the lack of the thermal -- basic 21 thermal data in the uncertainties area -- have we 22 reasonably large uncertainties as a result of that, or how 23 much is this hurting us?
24 MR. McCARTIN: Well, to date, the range in p) i 25 terms of what thermal reflux will do in terms of an u.J NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
111 1 increase in infiltration is rather extreme. I think you 2 could look at certain calculations done to date -- and I 7-( )
3 emphasize calculations, because generally these models are 4 done as homogeneous with very little heterogeneities in 5 them, and that could drive the whole system, but the 6 homogeneous models to date, some of them predict -- could 7 predict as much as 100 millimeters per year of 8 infiltration due to just refluxing. And maybe the numbers 9 could go up higher. That's a nontrivial amount of water.
10 The question is: What length of time? At 11 what time does it start? And how much?
12 For our particular approach, the reason we 13 have those two models -- and this is a nice trade-off (mx,) 14 between the KTI and the PA code. In the KTI, they're 15 questioning how much detail, how far should we go with 16 this particular detailed modeling, and in PA, we're 17 saying, Well, is there a way we can help you prioritize I l 18 this particular problem. So we have these models, these l l 19 two different abstractions. We can look at the l
l 20 sensitivity of these particular models to the final 21 result.
l
! 22 Does it make a big difference between the two 23 models? Does it -- possibly some of this -- maybe the 24 reflux occurs at an early time when the container is
( ) 25 intact, so it's not at issue. So it's a way of assisting NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l 112 1 the whole process. How much more work do they want to do?
l
,3, 2 Keith?
> \
, ; /
3 MR. McCONNELL: Keith McConnell. I'd just 4 ask, Dr. Hinze -- I think you're picking up on an 5 important point, that to narrow the range of these 6 parameter values that are going to come out of the thermal 7 refluxing, it would sure be nice to have data on that 8 process. But in the absence of that, I think we're going 9 to have to go forward with a range of effects in a lot of 10 areas, including thermal hydraulics.
11 MR. McCARTIN: Yes. There really isn't much 12 to -- I mean, you're looking at a highly complex process 13 in a very heterogeneous system, at a very large scale, and 7-~
iss 14 to date, the models are homogeneous and --
15 DR. HINZE: Over a very long period of time.
16 MR. McCARTIN: Yes. It's -- hopefully we 17 can -- the attempt is to hopefully bound the effects and 18 see, does it make a big difference to performance. If it j l
19 doesn't, then we're not too worried. If it makes a big 20 difference, then we have to go back and sharpen our pencil 21 and see, Gee, is this bound the right bound.
22 DR. HINZE: That's when even I will agree that 23 PA has a use.
24 MR. McCARTIN: Let the record show --
<"'N (j 25 VICE CHAIRMAN HORNBERGER: How likely is it NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005 3701 (202) 234-4433 l
l
113 1 that we'll have any data at all, let alone on a large f3 2 scale; how likely is it that we will have data from your
- i
~
3 tests?
4 MR. McCARTIN: Well, we --
5 VICE CHAIRMAN HORNBERGER: And I don't mean 6 for VA. I mean for LA.
7 MR. McCARTIN: Well, we've data from Peter's 8 tests. The question is that there'll always be is now 9 expanding that information to a mountain scale problem 10 with a much larger source, and I think there'll alwayo be 11 some uncertainty there that we have to live with.
12 You know, we're very interested to see what 13 impact these particular models have and very interested in
/ _N i
\ ') 14 the DOE thermal modeling as, once again, there's an area 15 where we can do some limited work in this area. They are 16 doing far more, and it'll be interesting to see as they 17 add heterogeneities, et cetera, do these numbers increase, 18 decrease. What do they think bounds the problem?
l 19 Other than that, I'm not sure. l 20 Margaret?
21 MS. FEDERLINE: Tim, Margaret Federline. I 22 think it's worth noting that we did send a letter to DOE 23 on their thermal hydrology program and made some 24 suggestions. I think one -- perhaps Mike or Keith could !
(~N,
(_,) 25 touch on it -- was a longer period of heating and a longer NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
114 1 period of cooling. And it's my understanding that they
,\ 2 did pick up on that and are planning to follow through on
)
'~'
3 that.
4 So I think we feel more comfortable over the 5 long run that we're going to get the confirmatory data 6 that we need. It's a case of making sure that the 7 sensitivity analysis for licensing gives that basis for 8 reasonable assurance that we could then look to the 9 confirmatory data to confirm.
10 MR. McCONNELL: And I would just add -- Keith 11 McConnell again -- that I think that based on our 12 technical exchange, that we've identified that as being 13 one of the key performance drivers that we would be 7m
'\_/) 14 expecting to work with DOE on developing not only the data 15 that's required -- and I think Norm will talk about this 16 this afternoon -- but also the conceptualization of the l 17 process over the next year.
18 MS. GILL: This is April Gill with DOE, if I 19 could just add something. I wanted to back up what 20 Margaret said. We did receive that correspondence. We've l
21 had good interchanges with the NRC and Center staff. We 22 did change our former hydrology testing program based upon 23 concerns that the NRC and Center staff had, and we 24 responded to your letter about three weeks ago, I believe.
/~
(s_- 25 I talked to Ron Green, and he said he was in receipt of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(20?) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
115 1 it. And we also have an exchange planned this fall to 2 discuss this in more detail.
I.,s 1
\'~) 3 DR. HINZE: April, is that the single leader 4 test or is that the drift test or both?
5 MS. GILL: It's both.
6 MR. McCARTIN: In terms of anticipated 7 sensitivities, if we look at matrix conductivities, right 8 now if we were to do the analysis today, the range we have 9 for infiltration is between 1 and 25 millimeters per year.
10 The fact that infiltration rates have risen quite a bit 11 since our last PA exercise -- one of the things we did, if 12 you remember, in Phase 2, we tried to account for the 13 matrix and the fracture systems both.
p
\m / 14 At that particular time, the infiltration 15 rates were lower, and so if you look at, say, the Topaz 16 Springs, 2 millimeters per year, we were on the cusp of a 17 little bit of -- maybe a lot of matrix, a little bit of 18 fracture, et cetera. For these ranges, we've abandoned 19 the desire to simulate that, those two systems, and the 20 reason is that if you look at the Topaz Springs matrix 21 conductivity of 2 millimeters per year, we think, well, 22 once you get into these high ranges, we'll just consider 23 fracture flow; we won't worry about the matrix.
24 Likewise, the Calico Hills vitric, the matrix
/^\
( ) 25 conductivity is 5,000 millimeters per year. That's always NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
116 1 going to be matrix flow, very low flow. Calico Hills 7s 2 zeolitic, 3 millimeters per year; the same kind of 5'"') 3 problem. Trying to simulate both continuum really didn't 4 do a lot in terms of the IPA Phase 2 work when you're in 5 the high infiltration rates, and now that the infiltration 6 rate has climbed, it makes a big difference, and you can 7 see what's going to happen there. And that's why --
8 And we brought up the point of retardation in 9 the Calico Hills vitric, and that is an important value.
10 I know in talking to Bo, he doesn't think it can rely on 11 much retardation in the matrix in the vitric. Now, I know 12 TSPA-95 -- I checked the numbers last night. TSPA-95 had 13 some rather substantial retardations for some nuclides,
,f s
\_/)
14 certainly not neptunium and technetium, but a lot of the 15 other radionuclides are highly -- are certainly retarded 16 in the vitric.
17 Now, maybe those numbers have changed. We'll 18 have to look at that, but you can see one of the aspects 19 of performance, where you have the Calico Hills vitric, 20 with some retardation, you can be in transport velocities 21 on the order of, you know, less than a millimeter per 22 year, and, in fact, for what it's worth, the problem we 23 were having with NEFTRAN, which was a nasty little bug to 24 find, it was a situation where it was four-member chain.
rx
!v) 25 It was retarded, and there were various retardations for NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
117 1 the different members of the chain.
,, - ~ 2 Some of those members were moving at the speed b 3 of about a micron a year, so although it was a problem l
4 with the code that we subsequently -- we believe we've 5 fixed, it was a -- one of the things that made it not 6 overly a big problem, if you will, is the fact that it was 7 such low flow that we really never expected to see any 8 significant doses, but it was the type of thing you could 9 see where some of the flow codes -- you got to a 10 particular part of the code where things just weren't 11 moving, and it was a part of the code that rarely gets to, 12 and that's where the error was. But that's where -- that 13 vitric unit is important, m
(\ ') 14 Pluvial infiltration: Right now in terms of 15 what will happen in a glacial maximum, we're looking to a 16 two to three-time increase in rainfall, up to ten degrees 17 cooler in temperature. When that happens, our 18 infiltration estimates can go as high as 30 percent of the 19 rainfall, which will be on the order of 150 millimeters 20 per year.
21 We're certainly -- that's if we did the j l
22 calculation today. We're certainly looking into analogous 23 environments to a cooler weather Yucca Mountain, to 24 understand, is that a reasonable value. It's a way of
(/~%)
ss 25 trying to get some confidence. Do we believe this number? I NEAL R. GROSS COUR1 REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l 118 1 Certainly the code produces it, but we can go to weather
- 7-)
\
2 environments, cooler environments than Yucca Mountain and Ns/
3 see, Is this to be expected.
4 And once again, the fact that we're assuming 5 retardation fractures and matrix diffusion is minimal is 6 certainly going to make for a difference between our 7 results and the DOE's. However, the book is not closed on 8 that particular aspect. I think we need to go back and 9 revisit the justification for that.
10 VICE CHAIRMAN HORNBERGER: Are you at least 11 going to do some sensitivity work on retardation and 12 matrix diffusion?
13 MR. McCARTIN: We certainly can. I now when a
t
\!' 14 we did the calculations for the NAS, we did look at a 15 retardation factor in the fractures. I mean, it's easily 16 done. It actually makes the code run faster; the slower 17 the transfer time, it actually speeds up the code, so it 18 isn't necessarily a bad effect.
19 But, you know, it's -- I think we have to look 20 at that and look at matrix diffusion. How much matrix 21 diffusion do you need to make a significant impact on 22 performance?
l 23 Looking at the clock, I realize I'm going very 24 slow if you want to finish by 12:00. Do you want me to n
() 25 talk very fast or skip some slides?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
I
119 1 CHAIRMAN GARRICK: Actually, we don't have to
,_ 2 finish by 12:00 if I read this agenda correctly.
t i
!' '/ 3 MR. McCONNELL: We have an hour after lunch.
4 CHAIRMAN GARRICK: We have to break at 12:00, 5 so you can change it to a break point around 12:00 that's 6 convenient to you.
7 MR. McCARTIN: Okay.
8 CHAIRMAN GARRICK: We can continue this 9 afterward.
10 MR. McCARTIN: Oh. It's been a long week.
11 The engineered barrier system -- and as you 12 saw yesterday, there is a lot of stuff in the engineered i
13 barrier system. There's a lot of processes going on.
,em
/ \
- (/ 14 There's a lot of important aspects to performance in that I
15 particular part of TPA code, many assumptions, many l
16 aspects of performance that are certainly going to be '
l 17 important: waste package corrosion, mechanical disruption 18 of the package.
19 The quantity and chemistry of the water l
l 20 contacting the waste form and the waste container, there's 21 no way to not have that important to performance. It's a i
22 very difficult aspect of the performance. To get dripping 23 into drifts is certainly one aspect of it.
24 And another part is given the drips onto a (N waste container, do you assume that all the water that i
s ) 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344 433
120 1 drips onto that container ends up inside the container on
,_ 2 fuel? That's also a part of the abstraction, very
- i i 3 difficult to try to determine what's the -- a defendable 4 number. It's easily determined what the most conservative 5 number is, but backing off that is very difficult.
6 Radionuclide release rates, and there we have 7 the dissolution rates and solubility limits. It varies 8 from nuclide to nuclide, what's more important. Something 9 like a neptunium and technetium that are very soluble, 10 very mobile, the solubility limit isn't that big a factor.
11 It's the dissolution rate that drives the performance.
12 For other nuclides where the solubility limit is very, 13 very low, well, then a high dissolution rate really
,m
\,)
s 14 doesn't matter, because the solubility limit restricts how 1 1
15 much can get out. !
16 In terms of our conceptual model approach for 17 the waste package corrosion, we're taking into account the 18 temperature, humidity, and the water chemistry at the 19 surface of the waste package, to determine the corrosion 20 of the waste package. And certainly we have -- I guess 21 one of the biggies is the galvanic protection, and there's 22 an efficiency factor that we have to -- that sort of l 23 varies the uncertainty and how much degree of belief we 24 have in the galvanic protection.
A)
(V 25 Mechanical disruption of the waste package is NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
121 I
1 considered. I i
1
,s 2 The quantity and chemistry of water contacting I i ! ,
\'~)
3 the waste package and waste form: There's two aspects to i
1 4 that. In terms of the quantity, once again, as I was 5 alluding to on the first slide, the engineering has a 6 major impact on water that's going to move into the drift, ,
l 7 and then, of course, the container, the characteristics of I l
l 8 that pit, how does water move through that pit into the j 9 container. j i
10 And there clearly is a lot of uncertainty l l
11 there. We have parameters in the model to sort of account i 12 for uncertainty. Now, these parameters are not physically 13 based, but they're accounting for -- for example, water ,
tN l
\_l 14 percolation going into the drift, some will be diverted by 15 capillary forces around the drift; some droplets may come l
16 in, move down the side wall of the drift and not contact 1
17 the container; some of the droplets will obviously drop 18 down and hit the container.
19 And there's some accounting for that fraction 20 of water that would actually hit the container. That's a 21 parameter that we have to look at how important is the 22 diversion of 10 percent, 20 percent, 40 percent of the j 23 water, or whatever. Like I said, there's virtually no 24 data on that; it's more anecdotal evidence, observations
,~.
( ) 25 in mines, et cetera. But it's something we'd like to have
! NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C 20005-3701 (202) 234-4433
122 1 a better idea of how important is it to performance.
s 2 Likewise, for the water that drips and l \
3 contacts the container, there is a parameter that tries to 4 account for not all that water will move intc the 5 container, and once again, we're looking at the -- an 6 uncertainty, capturing some of the uncertainty of what's 7 going on there.
8 Clearly we can set them both to zero or one, 9 depending on which way it works, to give the most 10 conservative value, that all the water into the drift 11 contacts the container and goes into the waste -- and 12 contacts the waste form. That's -- we know what that end 13 is. The question is: As you move away from that, how t /
\/ 14 does the performance react? j l
15 The radionuclide release rates -- oh, let l
16 me -- and then the chemistry, we have the chloride 17 concentration and pH of the groundwater. We have single I l
18 values we're inputting there. There is detailed modeling 1
19 going on in the near-field KTI to look at the evolution of 20 the groundwater chemistry, to try to give us a better 21 sense of, can we use a single value; should we use a look-22 up approach where it accounts for the variation of time, l
23 possibly due to the thermal pulse that's going to happen, 24 et cetera.
Or
( ,) 25 We're looking into that. We can account for
! NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
l (202) 234-4433 W ASHINGTON, D C. 20005-3701 (202) 2344433 l
t
123 1 it. There is a fair amount of uncertainty as to what it
,cw s 2 should be.
,)
3 Radionuclide release rates: We have congruent 4 dissolution of the spent fuel. One of the -- getting back 5 to Pena Blanca, the natural analog rele"7e model is 6 related to Pena Blanca, and we're looking at what kinds of 7 mobilization do we see of uranium in that natural ore 8 body. We'll specify that and see how that -- what does 9 that do to performance as one aspect that in the very, 10 very long term may be the release rate from the spent fuel l
11 will in part simulate the releases we've seen at Pena ,
12 Blanca.
_ 13 And then, of course, the -- we have the i $ ;
/ 14 bathtub and drip model for release model, and I have a 15 slight cartoon of that that gets into the -- one of the 16 important aspects of the dissolution of the spent fuel is 17 the surface area that's wetted, and one model we have is 18 obviously called the bathtub model. You get a hole in the 19 top, water's dripping in. It fills up to some location 20 where it then goes out.
21 The amount of fuel in surface area that's 22 wetted is based on the height of this pond, if you --
l 23 well, the height of water that you get to, to that 24 particular hole. This is a sample parameter. The height n
(.,,) 25 of that, that's our bathtub model.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
124 1 However, DOE, as you heard, has a slightly
,s 2 different approach. It's more of a dripping model, where 1
i
' /
3 there's a hole near the bottom. The drips come on, 4 contacts some waste, and continues on through. We can --
5 we're going to do both of those to what -- how different 6 are the releases between those two types of models. You 7 can certainly see that the amount of surface area that's 8 wetted is drastically different. This is most likely a 9 more conservative approach. How much more, how much 10 difference does it make, we'll try to look at that with 11 the two different modela.
12 VICE CHAIRMAN HORNBERGER: How long would it 13 take to fill up?
A I
\m.,') 14 MR. McCARTIN: Well, it's related to how much 15 infiltration and -- well, the diversion and et cetera.
16 Dick probably has a better idea. It's on the order of --
17 I thought we were getting on hundreds of years, maybe 18 longer.
19 MR. CODELL: I'm sorry. I didn't hear all of 20 the --
21 MR. McCARTIN: And be aware that the height of 22 this outlet pit is variable, but how -- on the average, 23 how long did it take for the bathtub model to fill up?
24 MR. CODELL: It was hundreds of years easily, x
(x/ ) 25 but that depends where the hole is. I'd like to make one NEAL R. GROSS
! COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C 20005-3701 (202) 234-4433 l ,
125 l
[ 1 comment, further comment --
- 7. 2 MR. McCARTIN: Sure.
i \
\',/
3 MR. CODELL: -- on the source-term model. The 4 DOE model, as you -- from the data I've seen to date, from 5 argon, would use a much higher release rate for dripping 6 than we would use for the immersion. And you might 7 picture this as here's a drop impacting a piece of waste, 8 and it's very active with the waste, whereas the immersion 9 model, the bathtub model, is pretty much stagnant.
10 So one of the considerations is we might want 11 to -- if we're using the drip-on approach, we would be
)
12 using a higher release rate from each square centimeter of 13 exposed fuel surface.
(
(_)s 14 DR. STEINDLER: Are you taking into account 15 any redox reductions or status in your model?
16 MR. McCARTIN: I'll have to pass that one off 17 to -- did you hear the question, either Dick or Sitakanta? I 18 MR. CODELL: Not exactly, not explicitly. The i
19 model itself is very simple, and it's just a number of i 20 rates which are highly abstracted from the fuel data. Of 21 course, the -- for example, the rate of dissolution of the 22 fuel is certainly controlled by redox considerations, but l 23 we're not handling that exclusively.
24 DR. STEINDLER: So there really is a potential
! ;f~~hv) 25 for being fairly far off from the real world because of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE.. N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
126 1 your abstraction, coupled to perhaps a lack of fixed 2 design plus basic information. So when you look at the 7-( ')
3 sensitivity of that source term, so to speak, if it comes 4 out to be highly sensitive, which is what one might 5 assume, that could represent a significant weakness of 6 your overall result, dose at the other end.
7 MR. McCARTIN: Yes. Absolutely. Yes. The --
8 that's what I said. There are many aspects to this EBS 9 situation that we're trying to do as much as we can to 10 understand how reasonable our release rates are, and it is 11 a very difficult part of the model, if just for the fact 12 of what is the chemistry of these droplets inside the 13 container. We are not accounting for how all the n
( ) l x_ / 14 chemistry might change inside that container. There are j l
15 many aspects that lead to a number of uncertainties.
16 In terms of uncertainties, once again, there's i 17 many. The corrosion, we certainly have the effectiveness 18 of the galvanic protection and critical relative humidity, l
1 19 which we heard yesterday at the technical exchange.
, 20 Quantity and chemistry of water contacting 1
! 21 waste forms, getting to the release rates: All -- there 22 is a lot of physics going on there and chemistry, and what 23 exactly is the right number is highly uncertain. I guess 24 only to say we're looking to experiments that DOE's doing n
() 25 on spent fuel, extrapolating from other things like Pena NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 127 1 Blanca, to try to get a better idea of what seems to be a f~s 2 reasonable bound for this particular part of the code, as i i
\") 3 it does have a very strong influence.
4 In terms of this particular KTI or in terms of 5 the KTI activities, you'll notice that I think we might 6 have every one there. I could be wrong, but I -- just 7 about certainly, and there's just a lot of activity, 8 trying to understand and best represent something to allow 9 us to understand everything that's going on.
10 And obviously there's the parameter 11 determination. There are the abstractions which you 12 certainly see a lot of them. I'll point out -- I got 13 tired of putting all the ones down, so for dripping, I n
! \
\_ / 14 have "every one." And I think every one -- it's just hard 15 to get away from the fact that trying to determine how 16 water gets into the drip, onto the container and onto the 17 waste form is -- everybody's interested in that one. And l 18 it truly is a -- it's a difficult problem.
19 And certainly, I guess, along those lines, 20 we're hoping some of the niche experiments that Bo talked l 21 to hopefully will help us in that regard with respect to l
l 22 certainly the water getting into the drift. There's still 23 the aspect of how water moves through the -- onto the 24 container and then into the container, and we're looking
(- 25 at that and CLST people are doing what they can to help us (a)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
128 l
1 out there.
f-3 2 Detailed modeling, we're certainly looking at Nj 3 the -- a multi-component, non-isothermal flow simulations 4 to try to look at that evolution of the water chemistry.
5 And that primarily's related to in the rocks, so it would 6 be the water getting onto the waste package. We are 7 hopefully going to look further and look at once it gets 8 into the container, what's the chemistry of the water 9 within the -- inside the container that impacts the 10 dissolution of the spent fuel.
11 In terms of anticipated sensitivities, in 1
12 terms of galvanic protection, in terms of a rough rule of l l
,_s 13 thumb, with no galvanic protection, we think we'll see ;
/ T i k 14 package lifetimes on the order of 2,500 years. With l l
l 15 galvanic protection, that waste package lifetime will move 16 out to the 10,000 year.
17 Now, these are all -- there's -- certainly the 18 water chemistry chloride concentrations, corrosion rates, 19 all affect some of those numbers, but, in general, if 20 you're looking, what does it buy you, that's the effect 21 that we think we will see.
22 Mechanical disruption of the container 23 certainly may be of minimal importance if we have a l
t 24 situation where we have limited galvanic protection and by q
( ) ,
25 2,500 years, most of the containers are failed by 2,000 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
129 1 years or so.
,~s 2 Release of spent fuel, this is getting at the i )
3 release rate problem. If we assume a release rate of 10 4 milligrams per square meter per day, that would result in 5 the dissolution of all the fuel in 27,000 years, assuming 6 a particle size of 1 millimeter. You make that particle 7 size smaller, the surface area goes up, and you would get 8 even a shorter release. Of course, that's an important 9 parameter for the release rate is that particle size, and 10 it is sensitive to that particular number.
11 DR. STEINDLER: Is 1 millimeter what you're 12 using, or is that just an example?
13 MR. McCARTIN: That's around what we're using, A
/ )
k/ 14 I guess. It's in the right ballpark. It'll be a sample 15 parameter, I believe, and I'll -- Sitakanta's nodding, or 16 do you want to add something?
17 MR. MOHANTY: Yes. My name is Sitakanta 18 Mohanty. I'm from the Center.
19 I would like to go back to the previous 20 question on the release model. We do take into account 21 chemistry to some extent, because at least one of our 22 models takes into account pH of carbonate concentration.
l 23 It has strong dependency on temperature. This is an l 24 empirical equation, so that is how we bring in chemistry 25 to some extent.
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
l 130 1 I'm going back to the particle size, that's l l
,_s 2 the sample parameter at this point. We maintain mean as 1 !
/ i
\
' '/
~
3 millimeter.
4 MR. McCARTIN: But the chemistry in the -- the 5 impact of the container contents, the degradation of that 1
6 on the chemistry of the water in the container we're not j 7 accounting for. l I
8 MR. MOHANTY: That is correct.
9 DR. STEINDLER: My point is that if you've got 10 a bathtub full of water sitting in an iron bathtub, you 1
11 need to take into consideration the redox potential, l l
12 because you're not going to get the kind of oxidated 13 medium that you think you're going to get elsewhere. And
,\ l
'w.)
14 that ought to influence strongly the rate of dissolution 15 or transport -- subsequent transport. That's all I'm -- )
l 16 MR. McCARTIN: Yes. And I think -- i l
l 17 CHAIRMAN GARRICK: Is this a good break for 18 lunch?
19 MR. McCARTIN: Yes. I'm at the end of the 20 EBS, and --
21 CHAIRMAN GARRICK: Yes. Why don't think about 22 it. The committee is very interested in the engineered 23 barrier system, and it seems to have grown in importance l
l 24 in the last few months. There's more questions about it
?s
, (v ) 25 that we'll surely search out tomorrow in the context of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE.. N.W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
131 1 the models discussed.
s 2 But maybe at lunch, you can think a little bit
( '
)
3 about the question of, is this an area given that it's a 4 higher profile issue now than it might have been in the 5 past, that NRC could have more resources to deal with it, l
6 because it's so fundamental. I i
7 And if Yucca Mountain is going to increase l
8 dependence on engineered systems containment, the whole i
i 9 containment strategy, the whole containment philosophy, l I
10 the whole issue of source terms, is sort of wrapped up in 11 the engineered barrier systems. How does the staff feel 12 about their capability to deal with this new, high-profile 13 issue? It's something maybe we could talk about some more
(-
t 1
\_' 14 after lunch.
15 MR. McCARTIN: I think we have some capability 16 in the code to look at different aspects of it. But, you 17 know, we're still -- defense-in-depth requires that we 18 want to see some performance of the geologic system, is a 19 primary reason we're digging a big hole.
20 CHAIRMAN GARRICK: Yes. And we're just for 21 the moment asking questions, and we asked the question of 22 the containment capability. If we're going to get to an l 23 engineered system, there's no question that the defense-24 in-depth philosophy has to be invoked here, but the idea
',3, (s_/) 25 here is how much containment are we going to get from each
\
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
i l 132 l
1 of the barriers.
l l .-s,
, 2 So this is going to require perhaps more l l
(~')0
! 3 extensive analysis than we were anticipating, I guess, is 4 my thought, and are we prepared to do that is part of the 5 question.
6 If this is an appropriate time, I think we'll 7 break for lunch and come back, let's say, 1:10. Would 8 that be fair?
9 Whereupon, at 12:10 p.m., the meeting in the i 10 above-entitled matter was recessed, to reconvene at 1:10 I
11 p.m., this same day, Wednesday, July 23, 1997.) ;
12 j i
13 I
,73 i )
is' 14 15 16 17 18 19 20 21 22 l
23 24
<~~.
w/
! 25 l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 1
133 1 A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N
,3 2 (1:10 p.m.)
\ 1 x._ j 3 CHAIRMAN GARRICK: Can we come to order and 4 continue.
5 MR. McCARTIN: Okay. Just a, I guess, couple 6 of thoughts. You were talking about the engineered system 7 and accounting for it, I guess, in our PA, and I'll give a 8 couple of quick thoughts that there is a lot of aspects to 9 the engineered system that has a lot of uncertainty, a lot 10 of information yet to collect in terms of what 11 contributions to performance it will create.
12 We believe in the code we have flexibility to
,, 13 try to determine how important some of these are, be it (x-) 14 the -- a factor for water moving into the drift, a factor 15 for water moving into -- through pits into the container.
16 We also have a parameter that's the number of 17 containers that get wet, depending on what we view in 18 terms of the engineering; also possibly Bovardson's flow 19 model. Maybe we see that, you know, given the 20 heterogeneities, we think maybe 10 percent of the 21 containers will actually get wet, so we have a parameter 22 to account for that.
23 That's not to say we can -- on the down side, 24 that's not to say that we have strong support for picking
/%
(_) 25 particular values, but I think as more information comes
! NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
134 1 in, as we learn more, we can account for aspects of the
,_s 2 engineering with these parameters.
/ \
\ l 3 Part of what we're doing in the sensitivity 4 analysis is trying to get a sense of, well, how important 5 are these parameters, and it'll be a way to look at some i
l 6 of the areas that we want to delve a little deeper into.
7 And other than that, I'm not sure there's a lot more we 8 can do at this time. l 9 There's very little information in terms of 10 the -- how the container's going to fail; how many pits l l
1 I
11 are going to grow with time; water moving through pits i
12 into the container; are the pits at the bottom or at the 13 side. There's just -- there is a lot of uncertainty, and
,m I )
\m / 14 the only thing I would like to add, in terms of -- and I'm l
15 not sure it was clear, but in terms of the chemistry i
36 within the container, within the near-field KTI, there is l 17 an acknowledgment that we are using the water chemistry of 18 the host rock, be it modified by thermal effects, and 19 assuming that's what it is in the container, they 20 certainly are aware of there's a lot of potential for 21 chemical changes of that water chemistry in the container, 22 and they are going to look into it.
23 But for now, for the code, we clearly don't 24 have a module that's accounting for that. That's O)
( ,
25 something that we envision that would be done outside the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
i (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
i
l 135 I i
1 code, and hopefully they can give us a sense of what are
,_s 2 some of the parameters for the release rates that we're
\
\' "/ 3 using.
l 1
4 Are they in the right range, or given this l 5 kind of change in the water chemistry, we would change --
1 6 it would be a lower release rate and use different I i
7 parameters, and that's -- it's an analysis outside of the l 8 TPA code, and that's where part of the abstraction work 9 gets into it.
10 But I don't think we can include detailed 11 evolution of water chemistry within the TPA code, but we 12 certainly acknowledge that some of that work needs to be 13 done to get us a better sense of whether these release m
j 14 rates are correct or not.
15 Some of the laboratory tests that we use to 16 look at release rates, some of them use potentially J-13 17 water and other types water. Well, how appropriate is 18 that to the inside of the container? Those are things l l
19 that need to be looked at. l 20 Now, I don't know if that gets to all of your 1
21 question, but it is difficult for us as a regulatory l 1
22 agency to ascribe credit to engineering where there's 23 little basis.
24 CHAIRMAN GARRICK: Yes. Some of us have a
() 25 little problem with that, and some of us believe that it's l NEAL R. GROSS
, COURT REPORTERS AND TRANSCRIBERS
! 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
136 1 easier to specify and verify an engineering design than it 2 is to characterize an actual setting. So in principle, 7\
~
3 one could be led to believe, if they believed that, that 4 we ought to be able to be more quantitative with respect 5 to the engineered systems than we are with respect to the 6 natural systems.
7 And so if there's one thing we ought to be 8 able to do, it would seem is to quantify the role of the 9 engineered systems with respect to the overall containment i 10 capability and the capacity of the repository.
11 And given that that's now something it looks 12 like we're going to have to depend on maybe more than we l 13 thought in the first place, that's the whole reason for
,m a
\
ks/ 14 raising the question and suggesting that maybe we need --
15 this committee needs to be convinced that we have the 16 capability to dc
- hat. And that's the reason for it.
17 It's just part of the package, part of the l
18 defense of that package.
19 MR. McCARTIN: Sure. And along those lines, l 20 what Dr. Steindler was suggesting about the environment 21 inside the container, we may have a better handle of what 22 that water is like inside the container than we do 23 potentially in the evolution in the rock. And so we l
l 24 should be able to look at the water chemistry in the
/~N
( ) 25 container and get a sense of how that should affect NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
137 1 dissolution rates.
,-, 2 And that -- hopefully we'll know what spent-
'~
3 fuel rods look like, et cetera, and things like that.
4 CHAIRMAN GARRICK: In the last couple of days, 5 we saw a discussion that involved a great number of 6 options for the design of the engineered systems, fillers 7 and backfills and drift protection. And somewhere along 8 the way, somebody's going to have to make a decision about 9 which of those options to actually implement.
10 Now, we're pretty close to that, as far as the 11 viability and assessment is concerned, but the license l
12 application is another entirely different matter or could 13 be for sure.
n
! )
\j - 14 MR. McCARTIN: And there's -- part of it j 15 depends on your approach in the performance assessment 16 code, and I know just going to our BTP on low-level waste 17 in terms of engineered barriers, there's always that 18 question: How long will it last? If we have container 19 lifetime on the order of a few thousands of years, at a 20 minimum, then a drip shield, if it is to exist -- you 21 know, we're looking at, what will it be like in 1,000, 22 2,000 years, not so much today; it really doesn't make 23 much difference. We have a container that'll be intact.
l 24 But in the thousands of years' time ff&me, I /*
(~- ) 25 what will it look like? And that's part of the dilemma NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
l l 138 1 with some of the engineering, projecting it much further, i
7y 2 and I guess -- well, with the Phase 2 code, we made an b 3 assumption that all the water that came into the drift got 4 into the container and it contacted the waste form.
5 In that situation, potentially a drip shield 6 has more relevance in terms of its impact on performance, 1
7 although be aware there aren't a lot of drifts at Yucca i
8 Mountain. Although we see the infiltration rates going l 9 higher, 10 millimeters per year is not a lot of water. It 10 isn't like you need your umbrella and a raincoat in there.
11 And so the drip shield is not shedding a 12 massive amount of water. It's shedding very, very little, 13 trace amounts, of water. And so there's the aspect of --
s
\ >'
14 currently in our approach, we now are looking at the 15 ability of the drift to shed some moisture, the ability 16 of -- there will be drifts, but when they contact the ,
17 container, they might shed it there, and so we're trying 18 to account for certain things, but it's not --
19 I know when I see the drip shield, it's not 20 like there's a flood of water coming down, being shed off 21 there. It still is a very low flow envirorment.
22 DR. STEINDLER: That's the normal situation, 23 but there are potentially abnormal situations in a 24 repository which might lead to a flow-type. If you're (3
(,) 25 talking about high conductivity, high transmissivity in NEAL R. G500SS !
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
l (202) 234-4433 WASHING'ON, D.C. 20005-3701 (202) 234-4433 l
t l
139 i
l 1 the fractures and faults, and you don't have the proper
,_ 2 setback, you don't know what these are, it could develop I) 3 into something more than just a drip, drip, drip.
l 4 MR. McCARTIN: There's certainly the potential 1
5 for some areas of the repository to be wetter than others, l
6 but we don't need a repository that's bone dry. But I 7 understand what you mean. Yes.
8 DR. STEINDLER: Could I ask --
9 CHAIRMAN GARRICK: It's sort of like, Keith, 10 the discussion, at least today, on the capability issue.
11 We might just toss it a little bit.
12 DR. STEINDLER: Well, I wondered -- certainly 13 it's true that there are a lot of oncertainties in tryiny
(_) 14 to decipher the geosphere.
15 The conventional wisdom has been that the EBS 16 would be a lot easier for us to get a handle on, to have 17 less uncertainties, but what I've heard here the last day 18 and what I'm getting from you, Tim, is that there are 19 increasing numbers of uncertainties, and I think your 20 point was that there was little basis for making some of 21 the decisions regarding the EBS.
22 Why is there little basis? Is this because 23 it's fundamentally impossible to obtain the information?
24 Is this requiring policy types of decisions? Is it es
!(v ) 25 because we've come at it too late? Why are we -- why is NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 13231HODE ISLAND AVE., N W.
(202) 234-4433 WASF 4GTON, D C. 20005-3701 (202) 234-4433
140 1 there little basis?
7-2 MR. McCARTIN: Well, I'll give my I
( )
\
3 understanding, and I'll ask others in the group if they 4 have comments to chime in. Not all aspects are that 5 uncertain.
6 Certainly corrosion rates, we have 7 experimental work and et cetera, and I think we're in 8 better shape with respect to the corrosion aspects. But l 9 when we talk about water moving into a drift -- and once 10 again, 10 millimeters per year -- it's very difficult to 11 conduct an experiment of anything that's relevant to that.
12 Likewise, water dripping onto the container, 13 then moving -- you know, we're talking single drips,
, . N
/ l 4
l (_/ 14 moving into the container. You're right. There isn't a i
15 lot of basis. We know -- assuming it all gets in there is j i
16 conservative. There's an aspect of -- that one of the )
l l 17 things that we're looking at is there's that initiation of l
! l 18 the first pit that maybe occurs, let's say, whatever, at l 19 3,000 years.
l 20 As the first pit forms, there are only --
l l 21 there just isn't one pit. There are multiple pits, and l
22 the question is: How long does it take until you get 23 many, many pits which could impact? Well, do you want to 24 take credit for water just rolling off this container, if rT
( ,) 25 after a couple hundred years, there's tens of thousands of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
141 1 pits?
,, 2 There's a fair amount of uncertainty, but some s /
3 of it's related to how quickly the container degrades to 4 many, many pit holes, and that's one part of it.
5 The dissolution of the fuel is one that, once 6 again, you get into the problem of accelerated testing, 7 and you're doing an experiment on the order of hours, 8 maybe days, and now you're extrapolating the dissolution 9 rate to tens of thousands of years, and was that the right 10 experiment. And that's always a dilemma.
11 There is information out there, but you really 12 have to use it very cautiously when you start 13 extrapolating to long time periods. That's why I gave the rw s- 14 one example for the EBS where if we assume this 15 dissolution rate and a 1 millimeter particle size, 16 everything's gone in 27,000 years.
17 The only way you can evaluate those 18 parameters -- they might have been based on experiments, 19 but you want to say, Does that make sense to me; do I 20 expect this to be all dissolved in 27,000 years. And 21 that's part of the extrapolation.
22 Margaret?
23 MS. FEDERLINE: I'm a little concerned about 24 what it is that we're defining as our objective here. I r 's i
sa ) 25 think everyone has agreed that we cannot accurately NEAL R. GROSS l COURT REPCRTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
142 l
1 predict over the long frame these kinds of things. What ,
,- 2 we're trying to do is define a reasonable test of
( )
i \
' '/
3 repository performance, given what we know today. i 4 I mean, I think we have to set an objective 5 that is doable. We can't set ourselves up to do something ;
\ !
\
l 6 that's impossible, and then say, There's no data. You 7 know, we -- chemical wastes have the same problem, and 8 they go through a five-year cycle, sort of looking at 9 whether this is acceptable. l 10 So I'm a little concerned. There was a whole 11 series of articles in Physics Today that discussed, do we 12 need to sort of rethink the whole thing, based on the fact 13 that data's not available. And I really think we have to p\
l 6
(.s'i 14 think about defining the problem. What is it that we're l
l 15 trying to do with these performance assessments?
16 And when I think Tim says we don't have data, 17 we don't have data and we're not going to have data for 18 10,000 years. You know, it's just not possible. So I 19 think we all sort of need to agree at the outset, what are 20 these performance assessments doing. And I think they're 21 more trying to design a test, based on what we know today.
22 CHAIRMAN GARRICK: I think that's a very good j 23 comment, Margaret.
24 MR. McCARTIN: Any other comments or --
rh
( ) 25 CHAIRMAN GARRICK: Any other comments at this w/
! NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS l
1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
143 i
1 point?
i
~s 2 (;No response.)
i 1 ,'~'j 3 CHAIRMAN GARRICK: You will have a roundtable 4 a little later.
5 MR. McCARTIN: We'll go to the saturated zone 6 flow and transport. The key elements, as you --
7 volumetric flow in production zones, basically getting at 8 the dilution of radionuclides at the wellhead, if you 9 will.
10 There's also the retardation in the saturated 11 zone, either the production zones of the of the tuff or 12 the alluvium, and there's a couple things to be aware of.
13 Hydrodynamic dispersion certainly tends to
/
\-)' 14 spread the plume out; flow channelization, which tend to I
15 constrict it; retardation in the production zones in tuff; 16 and retardation in the alluvium, as you get closer to the ;
17 Amargosa Desert, sort of beyond 15 kilometers from the l
18 site. We believe that that's about where the water table !
I 19 will go into the -- from the tuff into the alluvium. l 20 Our conceptual model approach is we're looking i
21 at volumetric flow in the production zone, and as Gordon 22 was here during the technical exchange was explaining, 23 we're looking at some thickness from the water table to 24 the bottom of the first production zone in the tuff l <
\
i s , / 25 aquifer as a thickness of where the plume would be NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005 3701 (202) 234-4433
1 144 l
l 1 dispersed over. We will not assume that the plume is l ,_s 2 dispersed over more than one production zone. That l ( ' ' ')
3 thickness was on the order of 40 to 120 meters.
l 4 Retardation in the production zones and l
5 alluvium: Certainly we're taking account for the 6 variation velocity from the repository to Amargosa Desert.
7 However, retardation in the tuff, we're assuming the zones 8 are fracture flow; we're not assuming matrix diffusion or l
9 retardation in the fractures.
10 However, when you get to the alluvium, it's a 11 porous media. We assume that there'll be significant l
12 retardation for a number of nuclides. However, as was 13 pointed out in the technical exchange, a couple of the
,r~
(_-) 14 nuclides that you're very interested in, neptunium, 15 technetium, will probably have limited retardation in the 1
16 alluvium. They're fairly mobile.
l i l
17 One -- drawing upon the figure that -- this l 18 sort of recreates the -- a picture of the two-dimensional 1 19 modeling that Gordon Wittmeyer's doing, and here's one of i 20 those abstractions where we are doing some detailed 21 modeling, and the repository's up here.
22 You can see the flow lines going down to the 23 Amargosa Farms area, and we have some distinct flow paths 24 that we're accounting for in different strata, and we
/ \
25 would be true to the saturated zone detailed model in our (v)
NEAL R. GROSS
! COURT REPORTERS AND TRANSCRIBERS I 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
145 1 abstracted model, which as Bob showed, is this pipe model l
- 2 going out. Those pipes are consistent with these flow
( )
3 lines in terms of the properties. We'd be looking at 4 different retardation potentially when you get into the 5 alluvium.
6 But that's -- you can look at this particular 7 abstraction, and to me it says that a 1-D pipe model 8 abstraction for this isn't too bad. That, I think, will 9 capture the movement of the plume across this zone fairly 10 well at this location.
11 Uncertainties: Certainly the volumetric flow 12 in the production zones and retardation in the production 13 zones, and once again if we ran the model today, we would g
i s
(_) 14 have no accounting for matrix diffusion or retardation in 15 the fractures for tuff. But that's one that, you know, 16 we're certainly -- we're going to be looking at that a 17 little more carefully.
18 Once egain, the unsaturated / saturated flow 19 under isothermal conditions, which is what USFIC stands 20 for, that particular KTI is primarily responsible for all 21 the abstractions in the saturated zone, be it the 22 parameter determination, the abstraction itself, and the 23 detailed 2-D modeling that's being done at the center.
24 In terms of anticipated sensitivities, once rx I ) 25 again, the fact that we're not having retardation or
%/
NEAL, R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
, (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433 l
l
I 1 146 i
1 1 matrix diffusion in the tuff, that's certainly important.
, s 2 We think we will see a difference between the different
\l \
l \
'- 3 locations, be it in the tuff or in the alluvium, where we i
4 will have retardation in the alluvium.
l 5 There'll be long travel times in the alluvium.
l j 6 It's basically, obviously porous media flow, so the
! 7 porosity is significantly higher than the fracture system, 8 so we would expect long travel times, thousands of years 9 in the alluvium potentially.
10 And one of the things, pumping rates at the 11 critical group -- we'll be looking at different pumping 1
1 12 rates. In general, I think we're looking at ranges from a 1
, 13 half to 10 million cubic meters per year, depending on the j ll/~s 1
!(/ 14 particular lifestyle of the critical group.
15 And our abstraction to date -- we are assuming l l
l l l
16 all the releases from the repository get to the critical 17 group at the location where they are, and then there's 18 this mixture into the pumping needs of the critical group.
i l 19 And it's at a later slide, but this is probably as good a 20 place as any.
l 21 There's really three volumes of water that we 22 consider in terms of determining the concentration.
l 23 There's the volume of water going through the footprint.
24 There's the volume of water in the production zones in the n
( 25 tuff, and that's where that thickness of 40 to 120 meters
! NEAL R. CROSS i
COURT REPORTERS AND TRANSCRIBERS l
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
147 1 would come into play. And then there's the pumping rate
,- 2 of the critical group.
3 We look at those three volumes. Whatever one 4 is the largest, we would use that to determine the 5 concentration at the wellhead for the critical group, and 6 depending on the particular location, the one will 7 dominate over the other. We're anticipating it will 8 generally be either the flow in the production zone and 9 the tuff or the pumping rate of the critical group.
10 It's hard to imagine that the flow through the 11 footprint will exceed some of the other things, but be 12 aware, obviously when you get into the pluvial case, where !
l 4
13 if you get into 100 millimeters per year of infiltration !
(~5 l
\_/)
I 14 through the footprint, you have approximately 5 million 15 square meters of footprint. You are talking on the order i 1
16 of 500,000 cubic meters of water, so it is a -- that )
I 17 footprint's pretty big, and so 100 millimeters per year j 18 does add up to a nontrivial amount of water. i i
19 VICE CHAIRMAN HORNBERGER: Tim, just a point 20 of clarification. You're talking about volume really l
21 being -- the volume being --
l 22 You've gone through a chain of calculations 23 that tells you what the concentration at a given time is 24 in the groundwater at Amargosa Valley, because you've gone f
( )) 25 through the unsaturated zone and saturated zone. Now you l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS I 1323 RHODE ISLAND AVE., N.W.
! (202) 234-4433 WASHINGTON, D.C. 20005 3701 (202) 234-4433
148 1 have a concentration.
py 2 Why do you need this volume to --
C )
3 MR. McCARTIN: Not really. For NEFTRAN, it --
4 it's transport. And it's treating it as a trace species, 5 and it's not -- we provide a groundwater velocity to it, 6 that it then uses to move the nuclides to the exit point.
7 It releases curies per year of the nuclides. That's the 8 only value it produces is a flux of radionuclides.
9 You then have to specify -- and we do it in 10 the context of the TPA code -- what volume of water are 11 these nuclides moving into, and that's where --
12 VICE CHAIRMAN HORNBERGER: I guess there could 13 be a disconnect here, because, yes, it gives you curies (3 i t
N/ 14 per year, but you've already told it what the flow rate 15 is, and therefore it's not a very difficult computation to 16 calculate it up to a flux of radionuclide.
17 MR. McCARTIN: It doesn't use the cross-18 section, though.
19 VICE CHAIRMAN HORNBERGER: I know; I 20 understand. You're only using a two-dimensional model.
21 But it doesn't take too much to figure if you have 120-22 meter depth of aquifer that you're considering, to scale 1
1 j 23 it up as well.
24 MR. McCARTIN: Right. Yes, o
k,) 25 VICE CHAIRMAN HORNBERGER: Presumably you're
! NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
149 l
1 doing some of these computations to just make sure you're
,s 2 not off, you know, doing something that's three orders of
(
I 3 magnitude different.
4 MR. McCARTIN: Well, right. But it's as much, 5 let's say -- we're assuming all the contaminant is in a 6 single production zone. I 7 VICE CHAIRMAN HORNBERGER: Right. ,
8 MR. McCARTIN: Ana let's say that the flow in l I
9 that production zone is 200,000 cubic meters per year, so 10 we can get X concentration in the aquifer. However, we're 11 at the wellhead, and if what appears to be a reasonable 1
12 assumption for the critical group that we've set, that 13 their water needs, for example, are a million cubic meters r 's
- 14 per year, they obviously are pumping other fresh water. l 15 And so then we would increase it to a million l l
16 to account for the fact that, yes, they're pumping the 17 plume, but they have to be pulling in other fresh water.
18 And it gets to the difficult problem of, Where do you 19 locate your critical group. I mean, for us, we're putting 20 them smack-dab in the middle of this plume. We're ,
1 21 counting on them being right there.
22 But their water needs could be greater than 23 what is just in the plume, and that's what we're trying to 24 account for, and it's -- you know, we're certainly being r~%
() 25 conservative in our placement of the critical group. But NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
150 l
1 the allowance for looking at the water needs of the
,s
, 2 critical group, I think, is taking into account other i a (s~#
3 aspects that, I think, are worthwhile.
4 Direct release for us is related to the 5 volcanic scenario, and there the key abstractions are 6 re]ated to the probability of the volcanism, entrainment 7 of the waste in the ash, and the air transport of the ash.
8 The probability of volcanism is based on the 9 Center model. Entrainment of the waste in ash, there's a 10 lot of things that go on there. The incorporation ratio 11 is used to account for how much stuff is brought up by the 12 magma; the grainsize of the spent fuel will impact the 13 calculation.
, /j
\-) 14 We're making an assumption that the waste 15 package does not affect the entrainment, and that you look 16 at the size of the activity and you look at how many 17 containers are affected. That amount of fuel is capable 18 for entrainment. There's no impact of a container.
19 The air transport of the ash is related to the 20 magnitude of the eruption.
21 Certainly the, once again, uncertainties are 22 with the probability. The magma / waste-package l 23 interactions, there's obviously a lot of tnings that could 24 be going on there. We're being -- taking a conservative
/^N
(_) 25 approach for now that there is no impact due to the l NEAL R. GROSS
! COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
1 151 1 container, et cetera, the cladding that decreases the
.s 2 ability of the magma to pull up radionuclides.
(
)
3 And the air transport of the ash, we're 4 looking at analogous volcanic events, and as Bob 5 mentioned, the Center researchers recently compared the 1
6 results of their model with Cerro Negro and got fairly '
7 good -- actually very, very good results in comparison. .
l 8 In terms of the KTI work on this one, it's 9 primarily igneous activity and the container life and i 10 source term people interacting with them in terms of the 11 grain size of the spent fuel.
12 In terms of anticipated sensitivities, the 13 energy of the eruption is certainly going to affect the q
\
x- 14 number of containers intercepted. For now, we're looking 15 at approximately five to ten containers will be affected 1 16 by an event. Ashfalls will be in range of centimeters to ;
17 millimeters at a distance of 30 kilome.ers.
18 Particle size and incorporation ratio 19 certainly affect the number of radionuclides dispersed.
l 20 Peak dose in volcanism is very sensitive to the time of l
l 21 the peak dose. This is a very interesting aspect to the 22 volcanism dose calculation relative to the other parts of l 23 the calculation.
l' 24 Generally, the groundwater releases, there's a g
( ,) 25 dose that goes up, and it plateaus for a very long time NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
152 1 period. It persists for hundreds if not thousands of 1 l
m 2 years.
.I s
\
/
'~'
3 Unlike that is the volcanism dose which tends 4 to decrease significantly ten years after the event.
5 There's -- primarily because you don't have a volcanic 6 event every year. You have one event. It puts down some 7 ash. You get a dose for that year. Year after year, l 8 there's more dilution, if you will, in the soil, et l l
9 cetera, and after ten years, the dose has decreased l l
10 significantly. I l
11 Also, the magnitude of that peak event is 12 related to how quickly do you start farming in the ash l l
13 after the event. And for now we're assuming that the l
/~~N >
I
< 1
(_/ 14 volcanic event does not disrupt the lifestyle of the l l
15 critical group, so that there'd be farming in that year l 16 after the event.
l 17 Also for a closer-in person, where they l
l 18 wouldn't be farming, there's factors of the resuspension l l
19 factor, and we're looking at -- it's also assuming that 20 the people are still living close in.
21 It's also very sensitive to the -- obviously, 22 the location of the critical group and the wind speed.
! 23 MEMBER POMEROY: Tim, does this computation 24 now include resuspension of material?
(q) ,
25 MR. McCARTIN: Yes. It's quite sensitive to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS i
1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
153 1 that, and that's also one of those parameters that 73 2 immediately after the event, there's a lot of fine ash in
( )
~
3 the air, and it decreases very dramatically, and 4 there's -- we're looking into the sensitivity of that and 5 what seems to be a reasonable parameter for calculating an 6 annual dose.
7 There's the person inside of 20 kilometers who 8 isn't a farmer, and there you have a direct and inhalation 9 pathway relative.to the farmer who there isn't as 10 critical, because you have the ingestion pathway which 11 dominates, and the direct and inhalation isn't important 12 there, but as you get in closer, that's where that tends 13 to show up.
[ i i /
N/ 14 But, yes. Resuspension factor is one of those 15 that is a sample parameter. We're looking into -- it can 1
16 vary by quite a bit in terms of what seems to be l 17 reasonable for this particular environment.
18 MEMBER POMEROY: I heard a few very large 19 numbers early on, because -- and I was curious --
20 MR. McCARTIN: For the resuspension factor 21 or --
22 MEMBER POMEROY: For the effect on the 23 ultimate dose.
24 MR. McCARTIN: Yes, yes. On to the dose TN
(,) 25 calculation, the final part of the TPA code: Some of the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
l l 154 l
1 things we've touched upon, certainly dilution of
,s 2 radionuclides in the groundwater, getting to the pumping
(
3 well characteristics, water use of the critical group, 4 dilution of radionuclides in the soil from the direct 5 release, plowing and surface processes, and the 6 resuspension factor, and ultimately the location and 7 lifestyle of the critical group, specifically the diet of 8 locally grown food for a farming community.
9 For our particular calculation, we're looking 10 at the dilution of radionuclides in groundwater via the 11 way I explained briefly before. I won't go through it 12 again, but it's once again keeping track of the volumes of 13 water, either flowing through the footprint in the k_j' 14 saturated Zone or at the critical group. We use the 15 larger of that particular number.
16 Dilution or radionuclides in the soil: For 17 farming practices, we're looking at the current farming 18 practices for determining those parameters.
19 Location and lifestyle of the critical group:
20 We're using a representative person to calculate the 21 average member to the critical group, so we have a --
22 we're calculating a single dose for this average member.
23 And the habits, characteristics are based on, as the 24 National Academy of Sciences recommended, cautious and
/
l (%)
%d 25 reasonable assumptions.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
155 1 We're looking at two locations, as I 7s 2 mentioned. Inside of 20 kilometers, we're looking at a 3 non-farming critical group; outside of 20 kilometers, 4 we're looking at a farming critical group. We believe 5 that's reasonable and cautious.
6 The reason for that is twofold. One, the 7 depth to water, that we don't think it's that reasonable 8 to count on a farming community drilling very deep wells 9 for water. Also, the soil characteristics, and we're 10 certainly looking into the reasonableness of that 11 assumption. We believe -- that's what we're currently 12 approaching the calculation with.
13 The uncertainties, the dilution of g
i
\_ / 14 radionuclides in the groundwater: We think we're using l
1 15 the minimum yields for the particular areas that we're l l
16 looking at; flow through the footprint. Our production 17 zone volumetric flow is based on a single production zone, 18 and the pumping of the critical group, we think, will be 19 appropriate for the critical group.
20 We'll certainly be looking at -- once again, 21 the DOE is doing far more sophisticated analyses in multi-22 dimensional space with respect to the saturated zone, and 23 we'll be watching that to see if you can justify a much 24 larger dispersal of the plume in the saturated zone.
( ,) 25 Right now, we're relatively uncomfortable assuming
! NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
156 1 hundreds of meters of dispersal in the vertical direction
-s 2 for the saturated zone. But we're looking at the
/ i
( /
'^'
3 information to date.
4 I mean, I'll point out the seawell complex.
5 If you look at the volumes of water they did in their pump 6 tests, clearly in the saturated zone via fracture flow, l l
7 there's not a lot of storage with the fractures. To get )
1 1
8 the volumes of water they did and not draw down to air, 1
9 there has to be a lot of conductivity between the 10 fractures, a tremendous amount of conductivity.
11 However, it's one thing to have the fractures 12 all connected and be able to pull water in from them, and 1
13 having the plume disperse in the vertical direction I
n 4
(.s/ 14 significantly. Now, part of that will be related to the i
1 15 fault zones and the structures there.
16 Once again, we'll be looking at a lot of the l
17 work that the DOE is doing with respect to that 18 abstraction, and also, I know, Bob Baca did some cross-19 sectional analyses where we looked at in two dimensions, 20 looking at the structure and the ability of structures to 21 disperse the plume in the vertical direction.
22 So we haven't necessarily said we'll only go l
23 to one production zone, but for now, it seems prudent to 24 look at dispersal in that particular area, and as you look
/ s (ss ) 25 at the pumping rates, it may or may not be that critical i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
I l
157 l
1 in terms of the overall vertical dispersal of the plume.
i
, -<~3 2 The dilution of radionuclides in the soil, 3 we're looking, once again, at the parameter ranges there, 4 and we think it'll be consistent with current farming j 5 practices.
l \
6 And location and lifestyle of the critical 7 group: As I mentioned before, we're putting the critical 1
8 group right smack-dab so they get all the radionuclides j 9 from the repository. We think that's just about as I
10 conservative as you can get. We're going to use the J l
11 reference biosphere and cautious and reasonable l 12 assumptions. 1 13 And as you noted in the technical exchange, i I t \/ 14 DOE has done a survey of that area. We certainly are 15 interested in looking at the information from that survey 16 to help us define, what are some of these reasonable 17 assumptions, and are they defendable.
18 Well, here's one of the few modules of the 19 code that I can champion the TPA group, the performance 20 assessment group. You'll see we're primarily related to 21 the dose calculation, and we're looking at the parameters, 22 the abstraction, and the detailed modeling with the GENII-l 23 S code, which is being done at the Center. One of the few 24 that's dominated by PA, you'll notice.
O.
( ,), 25 In terms of our anticipated sensitivities, i NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS
( 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005 3701 (202) 234-4433
158 1 right now we're doing a -- it's a table look-up as Bob s 2 showed in his overview of the code, and for now, we
' )
3 have -- we're assuming 50 percent of the diet is locally 4 grown for the farm location, and the water and milk is 100 5 percent.
6 We're using what we think are sort of average 7 lifestyles over the entire community. How these -- what I 8 mean by that is Pat LePlant [ phonetic] from the Center 9 looked at all the kinds of things that were going on in l
10 that area of Nevada and averaged them all together, so l 11 there's a little bit of everything that's going on there 12 is included in our 50 percent value. l l
13 And the resuspension factor will be -- does t )
\/ 14 have a direct impact on the inhalation dose, primarily for 15 the non-farming community where -- in volcanism.
16 And once again, size and location of the 17 critical group, once again we're assuming the fact that 18 our critical group gets all the radionuclides. We're not 19 looking to try to move the critical group and capture -- a 20 critical group is over here and gets maybe a portion of 21 the plume. It's -- for now, we think this is the easiest 22 way to go.
l 23 And with that, I'm done. I'll be happy to 24 answer any questions, comments by other NRC staff members
,ex
( ,
) 25 that want to correct any mistakes I might have said.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433 l
l 159 1 VICE CHAIRMAN HORNBERGER: Not wishing to beat l
7._
2 a dead horse totally, but if you have a certain number of I /
\'~' )
3 curies per year at the well, number one, and, number two, 4 you make the assumption that the total inventory is 5 intercepted, and, number three, you have a certain volume 6 pumped that year, how can there be any leeway at all for 1
7 determining the concentration? How can you not use that 8 volume?
9 MR. McCARTIN: The pump volume?
10 VICE CHAIRMAN HORNBERGER: Yes.
11 MR. McCARTIN: If the flow in the production 12 zone is greater than the pump volume, then we would use 13 the flow in the production zone. It would be more --
(._) 14 VICE CHAIRMAN HORNBERGER: Then how could 15 you -- then isn't that inconsistent with the middle 16 assumption, that the whole inventory is intercepted, 17 because if you don't pump it all out, then some of it --
18 MR. McCARTIN: That's correct.
19 VICE CHAIRMAN HORNBERGER: -- has to go 20 through.
21 MR. McCARTIN: That's correct. You're 22 correct. Yes.
23 A better way to state it is: The entire l
24 inventory's available to the critical group. Maybe that's t
_s) 25 a -- yes. A good point.
f NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
160 1 The only other thing, I guess, I would like to
,_s 2 say with respect to the PC workstation environment, in the I \
3 for-what-it's-worth-department, that moving to the PC 4 environment with this code is a useful thing for, I would 5 say, individual technical staff, to be able to use the 6 code with the computer environment they're used to, on a 7 limited basis.
8 The problem that -- I don't think we're ready 9 to -- and that's a very good purpose for it. But in terms 10 of the -- I want to at least recommend that for the ACS, 11 there still is a need for the ACS, in that the Sun 12 workstation, one of its claims to fame is it has massive 13 storage, g3
(_ '1 14 It also has the ability to tie in a number of 1
15 the machines to do the same problem, and as Dick Codell is 16 a master at squeezing the maximum efficiency out of our l 17 computer resources with tying all these together and using 18 it, when we go into -- in the PA, the total system 19 analysis, and do, say, 400 realizations, it will require a 20 lot of storage, a real lot of storage and a lot of CPU 21 time.
22 Tying all those Suns together is the main l 23 advantage of the ACS for that particular calculation, 24 because, remember, as Bob pointed out, there's a lot of 7 ~.
25 intermediate calculations that we're storing. We want to (v) l NEAL R. GROSS i COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l WASHINGTON, D C. 20005-3701 (202) 234-4433 (202) 234-4433
1 l 161 1 save that information, so that we see realization number
,s 2 129 looks a bit funny. Well, we have to be able to go
, s
( i
' 'j 3 back and look at the results, the intermediate results, so 4 we do save a lot.
5 And for the PC environment, I think even a 6 high-end Pentium would have a very, very difficult time 7 running this for the total system performance, and that's 8 where the ACS is of extreme benefit for us. And so, you 9 know, it's useful for the isolated user, debugging it, 10 using it, running it, and doing what-if games and doing 11 limited analyses, but when you get to the total PA 1
12 picture, it would stress a Pentium.
13 MR. McCONNELL: Our goal is to get it on Carl n
C'1 14 Paperiello's PC sometime next year. Very interested in 15 that.
16 That more or less is a lead-in to our next 17 topic which is the discussion of our planned sensitivity 18 studies, both at the process level and the system level, 19 and Virginia Colten-Bradley is the PA lead for that area.
20 MS. COLTEN-BRADLEY: I'm Virginia Colten-21 Bradley. I'm a geochemist with Farm Assessment Chain, and 22 I had the occasion to work with some of you last summer.
23 Today I'm going to talk to you about i
l 24 sensitivity and certainty analyses for this particular
(%
l (v ) 25 phase of the performance assessment at the NRC. I'm going NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 mWT9
r 162 l 1 to talk more about technique and how we go about tearing 7- 2 the code apart than some of the conceptual things that Tim t
' ') 3 just talked to you about.
4 Our objectives -- and this is an outline 5 primarily that I'm going to talk about. I'm going to talk 6 a little bit about what we hope to gain by doing the 7 sensitivity analysis, how we're going to do it in terms of 8 our approach, and also share with you what we did in IPA 9 Phase 2, and then what we're going to use the results of 10 that analysis for.
11 The primary objectives of doing a sensitivity 12 analysis from this subsystem level is to determine the 13 parameter importance to the output, and we also look at --
/3
! )
\_/ 14 we look at the output not only on the whole system or the 15 total system in the context of dose, but we'll also be 16 looking at the sensitivity to intermediate outputs that 17 Tim just described to you.
18 We're interested in quantifying that 19 sensitivity, as well as the uncertainty in some of these 20 parameters, and so by going in and looking at the output 21 from a variety of the different files, we get a sense of 22 what things look like.
23 We want to tear the code apart in the sense of 24 looking at some of the biases that go into doing the x _-)
s 25 modeling exercises. We've talked about a number of l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
163 1 conceptual models that have gone into deciding how we 7~- 2 model things today, and so what we want to do in doing i'"/
3 this sensitivity analysis is to go in and see how those 4 biases are affecting the results.
5 We have flags set to calculate things in 6 certain ways, and we can do the sensitivity analysis by 7 setting the flags one way or the other in terms of 8 determining how that affects the result.
9 We're also important in determining what parts 10 of the system are significant to the overall performance, 11 and that will feed in eventually to an importance analysis 12 that Norm Eisenberg is going to talk to you later about.
13 And the other purpose of doing this whole
/ i
\# 14 activity is to develop our review capabilities. Our job 15 in the end will be to review the DOE's performance 16 assessment model, and so if we've done it to our own, then 17 it gives us a chance to get practice in doing theirs.
i 18 Keith showed you this diagram this morning, l 19 and I just put it up here to illustrate that we're doing 20 sensitivity analysis on three different levels. We're 21 doing it at the key element level or the module level, to ,
1 22 test the process models and components of the system, and 1
23 then we're also going to do it at the total system level, l
24 so it covers all three levels of this diagram in terms of 7x
( ,) 25 how we're doing the sensitivity analysis, but we do them NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202; 234-4433 l
l
164 l
! 1 individually.
l l 7- 2 In IPA Phase 2, we used a variety of i (~~/'
)
! 3 techniques that Helton and Iman [ phonetic] had developed 4 at Sandia and had been already applied to performance 5 assessment sensitivity analyses. We just brought them in 6 and applied them to our own model. We used a couple of 7 other techniques that had not been used, and these are all 8 written up in the NUREG 14.64 which is the document that 9 describes our IPA Phase 2.
10 In order to determine the most significant 11 parameters, we used the step-wise regression technique, 12 which goes through and picks one parameter at a time and 13 adds it to a multi-dimensional model, and then based on i )
\/ 14 statistical tests, keeps it or throws it out.
15 We also used the Kolmogorov-Smirnov test, 16 which I'll describe to you in a couple of minutes. And we 17 used scatter plots, just the simple one-by-ones in terms 18 of one input versus the output, to see what kinds of 19 correlation and coefficients you might get, and you can 20 visually, very quickly, scan to see which parameters are 21 important.
22 We also determined the relative -- we were 23 also interested in the surrogates for the subsystem l
i l
24 performance measures that were outlined in Part 60 in IPA
( ,s) 25 Phase 2. We had several different ways of measuring NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D.C. 20005 3701 (202) 234-4433
i 165 l 1 travel time that we looked at. We had a container life
- 2 that we measured, and then we also looked at release
'7- \
l ( ;
~'
3 rates, the 10-5 rule, so we kept data on those particular 4 subsystem requirements, and then we compared our outputs 5 to those.
l 6 We were also interested in describing the l 1
7 variability in the output of the system, and so we used a l
8 variety of different distribution plots, such as the box 9 plots and histograms, which are real common.
10 The Kolmogorov-Smirnov test is a non-l 11 parametric test, and we got very good agreement using this l
12 technique relative to the step-wise analysis. The IPA 13 Phase 2 runs consisted of 400 realizations, and we took I l
r~x\ l
\
l 5%s/ 14 the top 40 realizations that gave the largest cumulative 15 releases.
16 In other words, we took the top 10 percent, l 17 the tail, so to speak, and of those 40 realizations, then 18 we took all the sample parameter values that corresponded 19 to those 40 realizations that gave the top cumulative 20 releases, and then we plotted the distributions that 21 corresponded to those particular parameters.
22 For example, in the case of this particular 23 parameter, which was one of the hydrologic parameters for l 24 fractures, the theoretical distribution's given as the
<^N
( ) 25 solid line, and then the sample values that came out of
' vs NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
166 )
1 the LHS sampler are given in the dets.
73 2 In the case of the infiltration parameter,
'( ~#
)
3 again, here's the theoretical distribution is the solid 4 line, and the dots represent those that were sampled by 5 the LHS sampler. l l
6 When there's a large disparity between those 7 two plots, then the parameter actually turns out to be 8 significant to the process, infiltration in IPA Phase 2 9 was our top driver for the system.
10 I also want to note that these plots and ell 11 the analyses that we did were done using an exploratory 12 data analysis program called S-Plus, and so we didn't have i I
13 to do any programming per se. All of the functions are i
/~h l
[ \ l
\/ 14 included in it. You can make scripts with it to automate l 15 the system and let it crank through everything for you.
16 You don't have to do everything individually.
i 17 Our step-wise regression analyses were 18 automated as a batch program that we could put into action 19 and let it go through and pick everything out for us, and 20 the Kolmogorov-Smirnov test was just a set of script that 21 we set, and it went through everything for us. We didn't 22 have to do anything individually.
23 Some of the types of parameters that came out 24 as important to us and to our model in IPA Phase 2: For
'q
( ) 25 the undisturbed scenario, that included no kind of change; NEAL R. GROSS
, COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON. D C. 20005 3701 (202) 234-4433 I
167 1 there was no seismic activity. It was just our baseline
,S 2 kind of present-day scenario. Infiltration was our top i ;
\
3 parameter. It did not matter which technique we used for i
4 selecting the most significant parameters. Infiltration i
5 always came to the top.
6 Corrosion parameters were important. We had 7 some retardation parameters that turned out to be l 1
8 important. And ironically they turned out to be i l
9 radionuclides such as thorium. They really weren't the !
l 10 important ones to the dose. Some of the dissolution rate 11 parameters also turned out to be important.
12 For the fully disturbed scenario, which 13 included climatic change, seismic, drilling and volcanic Y' 14 activity, the corrosion parameters turned out to be the 15 most important ones, and we saw some of the parameters 16 that would be specific to the different disturbed 17 scenarios, such as where the drilling was located.
18 DR. STEINDLER: If the thorium wasn't 19 important to dose, why does it show up as a significant 20 parameter?
21 MS. COLTEN-BRADLEY: It's the way -- it wasn't 22 necessarily the top six. It was like in the top 20, and 23 we were sampling close to 300 parameters, and so I was 24 just trying to find some categories of things that would O). 25 come to the top.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISMND AVE. N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l l 168 !
l 1 DR. STEINDLER: I guess I'm still confused.
l i
l r~ 2 Were there any retardation parameters directly related to N.
~.
,s) l 3 dose that looked like they were in -- roughly in concert, 1
i 4 somewhere between corrosion and dissolution rate, which is 5 where you've got it?
6 MS. COLTEN-BRADLEY: Well, let me back up for 7 a second. We calculated dose differently than we're 8 calculating at this time. We calculated population dose.
9 Does that matter to you?
10 DR. STEINDLER: Fine. No. It doesn't matter 11 to me.
12 (General laughter.)
13 DR. STEINDLEh: I can divide by 200 just like
- V i 14 everybody else. I'm just trying to figure out whether 15 retardation is an important issue, and the importance in 16 relation to dose is what I guess I'm asking.
17 MS. COLTEN-BRADLEY: Well, if -- there were a 18 couple of KTIs that came to the top in terms of, say, in 19 the top 20 parameters, but after number six, if you're 20 starting from number one looking down, there's not much in 21 terms of the real additive to the model. If you took the j 22 top six parameters, you could pretty much match the 23 variability in the model pretty closely. After that, l
l 24 things fall off pretty quickly, and it gets pretty noisy. l 1
i(j s 25 But the --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
169 ,
l 1 DR. STEINDLER: It must be the heat around I l
n 2 here. Let me ask the question again, maybe somewhat
( ' ~ ' /l s
3 differently.
4 MS. COLTEN-BRADLEY: Let me try the answer 5 again. It's not all that important, but it's there; it's I
6 listed. I 7 DR. STEINDLER: It's not number 3 is what 1
8 you're telling me.
9 MS. COLTEN-BRADLEY: No, no.
10 DR. STEINDLER: I'll go quietly. ]
l 11 MS. COLTEN-BRADLEY: And what -- the other l
12 thing I want to point out is the variance -- one thing we 13 noted in IPA Phase 2 is the variance on a parameter really
/3 U 24 makes a big difference in where it's going to come out in 15 the list. And it's -- you can accidentally set the l l
16 answer.
17 Scatter plots are often busy, but on the other 18 hand, they also can add a lot of information. As I 19 mentioned before, we have a surrogate for the subsystem 20 performance measures, and on these two plots, I've noted 21 the fractional release rate of 10-5 as a solid line. The 22 dots are one for each realization in terms of the Carbon l
23 14 releases. Carbon 14 is for all the gases and the 24 groundwater pathways added together.
fs 25 In the case of the Carbon 14 releases, all of
(]
- NEAL R. GROSS l COURT "20RTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N W.
(202) 234-4433 WASHINGTON, D.C. 2000S 3701 (202) 234-4433 i
l 170 1 them exceeded the fractional release rate of 10-5 On the I
\
l l
,- 2 other hand, there was only one realization where we had a
( l 3 fractional release rate of Americium 243 that exceeded 10' 6
4 5 And the nice thing about doing a plot like 1
i 6 this is then you can go in and pick out that particular l
7 realization and go in and see what other values are set 1
8 for the hydrologic parameters or the pumping -- or, you 9 know, something else that is driving that particular 10 realization, to give you a value that exceeds the 11 standard.
12 Now, as I noted, Dr. Steindler, this is a 13 different dose calculation than what we're doing this i
(/ _ 14 time. But these are box plots, and they show the 15 distribution of the values for the fractional l 16 contributions to the population dose by nuclide, and these j I
17 are not all the nuclides that we looked at. These are l
l 18 just the top seven or seven that we just sort of picked to 19 be representative.
l 20 The boxes represent the 25th to 75th !
l 21 percentiles. The white line in the middle is the median, l l
22 and the means are noted, and then what's nice about these 1
l l 23 plots is that individual values are shown out here in the 24 tails, and again that gives you a chance to go out and see
( ) 25 where the really extreme values are located and how they NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
l
l 171 l \
l 1 might be affecting other things in the model.
l ,
- s. 2 In this particular figure, the lead has the
! . ~')
3 highest contributions to the dose in some of the l 4 realizations. ,
l 5 By filtering the data, you can get a lot of I
l 6 information about how the system is working. I've plotted 7 here for you the carbon 14 releases for the entire system, 8 as a function of infiltration rate. This includes both 9 the gaseous and the groundwater pathways, and you can see I
10 there's not much correlation between the infiltration I 11 rate.
l I
12 Our bias or our intuition would tell us that j 1
13 releases should be related to the groundwater pathway if l es l
! \
x/ 14 they're flowing through the groundwater system, and so by l 15 filtering out the gaseous releases, we get a much better 16 correlation.
17 In the current way of doing things, we're 18 going to expand our number of hands who are going to be 19 doing this activity. Last time around, my colleagues Dick 20 Codell and Rose Byrne were my primary helpers in doing the 21 sensitivity analyses. This time, the KTI teams will each 22 be taking a part in this activity.
23 They'll be looking at the process models to 24 look at the assumptions that they've made and helping to im
( ) ,
25 develop those models. They'll be looking at the
! NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(
(202) 234-4433 WASH!NGTON, D C. 20005-3701 (202) 234-4433 i
172 1 parameters that they've chosen, the distribution that
- s 2 they've chosen for those parameters that go into those
,( )
l 3 modules, and even just the -- they'll be also evaluating 4 different ways of modeling things. For example, in the 5 EBS release model, we have several different ways of using 6 the release rate.
7 So at the subsystem or performance model --
l 8 excuse me -- process level model level, again the teams' 9 primary interest is to determine which parameters are most 10 important to any particular module. They'll also be 11 looking at the most important parameter to dose, and we're 12 using that primarily as a benchmark from one team to the 6
13 other.
kY~ 14 Even though all the other parameters outside 15 of a particular module will be set to a nominal value, the 16 teams will be calculating -- looking at sensitivity within 1
1 17 the module to output from that module as well as to the J 18 dose. l i
19 This sensitivity analysis can be done in two 20 different ways. The teams can choose to sample their 1
21 parameters that they're using for that particular module, l
22 or they can just do it in the what-if kind of mode, where l l
! 23 they just choose to change the value of one parameter by 24 some value. They don't have to sample everything that
,G
! ) 25 goes into that model.
\_/
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
i 173 ;
1 Now, one of the things that we are going to do l l
l g-~3 2 in order to sort of benchmark or to sort of even out the ;
Y.Y 3 playing field for all the teams in terms of doing their 1
4 sensitivity analyses so that we can interpret results in a j i
i 5 more even fashion is we will take that reference database i 6 that was talked about this morning, and we'll set all the 1 7 parameters to a nominal value that each team will use to )
8 start with, i 9 And it may be just the mean of the 10 distribution; it's something relatively simple. And we'll 11 set certain flags for calculating certain things in a 1
12 specific way. And the team, depending upon which module
, _ , 13 they're looking at, will change parameters accordingly.
i / h 5# 14 The results of that analysis -- yes, sir.
15 DR. STEINDLER: If you choose the what-if mod, 16 do you make the assumption that the parameter that you're 17 picking -- somewhat arbitrarily picking a number or l
l 18 whatever, is independent of all others?
19 MS. COLTEN-BRADLEY: Yes. Now, you could also 20 choose that it's correlated to something, too. And 21 hopefully you would vary that accordingly.
22 One example of the relationship might be l 23 with -- we have this infiltration parameter that's going 24 to have an effect on a lot of different things throughout O
( ,) 25 the repository, and while you could set the infiltration l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
1 174 l l
1 to a nominal value and leave it alone while you're fooling
,_ 2 around with things that happen with EBSPAC, you probably l \
3 want to change it too, just to see how the system is i
4 behaving as you change it, because the corrosion rates, I l
5 the relative humidities, all those things are also 1
6 interrelated to that infiltration. You don't just want to )
7 leave it at a single value.
8 As I noted in -- the base case is sort of our l 9 reference case. The difference between this base of our 10 performance activities and what we did in IPA Phase 2 is 11 that this tim the climate change is included in the base ;
1 12 case. We have time-varying infiltration. Last time we 13 did not include it as part of the base case. It was a
(_ 14 separate scenario.
15 Some of the flags that I talked about a little 16 bit are those where we choose to calculate something in a 17 certain way. For example, in the module EBS release, 18 there are three different dissolution rates that could be 19 used. One would be for the dripping model. One would be 20 for the bathtub. And one is using the natural analogs 21 data from Pena Blanca.
22 That base case then, again, will be used by l
23 all the different teams. They will also be updated after 24 the sensitivity analysis is completed, and we might have
,m
( ) 25 to refine some things based on the results, and then we'll NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
175 1 use that then to go into production runs.
,s 2 I included in your packet an example for all
- c i 3 five of the different process levels, but I'm just going 4 to talk to two of them, for sake of time. Give Dr.
5 Steindler one more chance to fire at me.
6 Anyway, the -- just to give an example of how 7 some of these parameters might be used in doing the 8 sensitivity analysis and how this ties in to our 9 abstraction process, in the waste package corrosion area, 10 some of the parameters that might be varied -- and, again, 11 this is only examples; these are not fully inclusive of 12 the list, because in EBSPAC, there are close to 200 13 parameters.
r~N
<( )
i
\_/ 14 The teams will be looking at chloride 1
15 concentration and the critical relative humidity in terms 16 of how that affects the output from EBSPAC. And then l 17 specifically they'll be looking at failure tim and number 18 of container failures.
19 To sort of address the questions of 20 uncertainty and lack of data, I've chosen to show you this 21 one, because we don't have a lot of data concerning 22 retardation coefficients for the alluvium, and so what the 23 sensitivity analysis can show us in this case is whether 24 or not it really makes a whole lot of difference.
x (R./ ) 25 We can choose a whole variety of different NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS I 1323 RHODE ISLAND AVE., N W.
I (202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 2344 433
176 1 values for the retardation coefficients in the alluvium
,s 2 and see whether or not it makes much of a difference to I 't
'~'
3 the final dose. And if it does, then that suggests that 4 we need to go back and think about things and make some 5 comments to DOE, and if it doesn't, then we can just leave 6 it alone.
l 7 After the teams complete their sensitivity 1
8 analyses, as you've seen, we've been doing a lot cf 9 integration as a group with the different teams. We'll be 10 meeting again in a large group or individually with the i 11 teams, to go over the results of the sensitivity analyses.
12 They'll be writing these up for their issue 13 resolution status reports and to include in the annual g~s i )
(_/ 14 report, but then we'll come back as a group; we'll .
I 15 evaluate the results as a group; and then we'll redo the 16 reference data set, based on those results.
17 In some cases, parameters -- because we have 18 the potential of over 1,200 parameters in our database, we 19 don't necessarily want to sample all of them for 20 computation time, and so things that turn out to be very 1
21 insignificant at the module level, we may choose to just l 22 set as constants and leave them as that for the final i
l 23 runs.
24 In some cases, we may choose, even though they
(~s l (v ) 25 appear to be insignificant at the modale level, we may NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, O C. 20005-3701 (202) 234-4433 l
l
177 l
l 1 choose to leave them as sample parameters for the final 1
- 2 runs, because we have a belief that they're still going to l
\ !
~'
l 3 be important at the overall system level.
4 We'll be using similar techniques to what we 5 used in IPA Phase 3 -- excuse me -- IPA Phase 2. We'll be l 6 using the step-wise regression technique primarily to sort 7 through that close to 300 or 400 sample parameters, which 8 we'll be sampling with the Monte Carlo or LHS sampler.
9 We'll use the Kolmogorov-Smirnov test again, 10 and hopefully we'll get good agreement. And now that I 11 know how to do surface plots, we'll be looking at 12 correlated variables and variables that have relationships 13 to each other in terms of how they af fect the output.
(
~-
) 14 We also have some very up-to-date techniques 15 that we're going to use here. I call this the group-force 16 method of doing sensitivity analysis, to prove to you that l
17 I do walk like one and I talk like one.
18 Anyway, what we're going to do with the 19 results of this analysis .is to, first of all, compare our 20 results with that of IPA Phase 2. It was mentioned this 21 morning that we'll be trying to run that data set through 22 this module as much -- in terms of what -- as much as we 23 can, because not all parameters are exactly the same in 24 terms of what we're calculating in Phase 2.
,y,
( j/ 25 But then we'll be using that result to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
178 1 identify the issues that we need to address as an agency ey 2 in terms of our regulatory needs, but also in terms of i I
%j' l
3 things that we want to raise up to DOE for their 4 consideration in terms of their TSPA-VA.
5 The results of this work will also be used to 6 prioritize our work in terms of identifying areas where we 7 need to do more study or more modeling. And then it will 8 help us to develop the acceptance criteria and for making 9 comments on the waste containment isolation strategy.
l 10 We'll be summarizing these, hopefully in the annual 11 report, and then we'll be using that analysis to l
l 12 develop -- start moving towards doing an importance !
13 analysis. ;
I ) l
\# 14 And I just want to touch on that briefly, 15 because I'll let Norm address this in more detail, but as 16 we move towards looking at which components of the system 17 are important and not just the parameters themselves, then 18 we'll be looking at which pieces of the system are more l l
19 impcrtant. This is that second level that was in that l 20 figure earlier. j 1
21 And so while the concept might be slightly l
! 22 different in terms of how we look at the importance versus l
l 23 the sensitivity to a particular parameter from an 24 analyst's point of view, in my shoes, it's really the same m
/ 1
(_,i 35 way in terms of how I go about implementing it. I turn l NEAL R. GROSS
! COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N.W.
[
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
179 1 flags on or off; I set systems on; or I just change
,- 2 certain tl' ngs within the model. So we're set up to go as k'~ ') 3 soon as we figure out what we want to do.
4 DR. HINZU: But you're going to be finished by 5 12/5/97? l 6 MS. COLTEN-BRADLEY: That's what they tell me.
1 7 DR. HINZE: Okay. And you haven't really l
8 started yet.
l 9 MS. COLTEN-BRADLEY: Oh, no.
10 MR. McCONNELL: Well, let me clarify.
11 MS. COLTEN-BRADLEY: Well, yes and no.
12 MR. McCONNELL: We are going to do the system- j 13 level sensitivity analysis -- this is Keith McConnell for l
(
\_-) 14 the staff -- by the end of December. We won't be doing 15 the importance - the, quote, importance analysis -- l 16 MS. COLTEN-BRADLEY: Right. i 17 MR. McCONNELL: -- that's more of a PRA 18 technique until sometime next year. That's 1c, I t h '.nk .
19 MS. COLTEN-BRADLEY: Are there any questions?
20 CHAIRMAN GARRICK: Sure. Mike?
21 DR. STEINDLER: This committee spent a 22 significant amount of time worrying about a couple of 23 processes and eventually -- you may recall that.
24 MS. COLTEN-BRADLEY: I knew it would come back l
~x l O) 25 to haunt me.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344 433 WASHINGTON, D C. 20005-3701 (202) 234-4433 M
180 1 DR. STEINDLER: Yes. Let me turn you to one
,_s 2 of your graphs in which you look at the varied input was i \
3 the infiltration of the function at climate. T don't much 4 care whether it's climate or whatever, but it's 5 infiltration. And then your output was the number of 6 containers that get wet, and it seems to me that's almost 7 a non sequitur, because the question really is the numbers 8 of containers that get wet with what.
9 And you've indicated that also the output was 10 a reflux value of some sort. That's not important either 11 unless you, in fact, take that one step further and ask, 12 what does the reflux do to the chemical composition of the 13 liquid that finally impinges on the containers.
/
k -) m 14 I guess what I'm asking is: In the absence of 15 some either estimation of process or coupling parameters 16 or analytical process for coupling of parameters, how can 17 you make some kind of claim on the validity of the 18 sensitivity analysis, unless you carry it all the way 19 through to that extent?
20 What I guess I'm driving at is the only thing 21 you point out, for example, is that the number of 22 containers that get wet are increased with the 23 infiltration rate. That per se, aside from being perhaps 24 obvious, is not a particularly useful -- thinking of the n
I v) 25 dose, again, as the primary output -- is not a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l
1323 RHODE ISLAND AVE., N.W.
l (202) 234 4433 WASH!NGTON, D C. 20005-3701 (202) 234-4433
l 181 I
l 1 particularly useful conclusion.
l l ,s 2 And in some cases, not picking these two but l ('
i \ '/
3 l
3 picking other things, they may, in fact, be incorrort. in
! 4 the sense that they focus your attention on what you think 5 is a high sensitivity of one parameter against the output, 6 but it's not necessarily true.
7 MS. COLTEN-BRADLEY: So what are you asking?
8 DR. STEINDLER: Well, what I'm asking is: How i
9 can you derive useful sensitivity, i.e., priority, l 10 information -- you indicated that your sensitivity l
11 analyses were, in fact, going to govern your program.
l 12 They can and obviously should, but only if you are able to l
13 couple the parameters to the things that are clearly l K. 14 important to the final output.
15 MS. COLTEN-BRADLEY: Well, and that's why you 16 don't just do the process-level sensitivity. We also do 17 the full system sensitivity, because if you only look at 18 it at the process level, then you would be focused on 19 these things that you've sort of taken out of context, 20 such as the number of containers that get wet, without 21 looking to the dose.
22 DR. STEINDLER: You intend to resolve this 23 problem at the system level rather than the process level.
l 24 MS. COLTEN-BRADLEY: Right. But by doing it
( ,) 25 at the process level, you g_, the teams a chance to get NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 182 l
1 their hands on the code and have some experience with it,
,e 2 without having to understand the intricacies of the whole i i
'~'
3 thing.
4 DR. STEINDLER: It's that latter comment that j 5 bothers me. I think understanding of the intricacies are 6 precisely the thing that validates the quality of the 7 output.
8 MS. COLTEN-BRADLEY: Right. But they know the 9 intricacies of their own particular module. They don't 10 necessarily know the intricacies of how the dose is )
1 1
11 calculated. Say, you're looking at the hydrologic group, 12 and so by -- it gives them a chance to think about the 1
13 intricacies of that particular module as well as whether l
/'~'T i
'w..) 14 or not those parameters that they're choosing to sample or i 15 to use make any difference or make any sense. l l
16 And as Bob mentioned this morning, it gives us 1 1
17 a chance to break down those -- to shake down the code in 18 a different context. But it's not -- if you do it out of 19 context, without doing the full sensitivity analysis for 20 the whole system, then it really -- then you can lose that 1
21 through the cracks in terms of not really understanding 22 that coupling piece of it, which is why some parameters 23 that may not appear to be very important at the subsystem 1
- 24 level would still be left as variable for the whole system
! / ~ s.
(v) 25 sensitivity, just because we think they are coupled or NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
183 1 they are important.
,m 2 Does that answer your question?
i \
V 3 DR. HINZE: Could I follow up on that?
4 CHAIRMAN GARRICK: Yes.
5 DR. HINZE: I -- as you well know and know 6 perhaps better than anyone else in the room, this 7 committee has been very concerned about coupled processes 8 being lost in the shuffle here and not being a KTI and 9 being concerned that it would be lost in the PA KTI.
10 And when I hear the discussion here, I don't 11 see a lot of this coupled process being brought in, and I 12 wonder what criteria you're using to make a selection of 13 how you decide what to do sensitivity studies on with (n)
14 relationship to coupled processes.
15 For example, there are the whole thermal 16 hydrologic chemical -- we've discussed this at length 17 personally, and I don't see these in here, and I don't see 18 the volcanic effects upon groundwater movement. Perhaps 19 there's justifiable reason for excluding these kinds of 20 things, but what criteria did you use to select which 21 coupled processes to investigate and which ones --
22 MS. COLTEN-BRADLEY: Well, actually those 23 kinds of things are decided upon when you decide what 24 you're going to model. And if you don't model it, you g-So if you don't
() 25 can't do a sensitivity analysis on it.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344 433 WASHINGTON, D C. 20005-3701 (202) 234-4433
I 184 1 have volcanism affectir.g groundwater flow, you can't do a
,_ 2 sensitivity analysis on it.
i ( ')
3 DR. HINZE: So how do you convince yourself 4 that it is important or not important to model it?
5 MS. COLTEN-BRADLEY: Well, that's done back at 6 the investigator level in terms of what they -- that's the 7 parts of the bias that goes into the process of the 8 modeling. Tim might be able to answer to add something to 9 that.
10 DR. HINZE: Is this getting at the point that 11 we are at the risk of losing coupled processes as a 12 concern here and -- l 1
13 MS. COLTEN-BRADLEY: Well, actually we have a
('~h :
(_ ,/ 14 number of things that address that. We have -- Jeff is 15 doing a lot of work on isothermal flow and the effects on 16 infiltration, so -- or into the repository, so I would --
17 in that regard, I would say that we haven't lost it 18 through the cracks.
19 Keith has --
20 MR. McCONNELL: If I could before Tim, Keith 21 McConnell for the staff --
22 I think we're actually moving closer towards a 23 recognition of the importance of coupled processes. When l
l 24 you look at the bottom tier of the diagram, in many cases, 1
And what we're trying to do is (v\ 25 they are cold processes.
HEAL R. GROSS COURT REPORTERS AND TRANSCR:BERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
185 1
! 1 make sure that we don't have stovepipe thinking, but those
,s 2 coupled processes and the inputs we need and the process-
,I \
\
'"/ 3 level models we need are coupled within the context of how l
l 4 we're going to do a performance asseesment.
5 DR. HINZE: I agree with you, and I was very 6 pleased to see this in these bottom boxes. But I was 7 also concerned about how those were selected and others 8 weren't.
9 MR. McCONNELL: Right.
10 DR. HINZE: What kind of criteria are being 11 used to make certain that we're not losing something here?
12 And I'm still concerned about the one-plus-one-plus-one 13 equals fifteen and not three.
n t
(_)
14 MS. COLTEN-BRADLEY: Tim, did you want to say 15 something?
16 MR. McCARTIN: Yes. I guess I would just add j l
17 in that there's certainly work being done, as Ginny <
18 indicated, in the KTIs to get a better sense of what to 19 include in the model. And, for example, the near-field is 20 looking at the thermal hydrologic chemical aspect of the l l
21 near-field environment. We have parameters in the model 22 to account for certain conditions.
23 MS. COLTEN-BRADLEY: Right.
l l l 24 MR. McCARTIN: Do we -- are those the right (n) 25 ones? They're doing the detailed modeling to suggest, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
I 186 l 1 well, can we get by with just a chloride concentration and
.3 2 a pH, but they're doing that detailed modeling to see how
'~'
3 variable things are, and so that's where the coupled 1
4 process work is going there, but we'll always hopefully l l
5 try and abstract a simpler model, just in the context of I 1
6 being able to run it in a reasonable amount of time. l 7 But we are fully aware of the fact that 8 sometimes that abstraction process breaks down, and maybe 9 you can't do it, but that kind of work is being done 10 within the KTIs. And thermal hydrologic / chemical one is 11 probably the best one in --
12 DR. HINZE: And most concerned one. l 13 MR. McCARTIN: Yes.
rh i I
\m / 14 DR. HINZE: So really these types of 15 sensitivity studies are being done on kind of a -- I may 16 not say that; I was going to say an ad hoc basis, but some 17 kind of organized activity within the KTI --
l 18 MR. McCARTIN: Right. And then that's really 19 the really nice integration aspect of the entire 20 performance aspect that we're doing. We have the code, 21 but they're intimately familiar with what we've included 22 in the code, and they're trying to do more detailed work 23 to suppcrt the bases for what we're doing and maybe l
24 identify areas where we need to improve it. And it's
, (7%)
us 25 all --
i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
187 1 That's the, I think, real beauty of the x, 2 diagram that Keith showed up in terms of what we call the 3 flow-down diagram, with that multi-box diagram that 4 everything's trying to pull together towards the same 5 goal.
6 MR. LESLIE: Dr. Hinze, this is Brett Leslie 7 from the NRC staff in Washington, D.C., and I'm the team 8 lead for the evolution of the near-field environment. And 9 I'm one of the PIs that is doing the sensitivity studies 10 within the TPA Phase 3.1 code, and frankly, I like it 11 because it allows me to use all the different modules to 12 evaluate what we think the effects of the coupled 13 processes will be.
( i 14 For instance, if we are concerned that the l l
l 15 interactions of thermal hydrology and chemistry on l l
16 concrete will affect all properties, I can go to the UZ !
17 flow and transport and change the porosity and 18 permeab'.lity to determine whether it has an impact on 19 performance, either by changing those values in a certain 20 horizon or varying the length that -- or length scale over 21 which those flow properties might change.
22 So the KTIs are involved in coming up with the 23 sensitivity within the codes.
l 24 DR. HINZE: Thank you. I think that's an
,/~'s
(_) 25 excellent illustration of the way it should be done.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
188 1 While I have the floor, can I ask another
,s 2 question?
/ \
\ l 3 CHAIRMAN GARRICK: Sure. I 4 DR. HINZE: I don't know whether it's !
5 appropriate here, but we've heard a very limited amount 6 about human intrusion. Bob Baca suggested that there will 1
7 be consideration of a stylized type of calculation. Are
)
i 8 there any sensitivity studies being done on the stylized l
9 calculations?
10 MS. COLTEN-BRADLEY: Tim's sticking his finger 11 up in the air. ;
12 MR. McCARTIN: Yes. As the rule KTI lead, for l I
13 now, we're sort of lying low in that there's enough flux l
,y
, I k_I 14 in terms of the legislation --
15 Let me rephrase that. Between the legislative 16 uncertainties and uncertainties with how the EPA standard, 17 for now we're not doing anything significant with respect 18 to human intrusion. However, based on our IPA Phase 2 19 work and subsequent work we did for the NAS standard, we 20 don't believe there's any great difficulty in implementing 21 a stylized calculation for human intrusion if required.
22 But for now, we're just holding back that it's i
23 doable. It's just, let's not go ahead and charge forward 24 and do something if maybe legislation says we don't even f% i
! 25 have to do it.
G l NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
l (202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
i i
1 189 1 DR. HINZE: But you can incorporate that and l
l ,.s 2 plan to in the TSPA 3.
/ T
\
\ 'l 3 MR. McCARTIN: At some future point, yes.
4 It's certainly possible.
l 5 DR. HINZE: Thank you very much.
6 CHAIRMAN GARRICK: Ginny, this is a question !
7 I'd much prefer to ask if I knew a lot more about how you I 1
1 8 did your sensitivity analysis and your uncertainty 9 analysis. But let me just share with you something that's 10 bothering me a little bit. It might be something we'll l
11 get on tomorrow, and that is the significance of !
12 sensitivity and uncertainty analysis in a calculation i 1
1 13 that's based on very conservative assumptions and the l k._,) 14 application of a conservative assumptions, not necessarily 15 evenly throughout the model.
i l
16 It concerns me that maybe it puts the focus on 17 results. I don't know. I have this vision of compounding l 18 of uncertainties and getting a distribution around each l
19 one of these parameters and these dose values, and sooner l 20 or later, you get out into a domain that is totally 21 unrealistic and not representative at all.
22 I can see great value in sensitivity analysis 23 and uncertainty analysis if you approach the model from 24 the point of v ew of a realistic model, because then you
/m
! can look at the tails and you can look at the
) 25
~ s' NEAL R. GROSS COURT REPORTERG AND TRANSCRIBERS
- 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON, D C 20005-3701 (202) 234-4433
190 l
1 distribution, and you can draw insights probably that
- s 2 are -- that have some physical meaning.
1
- t
( 'l 3 But I have less comfort and understanding of 4 the meaningfulness of distributions about parameters and 5 results, dose results, that are already extremely 6 conservative and how those results could be misinterpreted 7 and misused and what have you.
8 Have you got anything to make me sleep better 9 about this issue, anything to say?
10 MS, COLTEN-BRADLEY: I would agree with you on 11 that. There's -- Norm is smiling at me. I would agree 12 with you very much on that. The -- one advantage that 13 this kind of breakdown or shaking of the code allows you
,f S
(,)
14 to do is to figure out how unrealistic perhaps in a way, 15 certain ways, that you're modeling things are and then if 16 you want to step off, the unrealistic way of modeling and 17 do it more realistically, then you've got sort of a 18 benchmark to start from, or vice versa.
19 But the other thing it allows you to do is to 20 find out those pieces of the actual modeling exercise that 21 are really giving you very spurious or -- you know, 22 very -- results that you know just seem very, very much 23 out of whack and forces you to go in and look at all the l
24 pieces that went into deriving that one part tular result.
r~s
() 25 For example, we had a really interesting case NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C 20005-3701 (202) 234-4433
191 1 in IPA Phase 2, and I'm trying to remember all the details l
l ,- ~s 2 of it as I stand here, where for some reason the
. + )
~'
3 infiltration rates -- no. It was the groundwater travel 4 times changed very dramatically with just a very small 5 change in the infiltration rate, and maybe you can l
6 remember what that was, Tim. It was because things became 7 very horizontal. I can't --
8 But it was something that just was very 9 counter-intuitive to us, and so after Tim went back and 10 looked at it, we were able to figure out -- or he was 11 rather able to figure out what was making the system 12 behave that way, but at first cut, it just didn't make any l
13 sense to us at all in terms of what the output looked
/3
( \
't) 14 like.
15 CHAIRMAN GARRICK: It puts a lot of importance 16 to the documentation --
17 MS. COLTEN-BRADLEY: Yes, it does.
18 CHAIRMAN GARRICK: -- of this kind of 19 analysis.
20 MS. COLTEN-BRADLEY: Right.
21 CHAIRMAN GARRICK: So that people really know 22 what you're doing.
23 MR. McCONNELL: I would just, if I could --
24 Keith McConnell again -- clarify this, that we believe the (a) 25 input parameters and the processes we' re inclu. ling in the l NEAL R. GROSS l
COURT REPORTERS AND TRANSCRIBERS 13 3 RHODE !SMND AVE , N W.
(202) 234 4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
192 1 conceptual models we're including are reasonable, but we
,w 2 do have a tendency, particularly in the absence of data, i '8 3 to go on the cautious side or the conservative side of the 4 analysis.
5 And we think that's a prudent approach to 6 take, but we're not trying to make a prediction, but we're 7 just trying to develop a reasonable estimate of what the 8 performance might be.
9 CHAIRMAN GARRICK: Yes. But the 10 simplifications that you've introduced right at the 11 outset, such as one-dimension models and so forth, also 1
12 introduce conservatisms that could be rather pronounced. l 13 Okay. Any other questions for Ginny? I s l
/~'J
\ ~-
14 (No response.)
15 MR. McCONNELL: I think we're at a reasonable l l
16 break point if you're willing to take your break now and l 17 then we'll proceed with Norm's discussion of our future 18 activities.
19 CHAIRMAN GARRICK: Great. Thank you. We'll 20 do that. We'll take our break now for 15 minutes.
21 (Whereupon, a short recess was taken.)
22 CHAIRMAN GARRICK: Come to order now.
23 Norm, you're on.
24 MR. EISENBERG: All right. I'm going to talk
/s
(_) 25 about our plans for the review of the TSPA, and in doing i
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ; GLAND AVE.. N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
l
193 1 this, I --
.- 2 There's a lot of material that is in this set
'~# 3 of slides which those of you that were able to attend the 4 technical exchange have already heard, but I thought it 5 was important to outline what DOE is doing for the benefit l 6 of the committee, because it, in a sense, is -- we're 7 responding to what they're doing in our review.
8 So much of this material was taken from their 9 TSPA-VA plan, and they state the purposes of their TSPA-10 VA, and the first one is to serve as an input, of course, 11 to their decision on the viability of Yucca Mountain.
12 Second, they want to provide feedback 13 internally on site characterization and design to complete
,m
(_) 14 the license application if they decide to go ahead.
15 And they also give a nice statement of -- that 16 the fundamental purpose of the TSPA is to evaluate how 17 well the nuclear waste disposal system complies with given 38 eafety criteria. So that's the context of their TSPA-VA.
19 Now, the other side of that is the objectives 20 for the NRC review. You should recall that we have no 21 legislated manda.te to review the VA, so -- but we do have 22 a role vis-a-vis the VA, and certainly as a documentation
[
23 or snapshot of where DOE has come with their analysis of 24 the repository system, we should ;se it as an opportunity e ~x (u,- ) 25 to give them early feer >ack on au- identification and NEAQ. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N W.
(202) 234- # 33 WASHINGTON, D C. 20005-3701 (202) 234-4433
194 I
l 1 vulnerabilities in the case they're making.
1 p_ 2 Also, we do have a responsibility to consider 1
( )
3 the adequacy of the basis for their estimates of costs and 4 their scheduling, and so we would use it as an opportunity 5 to comment on that material also.
6 The specific objectives for the TSPA review as 7 opposed to the overall viability assessment: First of i
8 all, we anticipate that somebody somewhere is going to ask 9 us what we think of their TSPA-VA, so we're preparing to 10 answer the question when it comes, and we're almost sure i
11 that we're going to get it.
12 We would also like to use it to the optimum I 13 advantage to learn some lessons ourselves in assessing our !
~. l
( ) 1
(_j 14 license review capability; that is, the NRC's capability 15 to review ti>e license application, and in that sense, we 16 can use it to evaluate the TPA code, our computational 17 capability, our computer system, the technical teams that 18 we have, and in a sense, the overall management and 19 organizational structure, and whether we're really set up 20 to do this kind of integrated evaluation that also 21 involves quantification of performance.
22 And finally we expect to use the PA 23 information to assist in the review of their design and 24 site-characterization activities, especially pursuant to
,r m
( ) 25 the adequacy of those bases for estimating costs of
\~J NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON. D C. 20005-3701 (202) 2344411
! 195 l
l 1 schedules.
l j - 2 So I've put in here some key dates for the
! }
l 3 TSPA-VA. They've completed their abstraction workshops.
l 4 The end of August, they're going to freeze their process 5 models. End of September, they're going to freeze the 6 data set, and -- more or less, and -- well, we had some 7 discussion of that yesterday, and 1 think it's more or 8 less.
9 And they are going to complete the TSPA-VA 10 reference case at the end of January, and the complete 11 draft TSPA-VA, June 12. Now, we expect to get copies of 12 both of those intermediate products, as well as the final 13 TSPA-VA September 30 of '98.
f}/
(_s 14 Now, DOE has chosen to arrange their program
]
l 15 around several different structures, but one of the most I 16 central right now is this components of the waste I l
1 17 containment and isolation strategy. I don't know that !
l 18 I -- well, maybe I ought to go through just these five, 1
19 which are the primary features or attributes of the 20 repository they expect to rely on to demonstrate 21 compliance with the regulation and to demonstrate safe l 22 performance.
23 And they are the rate of water seepage into 24 the repository; waste-package lifetime; the rate of O. .
! (
L ,/
) 25 release of radionuclides from the packages; radionuclide NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, O C. 20006-3701 (2C2) 234-4433
196 1 transport through the engineered and natural barriers; and
,. 2 then dilution of the radionuclides in the saturated zone.
I i s
/
3 These are the primary components.
4 Dr. Steindler?
5 DR. STEINDLER: Yes. If DOE relies on the 6 waste package lifetime, are they counting heavily on the 7 10,000 years will remain in the EPA standard? It surely 8 is not the decay efficient product.
9 MR. EISENBERG: I'm not sure I understand --
10 try it again.
11 DR. STEINDLER: If they believe that the 12 waste-package lifetime is some kind of an important 13 attribute of this whole exercise, the only thing that I m
( )
(. / 14 can see is that they must be relying on a reasonably 15 close-in, namely 10,000-year, cut-off time for compliance.
16 Is that the sense that you have?
17 MR. EISENBERG: In a sense, although I think 18 they are moving towards -- and, you know, I can't say they 19 absolutely are doing this, but I think they are moving 20 towards a perception of waste packages where they have 21 substantially longer lifetimes.
22 Now, whether we think that could be justified 23 as another subject --
l 24 DR. STEINDLER: Longer than what, I guess is (x
) 25 my question.
I NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N W.
(202) 234 4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
- 197 l 1 MR. EISENBERG
- I would say fairly long l
7ey 2 lifetimes. They're talking about several engineering ii 3 additions. They're talking about a two-layer waste l
4 package. They're talking about the possibility of ceramic 5 coatings. They're talking about the possibility of drip 6 shields and other engineering features, so they're talking i 7 about taking measures to extend the life of the waste l 8 package.
9 And I guess I don't think we're sure and I'm 10 not sure that they're sure exactly how much credit they ;
11 intend to take at this time for a waste-package lifetime.
1 12 It could creep up to be longer than 10,000 years, I 13 suspect. I don't know if they are intending to do that.
/ h (s/ /
14 I don't know if they want to speak to that.
15 MR. BAILEY: This is Jack Bailey with the M&O.
16 We set a requirement in what we call our control design 17 assumptions document which says we're going to see what we 18 can do to mate it, and we've basically established that we 19 want to have a lifetime of the packages of 3,000 years 20 before the first through-wall penetration occurs.
21 And we have set a maximum of ten of the 22 packages being penetrated in that time frame, and we're 23 doing exactly what Norm said, and that is we're looking at 24 a number of features to try and extend that lifetime. And
) 25 the intent is to get it past the bulk of the thermal NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
1 198 1 pulse, if you will, to try and get the reaction rates for
,_s 2 the corrosion down to a range where you have a very long
(' ~') 3 lifetime left, because the corrosion occurs at such a slow 4 rate of the materials of choice at the moment.
5 We've done some math, and it's -- you know, 6 it's simplified types, but, yes. We show packages that 7 have lifetimes in excess of 10,000 years in some cases, 8 based upon being able to protect it from drips or to keep 9 it -- basically to keep the drips off of it and keep it in 10 a general corrosion region, and if you can steer your 11 temperature correctly, very long lifetime packages for the 12 materials.
13 And if we can go into the things like the ig)
(_,' 14 ceramics, the drip shields, that would be even lorip--
15 lifetimes. That's what we're trying to shoot at at this 16 point.
17 MR. EISENBERG: Okay. The next slide -- two 18 slides are a further elaboration of these aspects of the 19 repository, and under each are a set of hypotheses. There 20 are 12 hypotheses under the five elements of the waste-21 containment and isolation strategy. I don't know that it 22 pays to go through them. Maybe it's worthwhile to go 23 through the waste-package lifetime, since we were just 24 discussina it.
(j\
s x._-
25 They have three hypotheses: That the heat l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
199 1 produced will reduce the relative humidity in the vicinity j
,_s 2 of the waste package. Of course, that has implications I i
" ' 3 for wind corrosion will occur and the rate at which it 4 will occur.
5 The second is that corrosion rates are very j I
6 low at low relative humidity, and corrosion of the inner 1 l
l 7 barrier is slow, j 8 And, finally, the double-walled waste packages 9 will significantly increase containment times due to )
10 galvanic protection of the inner barrier by the outer 11 barrier.
12 Now, I have to add that these were taken from 13 the TSPA-VA plan which I think is September '96, and we
! \
\_ ,/ 14 were told during the technical exchange that DOE is l l
l 15 constantly revising and updating these hypotheses, so they !
16 might not be the same right now, and certainly in the ;
17 future they may change. But this is a snapshot of where 18 they were.
19 I should also add on the next slide that in 20 addition to the five elements of the waste containment and 21 isolation strategy, they've added a sixth category, which 22 is disruptive events, and they have three hypotheses about 23 the disruptive events. So this kind of covers the i
24 important issues, if you will, or the questions that will r~s
! ) 25 have to be suitably resolved in order to demonstrate
't./
l NEAL R. GROSS !
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433 l
200 1 suitable performance.
-..s. 2 So I think -- let me just add that I think the l \
\ J
'~~
3 approach that the NRC staff has taken is that, of course, l
4 we have to respond with the arguments -- to the arguments 5 that DO is making, and so their arguments are going to be 6 structured according to this outline.
7 We also have to see if there are imbedded in 8 what they're presented any issues or if they have omitted 9 any issues that we think may be important that they have 10 not considered, and we need to add those to our list of 11 what we need to focus on, so I guess I'm saying we focus 12 on the union of what they've articulated as important 13 issues and then any additional things that we think are i
n \
V 14 important.
15 Okay. The DOE has engaged, as we have, in an 16 abstraction process. They are moving from detailed 1
17 process models to abstracted models that will be used in I 18 their -- as submodels in their code which will be used to I I
l 19 estimate performance of the re.asitory. '
20 These are the ten process areas that they have 21 divided their activities into. Pretty much, I think, you 22 get a correspondence between these ten process areas and l
23 computer modules, although I believe some of these will l l
l 24 spawn several computer modules.
l -
() 25 So, again, I don't know that I need to go NEAL R. GROSS COURT PTTuRTERS AND TRANSCRIBERS j ts23 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
201 l
1 through each one of these. These are pretty much the
,s 2 suite of models that you'd expect, certainly that I would l \
\' /
3 expect to be elements of a total system PA for Yucca 4 Mountain.
5 Now, they've taken one additional step, which 6 is that they have prioritized the process models and also 7 correlated the process models they're considering with the 8 NRC KTIs, and that's shown in the next table.
9 And I guess I should say that the top things 10 they have are the site-scale unsaturated zone hydrology, 11 the drift-scale thermo-hydrology, the waste-package 12 degradation models, the site-scale unsaturated transport 13 models, and the site-scale saturated zone flow models. I r~ ~x l I }
k/ 14 Now, currently, I don't think the NRC staff 15 has a comment yet on whether we agree or disagree with !
16 this set of priorities. But we expect to accomplish that 17 in the next little while.
l 18 The next slide is an interesting one in that i l
19 it identifies areas in the TSPA-VA plan that these ten 20 process areas and where the DOE is developing abstractions 21 versus the NRC code and where we have developed submodels, 22 and as you can see, I think we've got the territory 23 covered. And in some cases --
l 24 And this is more, I would say, modeler's l
/'~'T
( ,) 25 choice. In some areas we have several nodules where they NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
202 1 may only have one, although their disruptive features, g3 2 events, and processes modela really will involve more than U 3 one module.
4 So I think we're looking very much at the same 5 aspects of the repository performance that DOE is in their 6 TSPA. So we expect that our TPA code will be a useful 7 tool in evaluating their estimates of performance.
8 We've been interacting with DOE rather 9 frequently to try to set a -- put us in a position where 10 we can give them early feedback on concerns we have about 11 their modeling approaches, any emissions we may have noted 12 from their modeling approaches, or concerns about certain 13 approaches they're taking that we may think are not either
,_. l
[t \
f l
x/ 14 appropriate or suitable. I hate to say " suitably l 1
1 15 conservative," because I know Dr. Garrick won't like it. I 16 In doing that, we have attended between the 17 NRC staff and the Center, we have attended the DOE 18 abstraction workshops and a sequence of expert I
19 elicitations. And I guess the side I don't have is that 20 they are engaged in three activities to help them 21 development their TSPA-VA code and evaluate it.
1 22 They are doing abstraction workshops in which 23 they get, I would say, inter-disciplinary groups of 24 specialists together to determine how to move from process
( ,) 25 models *.; the abstracted models to put in the TSPA code.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20N3701 (202) 234-4433
l 203 l
l 1 They also are assembling expert elicitations j 73 2 to try to define the significant uncertainties that would
/ t
() 3 be present in the modeling that they use in the TSPA code.
4 The third thing they're doing is they have a 5 performance assessment peer review panel which is 6 evaluating all the aspects of the TSPA development and 7 application as -- in a peer review mode, so they are 8 engaging in all those things, and we are trying to monitor 9 what goes on in all these different activities.
10 As the next bullet says, we've attended major 11 meetings of the performance assessment peer review panel.
I 12 We've reviewed their TSPA-VA plan, the waste containment i 13 and isolation strategy, and any other documents pertinent l es m-14 to the development and application of the TSPA code.
1 l
15 We're continuing our development of the TPA 16 code, which we fully intend to use to evaluate the results 17 of the TSPA-VA.
18 The continuing work on our issue resolution I
19 status reports is developing acceptance criteria as was 20 discussed this morning for various aspects of the modeling 21 and data handling, data interpretation.
22 And we've just had a technical exchange during 23 the last few days. We're anticipating another technical l
24 exchange in October. And I thought I heard a rumor that
/* h (d'
I 25 we would probably have one in February when we get the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
204 1 base case report and have had some time to look at it, to 2 discuss that.
J l 3 The general framework for the NRC review of 4 the TSPA-VA is, first of all, to focus on the key 5 performance assessment areas. We expect to define those 6 by using the independent, early runs of the TPA 3.1; we 7 expect to use the input from the KTIs and the KTI 8 subissues; and, of course, the information I just reviewed 9 on the waste isolation strategy and the hypotheses.
10 We expect to review both the abstractions that 11 DOE is using and their parameter choices. We intend to do 12 two things. We expect to use the TPA 3 code with 13 parameter sets that are similar or analogous to the ones r h. i
/ j l
(/ 14 that DOE is using, and then also run it with parameter 15 sets that we think are more appropriate, to see if that 16 provides any significant differences in estimated l l
1 17 performance.
18 For those areas where we do have significant, 19 identified differences, we expect to perform additional 20 modeling, which may be at the TPA level or it may be at a 21 more detailed level. It depends, of course, on what the 22 outcome of the initial exercise is.
23 And we intend to prcvide timely feedback to 24 DOE on their abstractions, parameter sets, the sufficiency n
(s- ) 25 of the supporting data and analyses, and, of course, the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433
205 1 credit taken for various engineering and site features.
,s 2 And then if you'll go not to what's in the
/ \
< )
3 package but this slide, I revised this based on the 4 technical exchange, and these are some technical areas 5 that we may focus on, based on differences that we have or l
6 cot erns that we have with their modeling approaches.
l l
7 And some of them are as follows: The matrix l l
8 diffusion, they indicated that they expect to perhaps rely 9 heavily on matrix diffusion, and we have questions about l 10 its applicability to both the unsaturated and saturated i
11 zone at Yucca Mountain, so that's an area that we are '
l 12 likely to focus on. l l
13 We would like to see a -- how they're going to KI 14 use their TSPA-VA to demonstrate an acceptance and use of 15 a multiple barrier approach. l We would like to see the 16 performances relied on -- the performance that is obtained 17 is relied on -- relies on several of the barriers and not 18 just one. And so we -- that's another area that we expect l
19 to focus on.
20 We had lots of questions on their support that 21 they had for the near-field modeling and the estimates of 22 the near-field environment. That's another area we expect 23 to focus on.
l l 24 We've found by their use of results of expert s
, ( ) 25 elicitations and the performance assessment, at least as NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 2344433
206 1 they've described it to date, that for the PVHA and the
, ,s 2 PSHA, there seemed to be questions about whether the i \
3 expert elicitations are gathering the data or the 4 information sets that are going to be easily used in the 5 performance assessment, or whether some additional 6 processing is going to be needed.
7 And we have guidance out, of course, that says 8 that before you conduct an expert elicitation, you should 9 take great pains to make sure that the scope and the 10 products are going to meet the end use. So we may focus 11 some more on that.
12 We had lots of discussions about the treatment 13 of variability and uncertainty in the consequence model, !
j
(')/
'N- 14 both how the terms are defined, how these variability and 15 uncertainty will be propagated in the models, and how 16 they'll be characterized at the end. So I suspect we'll !
l 17 focus on that also. l 18 DOE has modified the definition of waste-19 package failure. It used to be as soon as there was a 20 penetration that was it. Now it is expected to have some 21 integrity that lies beyond that point, and that's going to 22 be modeled, so there's now questions about how much 23 performance can be expected from one of these degraded 24 waste packages.
,a 25 Another issue that we have raised in the past
()
NEAL R. GROSS COURT REPORTER!, AND TRANSCRIBERS 1323 RHODE IGLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234 4433
t l 207 1 and which is an isnportant factor in calculating the 2
l7S l \ /
performance is the mixing that takes place as the 3 unsaturated zone flow enters the saturated zone. It still 4 appears that DOE is expecting mixing over a rather deep 5 interval, and we're not sure that that's appropriate, so 6 we'll be looking at that.
7 There are lots of -- always there are lots of 8 questions about the thermohydrologic modeling in the near-9 field. I suspect we'll continue to focus on that.
10 I'm sure we're going to have lots of questions 11 about the credit that can be taken for these engineering 12 innovations which are very interesting, but if I can just 13 take the drip shield as an example, it could prevent drips I,h
'N I 14 from hitting the waste package and providing a source for 15 corrosion with water. It would be corrosion perhaps in a 16 humid atmosphere.
17 But we're not sure how long a drip shield 18 could be relied on to last. We're not sure how it might 19 perform. So we have lots of questions about some of these 20 engineering innovations.
21 And another item cf considerable interest to 22 the staff is how water gets into the drift and how it's 23 being modeled.
24 So that's where we stand in our plans for
! /
(_,T/ 25 evaluating the TSPA-VA.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
i 208 1 MEMBER POMEROY: Norm, I've got several 7- 2 questions. I don't want to spend a lot of time talking to
( )
3 you about DOE, these plans, but I'm going to spend a 4 little time doing it, because perhaps I don't understand 5 quite what the table on page 9 does.
6 It establishes priorities somehow for TSPA-VA 7 in terms of -- are we talking about testing the process 8 model, abstracting the process model? I'm just unclear 9 about what that addresses first. Maybe you can clarify 10 that for me.
11 MR. EISENBERG: As I recall -- because I think 12 I looked this up last week -- DOE did not, in their 13 TSPA-VA plan, give a clear definition of what they meant
,rx
\
(_/ 14 by priority. I take it to mean that this is the 15 anticipated -- I hate to use the word "importance," but I 16 will -- impertance that they expect this particular 17 process model to hsve on total-system performance, and l 18 therefore, if you are --
19 One way to look at it: If you're living with 20 a reduced budget the way some of us are and you have to 21 say, Well, we can't do it all; what are we going to focus 22 on, you would focus on the things that you think are going i l
23 to have the biggest effect on total-system performance, t
24 and I think that's what this process -- I mean, what these 2%
(w JI 25 priorities mean.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
209 l l
1 1 don't know if there's anybody here from DOE s 2 that wants to get me out of this jam or not. Where did
/ T l iV )
3 Jack go? 4 1
4 MEMBER POMEROY: Well, I guess there's not,
, 5 and the abstraction and testing process is going on now, 1 .
l 6 as I understand it. Is that -- '
7 MR. EISENBERG: The abstraction process is 8 over.
9 MEMBER POMEROY: Yes. Of the model itself.
10 Right?
11 MR. EISENBERG: The abstraction testing is 12 ongoing right now.
13 MEMBER POMEROY: And that will be completed rN '
) l i._) 14 by -- at least prior to June 12 of 1998. l 15 MR. EISENBERG: Oh, I'm sorry?
16 MEMBER POMEROY: I'm asking -- there's some 17 completion date for the TSPA-VA document here that you've 13 given us, June 12, 1998.
19 MR. EISENBERG: Oh, yes.
20 MEMBER POMEROY: So that implies that all that 21 testing will be done by that time.
22 MR. EISENBERG: The testing will -- they have l
23 overlapping activities. They have a olide -- I think it 24 was yesterday -- that has these activities and their
/~m 25 bands, and so -- it for sure is going to end by the time
( )
NEUAL R. GIUDSS I COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 210 1 they issue or write --
1 2 MEMBER POMEROY: That's what I'm assuming, but i
/
3 I'm wondering how they can do that much work in that 4 amount of time. I don't ask you that question. Obviously 1
5 I'll ask them that question at some point in time. '
i 6 But the questions I'd really like to direct to 7 you are -- relate to the question that John brought up l
8 earlier today. What are the soft spots that you see 9 within the NRC program? )
10 MR. EISENBERG: Soft spots in the NRC program? l 11 MEMBER POMEROY: Yes, sir.
12 MR. EISENBERG: What do you mean by soft 13 spots?
-) 14 MEMBER POMERGY: Are there any points that you 15 can see where a lack of resources, a lack of computing 16 facilities, a lack of people, will result in NRC not being 17 able to be there when it needs to be?
18 MR. EISENBERG: Well, you know, I guess I 1
19 don't see it as a -- this may get into the discussion 20 tomorrow. I don't think it's a binary mode that we either 21 succeed or we don't.
i 22 MEMBER POMEROY: Of course.
I
- 23 MR. EISENBERG
- I think we'll either do better 24 or less well, and I think that the limitations on i r '\
(u) 25 resources is a continuing pressure that makes it harder l
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
f l 211 i
1 for us to do as thorough a job, as complete a job, and to l ,
cy 2 delve into all the details that we'd like to. I just
( l l
'~
3 think that there's no way around that. I 4 MEMBER POMEROY: You don't see identifiable j l
1 5 points. Mike, of course, mentioned the hardware question, 6 and we're cognizant of that. Are there identifiable 7 points in the science programs you can -- )
l 8 MR. EISENBERG: I don't think that I'd want to '
9 try to single out one particular area or another. )
10 MEMBER POMEROY: Suppose you were asked to 11 do -- at this point in time, you're doing extremely well 12 in my personal estimation. Suppose you were asked to 13 carry out some other specific tasks, major tasks, in the ;
/~N, :
i ! l
\_ ' 14 next two years. Would you --
15 MR. EISENBERG: In the performance assessment 16 area?
l 17 MEMBER POMEROY: Yes, sir.
18 MR. EISENBERG: We l . , I think we're thin as it 19 is, and if we were asked to do other accivities, it would 20 put additional pressure, and I think that the quality of 21 the work would suffer and the definitiveness with which we 22 are able to complete it would suffer.
23 MEMBER POMEROY: Okay. Thank you.
24 DR. STEINDLER: If somebody had available for I! p) 25 you an extra $400,000, what would you spend it on?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISt.AND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
i 212 1 MR. EISENBERG: A new house, 2 (General laughter.)
, w}
3 DR. STEINDLER: Let me rephrase the question.
4 In the absence of a new house --
5 MR. EISENBERG: What would I spend it on?
6 DR. STEINDLER: Well, let me put it 7 differently. What would -- what do you believe that the 8 NRC --
9 MR. EISENBERG: Mike wants to answer this one.
10 MEMBER POMEROY: Mike's going to come to the 11 rescue.
12 DR. STEINDLER: Well, if the answer's 13 hardware, Mike can sit down. He's already said that.
I 't
\> 14 MR. BELL: I mean, basically I think you're 15 asking Norm and you were asking Tim earlier what's 16 essentially a management decision, how you're going to --
17 DR. STEINDLER: No, I'm not. I'm trying very 18 hard not to do that, Mike.
19 MR. BELL: -- allocate resources.
20 DR. STEINDLER: The question is: Where's the 21 soft spot in the technical program?
22 MR. BELL: Well, I mean, we're all aware that, 23 you know, a year ago, as the results of the reduced budget 24 level, we cut out work at the Center in three KTIs. That (p _,/ 25 was based on the priorities as we understood them at the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344 433 WASHINGTON, D C. 20005-3701 (202) 234-4433
I 213 l 1 time.
l -, 2 And what we've learned during this year is as i/ \
3 soon as we cut out container life and source term and 4 radionuclide transport, issues in those areas started 5 coming up all through the DOE program. I mean, I think 6 one of our first priorities in fiscal year '98 is going to 7 have to be, you know, allocating resources in some of 8 those areas.
9 DR. STEINDLER: What are you telling me? Was 10 that a bad management decision, you know, that you didn't 11 do very well?
12 MR. BELL: It was at least a bad tactical 13 decision. I mean, essentially we had ten creas we were q(._/ 14 working in. We were faced with the choice of, well, do we 15 take this cut by whittling away all ten a little bit and, 16 you know, just do a little less in all ten areas, or do we 17 prioritize our work based on the way we understood the 18 priorities at that time and go full speed ahead on the 19 highest priority things and not work on what we considered 20 the lowest priority things, and we took --
21 DR. STEINDLER: So presumably Norm's answer 22 should be, Well, we need to increase the work on the 23 engineered barrier system. Is that what you're saying 24 then?
(m ; 25 MR. BELL: That's right.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
214 1
1 DR. STEINDLER: That's all I'm asking.
t 7- 2 MR. BELL: That's what we will do. I mean, i i
\
'~' /
3 that's what I anticipate happening in FY '98. I mean, 4 even if -- I'm not sure people generally know what the 5 budget situation is.
6 The Senate has essentially approved our full 7 request of 17 million. The House has given us a 13 8 million mark, which is 2 million more than we've had for 9 the last two years, but it's certainly not our full 10 request. So we should be able to do more than we've been 11 doing, and I think, you know, some of that increase would 12 go in those areas that we zerced out this fiscal year.
13 MR. McCONNELL: I would mention that this
/ \
ss 14 fiscal year and last fiscal year, we did have the 15 carryover situation.
16 MR. BELL: Yes.
17 MR. McCONNELL: So we were in perhaps a better 18 situation this fiscal year than we would be next fiscal 19 year if we don't get our request.
20 MR. BELL: And basically in deciding how we 21 would allocate resources in fiscal '98, we would intend to 22 use the results of sensitivity analyses that we'll have l 23 available to us early in the fiscal year, and I can't 24 imagine that some of the same things that are priority 1, T'S (v) 25 like total-system performance and unsaturated flow and l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
l l
l 215 l 1 thermal effects on flow, aren't still going to be among 1
s 2 the highest priority things, but some of the intermediate !
/ ) !
3 and lower priority areas may change, and I think we'll l
4 probably end up working on all ten areas. !
5 MEMBER POMEROY: As long as we've got both of 6 you up here, let me ask you another question, both of you.
7 Do you think the distribution of capabilities between the 8 Center and the NRC is appropriate at this point in time?
9 MR. BELL: Well, I mean --
l 10 MEMBER POMEROY: In the PA program now.
11 MR. BELL: Oh, in the PA program.
12 MEMBER POMEROY: Yes.
13 MR. BELL: You know, given the resource level
,- m i $
\ms/ 14 we're at, I think the mix is about right. I mean, it's ,
l 15 about equal, and, you know, if -- you know, when we do get .
l 16 an increase in resources, I don't see us pouring it all l
17 into the NRC staff and not --
18 MEMBER POMEROY: Yes. I guess I wasn't 19 thinking so much in terms of FTEs or something like that, 20 but I was thinking of the capabilities, the core science 1
21 capabilities.
22 MR. BELL: Well, I'm not sure you're aware.
I 23 The Center, as a result -- for some reason, some of which !
1 24 may be attributed to the bad budget situation for the last r
( ,\) 25 two years -- lost a couple of key people in the l NEAL R. GROSS
- COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W. ,
l I
, (202) 234-4433 WASHINGTON. D C. 20005 3701 (202) 2344433 l
216 1 performance assessment area.
f\ 2 MEMBER POMEROY: I'm aware of that. What I'm
\
3 asking is just this question.
4 MR. BELL: But I think the people they have 5 working on it now are doing the job we need.
6 Keith, do you want to --
7 MR. McCONNELL: No. I agree. I think that 8 the -- in both cases, that the distribution is good and 9 that we do have the people and the capabilities we need at 10 this time.
11 There are areas of weakness or soft spots, 12 and -- like container, like in source term, but should we 13 get increased budget, I think that the Center has
( _) 14 maintained the capability in that area, and that we could 15 build up in that fairly quickly.
16 So under the conditions that we have to work 17 on, I think we have a pretty good split and a pretty good 18 capability. We just have to do better with less.
19 MEMBER POMEROY: I'm sure the Commission feels 20 that you should do that. Of course, there's a point where 21 you can't do better with less. You don't do anything.
22 CHAIRMAN GARRICK: Any further questions? Go 23 ahead.
24 VICE CHAIRMAN HORNBERGER: One question for
( ) 25 Norm. On your very last slide, Norm, your item number 2 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
217 1 for focus said, " Demonstration of multiple barrier
,s 2 approach." Could you expand just a little bit on that?
i'~' )
3 What does this mean? Are you -- do you have some 4 quantitative notions in mind that you have to --
5 MR. EISENBERG: Actually, Budhi Sagar and I 6 have been working on some methods to try to quantify that, 7 but that's maybe a separate issue that I thought we would 8 get a little bit into tomorrow.
9 But in any event, the Waste Policy Act of '82 10 and our regulations require that DOE demonstrate a 11 multiple barrier approach. The current Part 60 has 12 specific requirements for the subsystems, but beyond that, 13 there is a requirement to demonstrate a multiple barrier
( )
'K '
14 approach.
15 We would expect them to speak to this 16 somewhere in either the viability assessment or in the 17 TSPA-VA and try to estimate how much reliance they're 18 placing on various pieces of their system.
19 Now, there are lots of ways to do it, but we 20 would like to see some discussion of that aspect. Is --
21 do you want me to say more?
l 22 VICE CHAIRMAN HORNBERGER: No. I mean, 23 it's -- your presentation is in the sense that your use of 24 TPA to evaluate in some sense, comment on upon TSPA-VA,
[ T
( ,) 25 and in concentrating on this, I was just curious what you NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
! 218 l
1 had in mind. Is this testing TSPA-VA for subsystem
, - 2 requirements?
\
\.
)
3 MR. EISENBERG: No. It would be perhaps a 4 situation where if they say, Look, we've got an average 5 lifetime for waste packages of 15,000 years and for 10,000 6 year standard, we're done -- we would like to see how much 7 redundancy is added by the behavior of the natural system 8 and how much additional performance gain they could get by 9 employing the natural system.
10 As I say, there's lots of ways to do it. You 11 know, you can think of, Well, how would the system perform 12 with the natural system there and with it removed, and 13 that woulo give you some measure of the impact of that
,o t !
(. / 14 natural system.
15 So -- and it need not be big pieces like that.
16 It could be the saturated zone, the unsaturated zone, the 17 limitation on the infiltration into the repository by the 18 unsaturated zone above the repository. There are all 19 kinds of elements that it could be partitioned into, but 20 we would expect to see some reliance on more than one 21 element of the system.
22 DR. HINZE: If I may, a quick point of 23 clarification on your slide 12, the second bullet: I l
24 assume that what you're getting at there is trying to
(~N
( ) 25 evaluate the TSPA-VA in terms of both processes modeling l
x-r NEAL R. GROSS COURT REFORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
219 1 and the parameters being used. And I'm wondering here.
,s 2 It was my impression that the NRC was having a somewhat
/ \
3 simplified version of the models being used in the TSPA-VA 4 by the DOE.
5 Can you, in that first dash there, can you 6 really develop parameters that are similar to DGE's, 7 considering this simplification? Is this a concern? Or 8 how are you going about that?
9 MR. EISENBERG: I'll tell you what I think 10 this means. It means that if they are going to have 24 11 pieces in their repository, that they model individually 12 with individual order infiltration rates and chemical 13 reactions, and we only have six, we would obviously have (O _-
> 14 to average some of those conditions to apply to our 15 partitioning.
I l
16 We just wouldn't have as many partitions, so 17 we couldn't use necessarily their same data set, because 18 it would be either on a finer discretization or possibly 19 they would be concerned with different behaviors. They
)
20 might try to separate packages that got real wet from ones l l
l 21 that were not wet at all, whereas we might consider just 22 an average condition.
23 So we would have to take their parameters and 24 modify them for suitable use in our model.
fm Is my assumption correct that the
- 25 DR. HINZE
N.-)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
220 1 first part of that is to look at their process modeling
, 2 and che second part is then to evaluate the parameters?
('~' )
3 MR. EISENBERG: I think we could look at the 4 abstracted model and see what they mean by their 5 parameters and see how that compares to what we mean by 6 our parameters and our abstracted models. I think we're 7 talking about doing it at that level, not at the next tier 8 down.
9 DR. HINZE: Okay. Thank you.
10 CHAIRMAN GARRICK: Any other questions for 11 Norm?
12 (No response.)
13 CHAIRMAN GARRICK: All right. I guess we're n
I
'xs) 14 ready for a wrap-up here, a summary.
15 MR. McCONNELL: Now that you've asked all the 16 hard questions to Norm and Mike Bell, I can summarize.
17 I've handed out two copies or two sets of papers. One is !
l l
18 our summary slide, and another is, I think, in response to 19 Dr. Garrick's request for perhaps a little bit more 20 clarity in the schedule for the TPA-related activities.
l 21 And I don't intend to talk directly to the table.
22 Basically what we've attempted to do today is l 23 to outline the approach that the division of waste 24 management performance assessment program is taking and r~x
.l ) 25 demonstrate that the program is integrated across key l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
WASHINGTON, D.C. 20005-3701 I (202) 234-4433 (202) 234-4433
221 1 technical issues, basically that we have an integrated 7s 2 program.
i ,I s
3 Second, assuming we get our FY '98 budget 4 request, we've demonstrated that we have the capability, ;
5 both in human resourceu and computational tools, to 6 evaluate the importance, the relative importance of NRC's 7 issues and subissues, and also review DOE's TSPAs. i 8 Now, let me Just say that some things that I l
9 keep me awake at nights is a scenario that I could 10 construct where we could be tasked to continue development 11 of the TPA code to keep up with DOE's, I think, increasing 12 emphasis on the engineered barrier system, but also tasked i
13 to do a major rule-making, develop a standard review plan.
's / 14 review a TSPA-VA type document, and review an EPA i l
15 standard, all in a relatively short period of time.
16 I think if that scenario, how unlikely it 17 might be, should occur, I think there would be some things 18 that get seriously delayed or don't get done at all, 19 and -- at least in the short term.
20 So there are things coming up on the horizon 21 that have a pretty large resource implication to it, and 22 it just depends on how the timing comes out as to how we 23 will address them., I think, and I think that as a section 24 chief, I think we've talked to our management about it.
?x 25 They know the problem, and they know the situation. I
(%s)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344 433 l
l
222 l l
1 assume if that scenario should occur, we'd have to go back
,y 2 and rearrange the deck chairs.
t :
)
i
'\', l 3 And, finally, the last goal of the 4 presentation today was to inform ACNW of the schedule and 5 timing of DWM's performance assessment activities. We )
6 tried to do that in the larger scale table that describes l
l 7 all of our activities in the high-level waste program, and 8 also in this more recent table that I've handed out, that 9 talks to the specific TPA activities that we anticipate.
1 10 And that's really the end of our presentation. '
11 CHAIRMAN GARRICK: All right. I think some of i
12 the committee members may have some general observations j 1
13 to make about the proceedings today. Maybe it would be
' / )
! -- 14 just as well for you to stand by to be the traffic l l
l l 15 director for those questions. !
l 16 Let me start with yc4u, Marty. j I
17 DR. STEINDLER: Okay. There exists a thing l i
18 called the Frye Rule, which you may have heard about.
19 It's a standard for the admissibility into evidence of l 20 computer programs. Let me just give you a short synopsis l
l 21 and make a comment on it.
\
\
22 The statement is generally made that judicial i 23 acceptance of a scientific theory or instrument that l
24 includes models, includes some computer-generated models
/~'N !
i
(_,1 25 and simulations, can only occur when it follows general NILAL R. GIUDSS -
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
i 1
1 223 i l
1 acceptance by the community of scientists involved.
2 It seems to me that since you're likely to
/-s\i t
I
\
3 eventually get to adjudicatory business, that that forces j 4 you all to think about -- and I guess I'm asking whether 5 you have -- a couple of things: One, how well have you 6 publicized, i.e., published in peer reviews?
7 And, two, can you defend the general notion 8 that a model of a phenomenon run by the NRC staff that 9 looks different than a model for the same phenomenon run 10 by DOE can legitimately be accepted on equal terms by 11 whatever internal, you know, scientific community in this 12 kind of business exists?
13 And, finally, shouldn't, under these
/'T (s-) 14 conditions -- and perhaps you do, and I guesa I'm asking 15 is whether the staff ought to consider seriously, once you 16 get reasonable confidence in your modeling processes --
17 shouldn't you be thinking about rule-making to avoid the 18 endless, potentially endless arguments that you could 19 engender once the license application process begins?
20 MR. McCONNELL: Oh, I'm not certainly !
21 qualified to talk about the kind of quasi-legal aspects of 22 codes and how they would be introduced within a hearing.
23 What I can tell you is that we do intend to l
24 publicize, both in the peer-reviewed literature and as 1
/"N l
( ) 25 documents in public document room and also in interchanges NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISMND AVE., N W.
(202) 2344 433 WASHINGTON, D C. 20005 3701 (202) 2344433 I
224 1 like this and other venues, publicize our code, put it on
,_s 2 the table, have people evaluate all aspects of that code,
! \
\'-) 3 comment on it.
4 We would then take those comments back and 5 make the appropriate, I think, changes, both to the 6 conceptual basis, as well as the coding itself.
7 As far as rule-making, one of the goals that 8 the division management has set for the strategy that 9 we're developing is that we do make the approach to 10 demonstrating compliance as clear, simplistic, and logical 11 as we can. Now, whether that involves speaking to the 12 issue of using specific codes or models, I can't say right 13 now, but certainly we can take that back and think about
/"'N b 14 it.
15 And when we come talk to you about the 16 strategy, which we will in the fall, I believe, or are 17 planning on, maybe we should readdress that issue at that 18 time.
19 DR. STEINDLER: Well, DOE has a similar 20 problem, of course, but at the moment, I'm more interested 21 in how you can handle it.
22 MR. McCONNELL: Right.
23 CHAIRMAN GARRICK: You had a KTI comment, I 24 believe.
rm
( DR. STEINDLER: Oh, yes. Let's see. I've
) 25 i
- NEAL R. GROSS
! COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE,, N.W.
l
! (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
225 I forgotten what it was. Yes.
-s 2 If -- it follows along something that John
! \
l
("/ 3 mentioned earlier.
1 You're currently focused on the KTIs ;
4 as they currently exist. How do you know that list is 5 comprehensive enough?
6 MR. McCONNELL: Well, I think that, as I l
7 mentioned this morning, that list was derived through, I ;
8 guess, kind of a quasi-expert elicitation, using 9 performance-informed judgments. I think we're finding now !
1 l
10 that it -- based on our work with our own code development 1
11 work and our interactions with DOE, that perhaps, j l
I 12 particularly in the subissue area, not necessarily the l
13 overall issue, but in the subissue area, there are some G
l b 14 areas where we need to shift ecphasis.
15 The one that's been brought up here today is 1
16 the waste-package area. There are aspects of waste-17 package performance that we haven't looked at and that we 18 need to go back to the drawing table, I think, and think 19 about a little bit more. So I think --
l 20 The idea of KTIs is that they're not static, i
21 that we're going to use the sensitivity analyses and re-l l 22 evaluate those on a yearly basis, I think, as Margaret 23 indicated; that they're not set in stone. We go back 24 every year and use the knowledge gained and reprioritize n)
( 25 and re-evaluate what are the key issues.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
226 l
1 And I think hopefully you'll see the results ;
I s 2 of that in these IRSRs, issue resolution status reports, )
I \
~'
3 that are intended to come out this fall and early next i i
l 4 year. You'll see some of that transition.
5 DR. STEINDLER: I guess what I'm asking is a 6 little bit oblique from that.
7 MR. McCONNELL: Oh , I'm sorry.
8 DR. STEINLLER: The scope of your activities 9 at the moment are limited to the kind of phenomenon 10 normally covered under your KTIs. What I guess I would i
11 ask, if you're calling this thing a total-system 12 performance assessment, how do you know that that's a 13 complete set?
/y
( I
\_/ 14 MR. McCONNELL: Well, I think it's based on 15 our past IPA work. You're correct. It's not total. It 16 doesn't include extreme erosion or things like that.
17 DR. STEINDLER: For example.
18 MR. McCONNELL: But what it is based on is we 19 believe we have captured those factors that are most 20 important to the dose calculation.
21 DR. STEINDLER: And that judgment is the 22 expert solicitation, your general experience, et cetera, 23 et cetera, that you can't think of anything that is a l
24 show-stopper that you haven't already considered. Is that i rx
(,) 25 basically the point?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
227 1 MR. McCONNELL: That's correct. And also we l
gS 2 will evaluate that when we do our sensitivity studies.
i )
Q/
3 Have we really captured that? And that's why this 4 emphasis on waste package might change that 5 interpretation. And maybe I should -- I don't know that 6 I'm answering your question.
7 DR. STEINDLER: No. That's fine.
8 MR. McCONNELL: Norm, did you want --
9 DR. STEINDLER: I just wanted to raise the 10 point.
11 MR. EISENBERG: I just wanted to say something 12 about the statement about the use of the computer codes.
13 My understanding of the situation is that the staff is
, /~'s
' (\ -) 14 going to use our computing capability to probe the DOE 15 analysis. We are not in competition with DOE. We don't 16 have tc, unke a case for safety; they have to.
1 l
17 Co I don't know that the same demands on l
18 acceptance have to apply to us that they do to DOE. Of )
19 course we want to have something that's reasonable, and of 20 course, we want to have something that's accepted. But I l l l 21 don't think we have as high a standard as DOE, because i l l
! 22 we're using it in a different way. We're not using it to 23 make the case for safet y. We're using it to probe their 24 case and to make sure they haven't left anything out or, A
( ,) 25 you know, misinterpreted something.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
228 1 And to do that, I don't think we need the same
/,,\
2 degree of peer review and external acceptance that DOE is t )
~# 3 going to have to have. In fact, our use of our code, I 4 think, is part of the process to give that to DOE.
5 DR. STEINDLER: Okay. I don't agree with your 6 interpretation, but there's no point in --
7 MR. McCONNELL: We actually have started more 8 a peer-review type of approach in that we have provided to 9 DOE at their request the ash plume module. And I'm sure 10 that they're running it and evaluating it, and I assume if 11 they have comments, they'll get back to us on that. And ;
1 12 I think that we would do that with all of our codes, all 13 of our modules.
( )
(s./ 14 Wes? ,
l 15 MR. PATRICK: Keith, if I could just add one 16 piece that may fill in a little bit of Dr. Steindler's i 17 question -- this is Wec Patrick at the Center. Predating 18 probably most everybody in the room -- you may remember --
19 we did some very comprehensive systems analysis, analyses, 20 on behalf of the NRC.
21 One set of those analyses examined the 22 regulation as it is currently formulated. Another 23 independent set looked at the physical systems. There was l
l 24 a so-called -- if I can remember the terms right -- a 25 repository operational criteria evaluation that was done, I
k-)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE.. N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
229 1 looking at pre-closure issues.
_ 2 There was also a repository isolation criteria
! )
3 study that was done, looking at post-closure issues, and 4 each of those two later studies looked at structure 5 systems, components, processes, and parameters that were 6 very broad in scope. And coming out of those were a set 7 of -- I don't recall now -- 56, 60, some such number of 8 key technical uncertainties.
9 And the ten key technical issues we now have 10 are a very small subset and in some cases a compendium 11 where there were like uncertainties, they were drawn 12 together, collected, and then we came up with these ten.
13 So a very broad net was cast initially, and n
k_) 14 things like the example that Keith mentioned were set 15 aside, and they were generally set aside as -- based on 16 the results of IPA 1 and 2, showing that they were things 17 either performance was insensitive to or in some cases, 18 people could not imagine that performance would be 19 bothered at all by any reasonable range of those 20 uncertainties.
21 So there is some historical context there for 22 much broader look than I think as we sit here today and 23 think of only ten things and try to imagine that we've 24 only -- we only thought of those ten or only thought A
through the process for those ten --
(v) 25 NEAL R. GROSS COURT REPORTERS AND TRANSCR:BERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
1 230 1 CHAIRMAN GARRICK: Yes. I think one of the l
,- 2 concerns that some of us on the committee have is that you !
t )
I 3 have developed these tools for telling you what's i 4 important with respect to performance, and that ;
i 5 development has been going on for some time and 6 considerably following, if you wish, the establishment of 7 this list of key technical issues, and the list doesn't 8 seem to change.
9 And one concern, of course, is that we have 10 become fixed on these as the issues, to the extent that 11 the processes that we developed for ferreting out what's 12 important in perhaps a more systematic and deliberate way 13 is not being applied and may not be all that relevant. .
, i 1
i :
\_,/ 14 That's the concern that, as I say, some of us have.
l 15 And we just don't want to see the effort that l
16 you've gone through to develop a method that has lots of
- 17 promise, of telling you what you should be looking at from l
l 18 the point of view of bottom-line performance measures, not i
l 19 be a factor in deciding what you focus on. That's at 20 least one anxiety that I have.
l 21 MR. McCONNELL: I think we her you and I 22 think -- I don't -- it's not an intention to put you off, 23 but I think we are in a position to document those changes 24 and make those changes in priorities in the next couple of fg
) 25 months.
(J NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
231 1 MR. GREEVES: Dr. Garrick, this is John
, ,-. 2 Greeves. Can you hear me?
( )
3 CHAIRMAN GARRICK: Yes, sir.
4 MR. GREEVES: Just to interject, I've been in 5 and out of the room, but I think you were asking, you 6 know, where soft spots are. I don't know whether you 7 talked about it today, but Keith has an equally long list 8 of performance assessment responsibilities outside of 9 high-level waste, so you just need to -- you know that; 10 I'm just reminding you of that.
11 And there are things coming at Margaret and I 12 that we did not anticipate. There is talk of an 13 application from EnviroCare for an NRC review. There's r^x
/ )
\s / 14 talk of increased NRC participation on Ward Valley. And I 15 just want you to be aware there are other demands being 16 made on these same resources that I think you just need to 17 be aware of that. I l
18 I think you are, but today you've been 19 focusing on high-level waste.
20 CHAIRMAN GARRICK: That's right. Thanks, 21 John. In fact, we've been focusing on performance 22 assessment capability and high-level waste. So that's 23 correct, and I think that's one of the issues here that l 24 we're trying to get a sense of is the resource situation
,\
(s_- ) 25 as it applies to these other things as well.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l l 232 1
l 1 MR. GREEVES: These very same people get drawn l
s 2 into those other issues in varying degrees.
\
3 CHAIRMAN GARRICK: Yes. Thank you.
4 MR. GREEVES: Enough said.
5 CEAIRMAN GARRICK: Paul?
6 MEMBER POMEROY: I just had one comment that 7 follows nicely. Thank you, John, for the lead-in or 8 Keith's insomnia problem.
9 It seems almost certain that somebody will ask 10 you to work on a revised Part 60 at some point in the near 11 future, I hope. If they do, I trust that there is some 12 contingency plans on what you do at that point.
13 If they don't, then it's my understanding from 6 <
\~/' 14 something Bill Reamer said in a previous meeting to us i
15 that either the existing Part 60 would apply to any )
l 16 application process and that the DOE went through, or that !
l 17 a draft rule would be selected in its place by some l 18 judgment panel. .
l 19 I guess I'm concerned only that somebody is 20 thinking about those contingency plans, and I guess 21 that -- it's nice to have you here, John, because I can 22 address that to you, too, that I hope you do have 23 contingency plans for those issues. And I know you do.
24 MR. McCONNELL: Maybe -- I'm sure John wants f \
(,) 25 to speak to that. I'd welcome you to, but we do have
! NEAL R. GROSS i COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(702) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433 l
i
233 l
1 1 contingency plans -- well, not contingency plans. We are
,.s 2 aware of the concern. But, in fact, we are developing the
! \
'] 3 strategy for the rule as we speak, the implementing rule.
4 We do anticipate moving forward, assuming we 5 get the Commission go-ahead to do that, with the 6 implementing rule next fiscal year. So I think one 1
7 additional --
l l
8 We have the ability to go out and tap other 9 people within the division to help us in that area, and 10 Janet Kotra is a prime example. She's the primary lead on 11 developing the strategy for the rule right now.
12 but it's the multitude of tasks that could l
- 13 come along, should we compress the schedule, either
/~
l k ,N) s 14 through legislation or other means. l l !
15 Sorry, John. Did you want to talk to this?
l 16 MR. GREEVES: No. You're fine, Keith. And I 17 think maybe this is something that Margaret and I can talk 18 to Dr. Garrick about in some future session. It's -- I 19 just wanted, Keith, while you were up there -- because I 20 know I am asking you frequently what's the full platter, j 21 so that Margaret and I can try and go fight for those 22 resources, and the picture changes a little bit with time.
23 A month ago I had no concept of EnviroCare 24 saying they were going to apply for a low-level waste
,-m (x- ) 25 license to NRC, so I just want to keep the committee NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
234 1 informed. And I think, Dr. Garrick, next time you're in
,_s 2 town, we could go over some of these issues with you.
I i i
"' CHAIRMAN GARRICK:
3 Thank you.
4 George?
5 VICE CRAIRMAN HORNBERGER: Yes. I have two.
6 First let me pose a question to you. We're been reviewing 7 capabilities, and we've repeatedly asked, quote / unquote, 8 soft spots. Let me give you an impression that I've 9 gotten, and then you can respond. Okay?
10 The impression that I have is that, in fact, 11 you probably have done less work in the area of the 12 engineered barrier systems and the analysis of the 13 containment systems themselves than you have in the ;
(~'s !
(_.) 14 geological area, and that, of course, may reflect again 15 the choice of KTIs.
16 Looking to the future, is this -- do you have 17 the capabilities to reverse that in the future?
18 MR. McCONNELL: I think we do. I'll let Tim 19 talk directly to the code. I think technically we do. I 20 think the Center has maintained the capability in this 21 particular area to build it back up should that decision 22 be made by management to change the priorities.
23 To address the first part of your question, I 24 think we're also -- I think as Tim mentioned yesterday, r^x
( ) 25 we' re kind of maybe two or three steps behind DOE, because NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
235 1 they're on the forefront of the data collection and how
,_ 2 they're going to proceed in developing a license
('~' )
3 application, so we're also responding to what appears to 4 be, I think on their part, an increased emphasis on the 5 engineered barrier and developing all these design 6 alternatives.
7 Did you want to speak to the --
8 MR. McCARTIN: Sure. While what we have in 9 the TPA 3.1 code is abstracted and simple, I still would 10 argue it is an improvement. If you look at what we had in 11 Phase 2 and what we have today in Phase 3 -- as Bob Baca 12 mentioned, we brought in the EBSPAC code that was many 13 years in the development within the container life and i
73 i
'(.,/ 14 source term KTI or that particular element if you go back.
15 The terminology is wrong if you go back too far in time.
16 But that particular group of people spent a 17 lot of time developing this code in terms of the ability 18 to look at galvanic protection and some of the corrosion 19 models, et cetera. This is a significant step forward 20 from Phase 2.
21 We still aren't all the way there, and 22 certainly there might be features that we would have today 23 if that funding wasn't taken away a while ago, but it's 24 still -- we still are moving forward. I think the fact O) iv 25 that we can look at chloride concentration and other types NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W l (202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 1
i l
236 1 of things of the near-field environment and chemistry, 7- 2 that we're not all the way there, but we certainly -- I ;
~
3 think we provide a necessary --
4 As Brett Leslie mentioned, there's certain 5 aspects of the code that he can use to see how it affects 6 performance, and I think that is a significant 7 improvement, although not all the way there.
8 MR. McCONNELL: Dr. Hornberger, I think your 9 comment was generic, but speaking directly to container 10 life and source term, while the Center's work in that 11 particular KTI was cut out, some of the work actually was 12 embraced by other KTIs.
13 There's some work on galvanic protection being I
/; ,
l (_-) 14 done within the TSPA KTI and also some being done in the )
1 1
15 near-field KTI. So we still -- that's where the l
16 maintenance of the capability comes from.
i l 17 VICE CHAIRMAN HORNBERGER: The second, if I 1
18 may, is really a comment. Again, it's an impression. You 1
19 can respond if you like.
l
! 20 I share the concern that I think that John l
21 voiced before our break, and that is: The impression that l
l 22 I have listening -- yes, it's clearly that you've made 1 23 some great strides and the logic that you've laid out for I 24 us is great. Nevertheless, it does strike me that n
( ,) 25 there -- at every turn, you are, again because of the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 g
237 1
1 nature of the NRC, going to make conservative assumptions, 1 l
,y 2 if you will.
,( )
~'
l 3 And the concern that I have in terms of 1
4 performance assessment is that these, again -- they can 5 get compounded, and this is still all well and good, but i
l 6 no matter what Norm says about interrogating things, the
)
7 end product is still going to be a CCDF when you go to the 8 total-system evaluation.
9 And I do have a concern that you can get 10 locked in, if you will, to believing these numbers to a 11 certain level. In other words, you tend to mix a worst-12 case analysis with a realistic analysis of uncertainty in 13 the mix, and I think that one has to be a little careful
's #
14 in how that plays out.
15 MR. McCONNELL: Well, I think we're aware of 16 that pitfall, and what we try to do is come up with 17 reasonably cautious approaches. You can call them 10 conservative, but we try not to be extremely conservative.
19 Sometimes you're driven to that, but certainly those are 20 identified, put on the table, for you as well as other 21 peer review groups and commenters.
22 And that's when we can go back and forth about l 23 maybe it's too conservative.
24 CHAIRMAN GARRICK: Bill?
/~%
k ,) 25 DR. HINZE: Well, two comments mostly of a l NEAL R. GROSS l COURT REPORTERS AND TRANSCRISERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
238 1 positive nature. At the review three years ago, one of my
,_s, 2 serious concerns was the interface between the analysts l 1
'"]
3 and the scientist-engineer, and I certainly see within the i
4 NRC work that this problem has been largely alleviated.
5 I'm very pleased to see that, and, you know, Ginny Colten-6 Bradley's activities, for example, are a good example of 7 that.
8 Another aspect is we have two units here that 9 are doing the PA: the NRC in Rockville, the Center here !
l 10 in San Antonio, and I wonder about the communications and ,
i 11 whether communications are in any way inhibiting the ,
1 12 capabilities in these areas.
13 I have looked for that as we have moved i
/s l
( 'j
/
14 through today and read the material, and I haven't seen l
l 15 any in those things which I know most intimately. I sense l 1
16 a very good communication, and I think that's something 17 that we as a committee should think about is the .
l 18 communication aspect of it.
19 The -- in terms of capability, the -- it's 20 obvious that the NRC has to do some kind of simplification 21 for their TSPA work, but what we're talking about is a 22 simplification or, if you will, an abstraction of an 23 abstraction. And I wonder how far this process can go and l 24 really be a meaningful process.
ts 25 How much are we gaining with this? How much
( (s_)
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 2344433
239 1 are we losing with this abstraction squared? Is this --
s 2 how much of this is driven by resources that are k')
3 available? These are questions that I'm not asking of the 4 NRC, but they are certainly questions that we ought to 5 keep in mind.
6 The -- and the whole abstraction of an 7 abstraction leads us to the question: So what? When we 8 get to the results and they compare or they don't compare, 9 then does this mean that it's just the extent to which the 10 work has been abstracted?
11 On another item, I think the work that's being 12 done in terms of identification of uncertainties and the 13 prioritizing of the critical items with the sensitivity
(~/
v-
)
14 studies is really great. The -- I think, though, that --
15 and I realize it's being done, but I'd like to see more of 16 an emphasis being placed upon the trying to identify those 17 uncertainties that should be cut down.
18 We aren't just interested in establishing i
l 19 uncertainties, but there are certain uncertainties that j 20 will give us that much warmer, fuzzier feeling if we can 21 cut down. And that brings us to the EBS.
22 I think that one of the things I've learned 23 here in the last two days is that this conventional wisdom 24 about the EBS as being something that would have much less
/m
( ) 25 uncertainty of concern with than in the case of the LJ
, NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS i 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
240 1 geosphere has been exploded, that there are many
- 2 uncertainties dealing with the EBS which I think places
'^
3 some new insight, at least for me, in terms of the 4 defense-in-depth and some of those concepts.
5 It seems to me that it's unfortunate that what 6 I'm hearing is that there's little basis for much of the 7 work that we are -- or many of the decisions that have be 8 made regarding the EBS, and because of the long time 9 frame, we may not be able to decrease those uncertainties.
10 But I think we need to -- I think there's a need to look 11 at those uncertainties and see what can be done to 12 decrease those.
13 That's the first group. Now let's -- no.
p
\m l 14 That's enough.
15 MEMBER POMEROY: That's your prefacing I l
l 16 remarks.
17 DR. HINZE: That's the preface. Right.
l l 18 MR. McCONNELL: Thanks for the positive l l
19 comments. l 20 CHAIRMAN GARRICK: We have a number of invited l
21 experts and also a consultant or two that are going to be l
l 22 highlighted tomorrow, but if some of them have a comment 23 or two they'd like to make on the basis of what's 24 been said today, it would be timely to do so.
(~\
(j 25 Of course, Ray, we're looking for a different NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 241 1
1 role for you later on, and I don't know if your exposure, l 1
7s 2 limited exposure here has inspired you to make a comment
! ) ,
3 or two, but we'd sure welcome it if it has. I 4 MR. WYMER: I'm mildly inspired. I'd like l l 5 to -- I suffer the disadvantage you always have of being 1
6 sort of at the end of the line. Everything has been said, l 7 and I should mention: My background is a chemist and a l l
8 chemical engineer, so anything I say will be out of that 9 context. And much of what was said by the other chemist, 1
i 10 the chemical engineer down at the other end there, Marty ;
i 11 Steindler, but I think -- !
12 VICE CHAIRMAN HORNBERGER: We have to put up l 13 with two of them today. l e
f) i
' w ./
14 MR. WYMER: I think it's worth repeating, 15 though. It's -- from my point of view, it seems like 16 there's not been enough attention paid to what I would 17 call the chemistry involvement in these modules.
18 I think Marty was right on target when he said 19 that or he was implying at least that there's a lot more 20 to the chemistry than the effect of temperature on pH and 21 in the chloride ion attack on the particular technetium 22 and neptunium wire.
23 Or elements with a lot of valance dates and 24 the chemistry of the repository and the containers, the
(-
(. ,) 25 cladding materials and so on are really going to affect NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
242 1 this, affect the chemical nature of these species, and em 2 they're going to move differently than what is presently (v) 3 being modeled.
4 And it seems to me that that's something that 5 sooner or later you're going to have to pay some attention 6 to, chemistry inside and outside of the fuel.
7 And I don't know whether it's important or 8 not, but I didn't see any mention of any modeling of the 9 chemical interactions of magma with these containers. It 10 was more -- sounded to me more like a dispersion 11 discussion than the possibility of a reaction of this red-12 hot magma chemically with stuff and perhaps a binding it 13 up even, rather than releasing it. So there was something N- 14 there. ;
I 4
15 I had sort of a half facetious comment I l l
I 16 wanted to make with respect to the interactions of --
17 DR. HINZE: We need one today, 18 MR. WYMER: It's timely. -- with respect to 19 the interaction of the NRC and DOE's modeling activities.
20 I'm reminded of what happened about the turn of the 21 century when people were measuring the speed of light, and 22 they were measuring it, getting the wrong values, but 23 everybody agreed that the values were right, and there was 24 a consensus, even though they were wrong, because all of I
/m i
(,) 25 the right results were considered spurious and they were l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l l (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
1 1
243 1 thrown out.
7_ 2 I would suggest that you be a little careful
.:'-)
I 3 that you don't get too much consensus and throw out the 4 spurious right things we should be considering. That's 5 all I had to say.
6 CHAIRMAN GARRICK: Okay. Thank you.
7 Any of the invited experts want to make a 8 comment today or -- Stan or Chris?
9 MR. WHIPPLE: Yes, John. I'm happy to make 10 one. Chris Whipple.
11 You and George both commented on an issue that 12 I've been thinking about recently, and that's the issue of 13 whether and to what degree it's possible to do a
-x i ) 1 N/ 14 sensitivity analysis in the framework of a repository l
l 15 performance assessment.
16 It's a very complicated system, and as you all 17 have noted, it's a combination of realistic and bounding 18 analyses kind of held together with baling wire and Scotch l 19 tape. And to extract from that an unbiased evaluation of l
20 sensitivities, I think, is a task that is largely 21 judgmental, not analytical.
22 And I want to try to illustrate that with a j 23 few examples from the WIPP-PA. In the WIPP-PA, for a 24 number of iterations, point estimates were used for
) 25 critical values in an attempt to be conservative, to use a i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
l l
t' 244 1 bound. And as a result, there was no opportunity to l ,~,
2 identify those as sensitive parameters.
1 3 I used to joke that the rock mechanics people 4 at WIPP got to the meeting first and grabbed all the 5 margin, and the chemists got there late and had to do 6 realistic work. And there was an assumed permeability in 7 WIPP 5,000 years after it's closed that showea the salt 8 was seven orders of magnitude more permeable than the time 9 the repository was operated.
10 And if you ask the rock mechanics people if 11 they really believe that, they'd say, of course not, but 12 we don't know how to defend another analysis exactly, and 13 it's conservative and we all know that's a good thing.
/~N ,
b
(_)
? l 14 And what it really was was a thing that caused l l
15 a lot of time and money to be spent on'the wrong areas of I
16 analysis.
17 A related example: They spent over a million 18 bucks looking at climate change at WIPP. Well, the WIPP 19 repository is 2,000 feet below the water table. It 20 doesn't know if it's raining. And when we finally said, 21 So what; what if you get climate change; what if it rains 22 three times as much, well, they said, We assume that the 23 water table rises all the way up to the ground surface.
j 24 And we said, So what. And they said, Well, the flow in i
(A)
.v 25 the Culebra Aquifer is up to 5 percent faster. And we i
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
245 1 said, so what. And they said, Well, we just thought you
, ~.s 2 should know.
(x /)
3 (General laughter.)
4 MR. WHIPPLE: And I think there's some of that 5 going on in the Yucca Mountain project. It's inevitable.
6 People get into an area. They get good at it. They like 7 doing it. And it's very difficult to say, Enough -- this 8 may be fun, but it's not going to be important in the end.
9 And I heard a number of comments in the last 10 two days that amplify on the need to pay attention to 11 this. First is that the NRC resources are limited, both 12 with competing demands on the staff time and attention, 13 and or, the budget.
,g k,_) 14 And the other comments were that we need to 15 include all the parameters, to do sensitivity analysis on 16 everything that moves. And you can't do both. You can do 17 some things well, or you can do everything superficially.
18 And I think it's a key management task to avoid the 19 temptation to say, yes, to do everything, and to really 20 find a way to close issues.
21 Ray brought up the issue of magma. I will 22 admit my own cynical reaction to that presentation in the 23 technical exchange meeting a couple of days ago or 24 yesterday, whenever it was.
T'N
()
s 25 We spent a lot of money figuring out how to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
246 1 take defense wastes and encapsulate them in a magma-like 7- 2 glass, because we think that makes them safe. However, if t.v) 3 magma were to encapsulate the spent-fuel canisters, 1
i 4 somehow that would be a major risk for the program. !
5 And I don't think we got the sign right on one 6 of those calculations. So I just think that there's an 7 effort to prolong issues that should be closed in the name l 1
8 of conservatism that keeps money going into the real ;
9 stuff.
10 And I think we identified a lot of things i 11 today that are real stuff that you all are working on, but 12 where the sensitivities are critical, the basic flow 13 models, saturated and unsaturated zone, are key to the
/% j
-- 14 performance of Yucca Mountain, and you're working on them.
l 15 But if you had more staff time and money to work on those, 16 you could probably do a better job. You might have time 17 to invent creative solutions that DOE's missed and to l
18 really stand back and understand what's going on.
19 Okay. One final set of issues that is part of 20 the background that both NRC and DOE have to work under, 21 which is doing performance assessment when you don't know l
22 what the performance measures are is a tough task. And we l 23 don't have a standard yet.
24 And the way that it's been done by DOE -- and
,a
( ,) 25 I'm less familiar with NRC -- has been to pick a number of
! NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
l
247 1 end points, which I think's a good thing, but sometimes
,-~g 2 that's -- you have to be careful how you do that. And
( /
~
3 I'll illustrate.
4 In TSPA-95, DOE calculated individual doses at 5 10,000 years and a million years. Well, what they found 6 was at 10,000 years, they ran lots and lots of 7 realizations and had only a few that had non-zero hits.
8 So that didn't really tell them much about sensitivity of 9 what matters in the performance.
10 At a million years, none of the engineered 11 barriers matter. They've all failed by then. So those 12 two measures are really no good at giving you insights as
,, 13 to engineered-barrier perforn ance. It doesn't mean it's
's 14 unimportant; it means you need a different measure. I 15 don't know what that measure is: time to significant 16 release, however you define that or something.
17 But I think that's something you need to think 18 about. Similarly, in the lack of a defined standard, I'm 19 struck by the difficulty in picking up the phone and 20 calling EPA, not that they can tell you what they're 21 doing, but a lot of what we heard today about the l
l i 22 saturated zone dependo critically on the -- what the EPA 1
l 23 does on the biosphere side, absolutely critically.
24 And I just hope somebody can impress on EPA
, 25 that anything that they can do to shed light early will NEAL R. GROSS COURT REPORTERS AND TRANSCR!BERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
248 1 tell you what they can about what to assume about critical
._, 2 groups or where they are, how much water they pump.
/ s 3 If they pump a million cubic meters a year, 4 maybe dispersivity's unimportant. If they pump 10,000 5 cubic meters a year, you better get to work on 6 dispersivity in the saturated zone. And you won't know 7 whether it's important until EPA comes out with a 8 standard.
9 In the peer review that the TSPA peer review I 10 panel that I'm involved with just sent in -- sorry we had 11 nobody from NRC at our meeting, because we had a similar l
12 comment on how the DOE folks' hands are tied from doing 13 competent criticality analysis, absent a clarification
(_ / 14 from NRC on how to interpret Part 60's criticality i 15 standard, let alone what will replace Part 60 down the 16 road, how it'll be revised. l l
17 But even the current standard is unclear, and I
18 the DOE folks are guessing at what it might mean, and 19 they're spending a lot of time and motion, I think, in l 20 perhaps a fruitless direction as a result.
l 21 So with those comments, I'll quit.
22 CHAIRMAN GARRICK: Thank you very much.
23 Stan, did you want to wait till tomorrow?
24 MR. KAPLAN: Yes. For the most part, but I'll
,m I 25 just say I'm happy to hear of the increased emphasis and NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
249 1 attention going to the engineered-barrier design,
,_ 2 recognizing the caution that there is lots of
! )
3 uncertainties in that area.
4 Still, I feel more comfortable with the active 5 approach. In the case of the geological barriers, there's 6 not much we can do. If we turn our engineers loose on 7 designing containment system, I have great confidence that 8 we can achieve darn near whatever performance we want.
9 CHAIRMAN GARRICK. For an applied 10 mathematician, that's a very good endorsement of 11 engineering.
12 MR. KAPLAN: Well, you have lots of capability 13 to do something in one case. In the other case, we're e
f3 a
(_) 14 passively sitting here, hoping that the good earth --
15 CHAIRMAN GARRICK: Okay. Unless anybody elce 16 has a comment to make, I think we're pretty much through 17 with this phase of our meeting. I think what we'd like to 18 do now is while we're still on the record, is talk a 19 little bit about future letters, and then go off the 20 record to process a letter that we've been working on for 21 quite some time.
22 I guess I'd like the committee's views on what 23 they see on the plate right now, as far as letter-writing 24 activities are concerned. There's clearly the defense-in-
/~
(3) 25 depth letter. We have heard considerable from various NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
250 1 agencies, organizations the past three months on the issue l -
,s 2 of subsystem requirements, and I think that we've heard 1\
l
' ') 3 encugh now to probably write a letter on that.
l 4 I thought that Janet Kotra highlighted the l
5 options that exist with respect to subsystem requirements i 6 pretty well, and those options have been preserved for us l
7 in a set of notes developed by Lynn Deering.
8 But she noted the possible options for i
l 9 revising the subsystem criteria include: 1. A rtandard 10 for the overall system performance only; 2. A standard for l
4 11 the overall system and quantitative subsystem i l
12 requirements, but with a rigorous nexus; 3. A standard for l l
i I
- 13 overall system performance wita qualitative subsystem !
l /~'i l l / 14 requirements; and 4. Multiple barrier options that provide 15 either multiple redundancy, engineered barriers as 1
l 16 supplements, or partial redundancy. And, of course, '
17 there's always the no-action option of status quo.
18 So I think that probably it's incumbent upon 19 the committee to prepare a letter on that subject, and I 20 guess I'd like to get the opinions of the consultants and l
21 committee members on that.
i
( 22 Have you got a feeling about that one, Marty?
23 DR. STEINDLER: Yes. I view the whole 24 question of whether or not subsystem requirements are to
(~h 25 be inserted or somehow touted as a good thing simply a
(
)
NEAL R. GROSS
! COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
251 1 translation of defense-in-depth, and the issue of defense-
,-~ 2 in-depth is, I think, easily divisible into two parts.
1 ( )
l '~' 3 On one side, you could argue that good 4 regulations would simply require the applicant -- and by 5 the way, this is clearly not limited to high-level waste.
6 It would require the applicant to conform to a set of end 7 product regulations, based on risk presumably, and it 8 doesn't make any difference how he gets there or she gets 9 there, but it does make a difference how uncertain that 10 analysis of the applicant is.
11 And one presumes that you want to build in 12 some kind of quantitative requirement for certainty. That 13 may turn what little hair you have left, John, to gray,
,m i 1
'w/ 14 but let me continue.
15 The other point that I was --
16 CHAIRMAN GARRICK: I wish you would.
17 DR. STEINDLER: The other point that I would 18 make is that we've heard enough here on the approach of 19 performance assessment and particularly probablistic 20 performance assessment that I find it difficult to see 21 where the NRC could answer the question, Are you sure that 22 the applicant has considered all the things that they need I
23 to consider. And that's the same question that I posed 24 regarding the scope of the activity regarding existing 25 KTIs, which came from KTUs, and it doesn't make any
(%..-)
! HEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
252 i
1 difference how they got there.
s 2 And in the absence of that, Mal Knapp gave us
. ! )
l Li l 3 a lecture sometime back, gave the committee a lecture. In 4 effect, he said, the reason you want that defense-in-depth 5 is because there may be something you didn't think of.
6 But he also pointed out there may be something 7 that doesn't work very well, and so as a consequence, you 8 want to be able to sequential failure problems under 9 control, especially if the barriers are important. I 10 My difficulty at this stage of the game is I
11 that if, in fact, it's true that the commission, for one 12 reason or another, is moving toward this risk-informed, 13 probablistic performance assessment-related process for
' \ \
2 14 their regulatory program, they cannot easily in any of the 15 things to which you might apply this address the question, 16 Has the applicant provided analysis of all possible 17 events, situations, et cetera.
18 And it's for that reasc7 that I think -- and 19 only for that reason; in other words, it's a methodology-20 driven reason -- that the NRC must continue to demand 21 defense-in-depth subsystem criteria, carefully worded, not 22 the way the Part 60's written, by the way.
23 CHAIRMAN GARRICK: Right.
24 DR. STEINDLER: But some kind of so-called s
(,,) 25 subsystem criteria defense-in-depth. I think it has to be NEAL R. GROSS COURT REPORTERS AND TRANSCR!BERS 1323 RHODE ISLAND AVE , N.W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234 4 433
253 1 that way. Otherwise, you can't defend the regulatory
,_ 2 process that the NRC is going through.
i )
! 3 CHAIRMAN GARRICK: Paul, do you have a --
4 MEMBER POMEROY: Well, I'd just like to make a 5 short, very short comment basically. I certainly think we 6 should write that letter. The only question that I really 7 have that we ought to explore is the timing question.
8 As I understand it, there is the staff paper 9 that's going forward to the Commission in September or i 10 something like that time frame. At that time, that would l
11 presumably then be a public document. We could respond to l 12 that. We could respond to what has been said about -- to 13 us here, and we can respond to what the Commission has
('h l
(_,/ 14 said already if we chose to do that, in terms of what i 15 they've said in the direction-setting issues paper.
16 I would personally like to hear what the staff 17 is going to say in its finalized form as it goes up to the 18 Commission before we write a response, but since my 19 situation is precarious, I think you ought to think about 20 it.
21 CHAIRMAN GARRICK: Yes.
22 MEMBER POMEROY: That's the case. That's all i
23 I have.
24 CHAIRMAN GARRICK: George, how do you feel
[)
q ,i 25 about the defense-in-depth letter?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234 4433 l
l
l
' 254 l 1 VICE CHAIRMAN HORNBERGER: Well, I mean, I 7_ 2 think we should write a letter. And I don't -- I believe
/ !
'~# 3 that we should -- we need to think about the issue that 4 Marty raised. I don't know the status of that. We should 5 perhaps learn something about the timing.
6 Other than that, I don't -- I think that's 7 fine.
8 CHAIRMAN GARRICK: Bill?
9 DR. HINZE: Oh, absolutely. I think this is a 10 great opportunity for the committee to serve the 11 Commission and probably needs an outside view. There can 12 be no question, as Marty has well put it, that we do need 13 a defense-in-depth, and that's been, I think, really
(_) 14 amplified by the last couple of days here.
15 The -- and certainly I think the country 16 doesn't need the stringency of the present 60, but also 17 I'm very concerned about making those requirements too 18 qualitative and really inviting all kinds of mischief that 19 could be developed associated with it.
20 I would think that there's room here -- I 21 don't -- about the timing will have to be taken into 22 consideration, but I think that there may -- yor may wish 23 to set aside some time at the next meeting for further 24 discussion of the -- going through the uses of defense-in-i r~N 25 l (x_.-) depth, how it is implemented, and the various NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
255 1 alternatives, and look at those options and discuss that.
! ,, 2 CHAIRMAN GARRICK: Well, I would hope we may
- \
i 3 even be able to do better than that. I would hope that we 4 could maybe consider having a letter developed at our next 5 meeting, and if we're ever so lucky, maybe even finishing 6 it and getting it approved.
7 I think that what we're all saying is that 8 defense-in-depth is a sound underpin of regulatory 9 philosophy, that what we're talking about mainly now is 10 how to prescribe it, how to make sure that the concept 11 continues, and of course, in the presence of a l
12 performance-based and risk-informed world, we should be in l
13 a much better position than ever before, it would seem, to
'v/ 14 be a quantitative relative to the effectiveness of the 15 various barriers or lines of defense in the repository, 16 for example.
17 So I think that the direction, it seems to me, 18 that we're going is the one to strongly endorse 19 continuation of that policy, if you wish, and, two, to 20 employ the tools and thought processes that we have at our 21 disposal to give greater visibility into the effectiveness 22 and the specific role of these lines of defense.
i i
23 And these two points, it seems to me, is --
24 constitutes a basis for us to craft some sort of a letter gg 25 (v) which, I think, we should go ahead and do as well, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
256 j 1 DR. STEINDLER: Let me make one other point.
7- 2 If you write a letter concerning the advantages of l
() 3 defense-in-depth, as Lynn, I think, very carefully has 4 written on here, what that tells you is that you don't 5 have a whole lot of faith in performance assessment.
6 Now, that's fine. I don't have any problem 7 with that conclusion, but I think you -- I had to say l
8 that -- but I think we need to recognize that's, in I
9 effect, what you're saying.
i 10 CHAIRMAN GARRICK: Yes. But I read that 1
11 entirely differently.
! 12 DR. STEINDLER: I would be surprised if you l
13 didn't.
' (ms) 14 CHAIRMAN GARRICK: I think that -- )
i l l
15 DR. STEINDLER: But you do have to address the 16 issue --
17 CHAIRMAN GARRICK: I think performance 18 assessment gives real meaning to what we mean by defense-19 in-depth, because it gives us the wherewithal and l
20 resources with which to be much more specific in terms of I
21 the effectiveness of the various lines of defense with 22 respect to the overall performance. And, you know, that's i
23 the angle that I think maybe our letter should gite.
24 Okay. The other thing before we go off the em (s.- ) 25 record here that I wanted to briefly discuss: We have the l NEAL R. GROSS
! COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
257 I possibilities on the basis of today and tomorrow two more l
lgy 2 letters, one with respect to staff capability and one with
! ( '~' /
l 3 respect to the PA technology and its general status, I
4 particularly in relation to how to increase our confidence 5 in its use.
6 I would -- I think probably there should be 7 two letters written. The one possibility exists is to l 8 write one letter and combine both issues, but I think 9 there's an advantage in treating them separately, and one 10 advantage is that we may be able to write a shorter 11 letter, and they're more likely to be read.
12 And the other advantage, I think, is that the 13 subjects are quite diff2 rent, and we can be maybe quite
\s' 14 precise with respect to those two subjects with two 15 letters better than we can with one.
16 So I would guess that the committee is in 17 general agreement that we should move in the direction of 18 generating those two letters. Okay?
19 MEMBER POMEROY: You're right on target with 20 that, John. I think we should try those two letters, not 21 one.
22 CHAIRMAN GARRICK: Okay. Excellent.
23 Well, I think that with that, we'll probably 24 go off the record now and work on the one letter that we
(-
( ,/ 25 want to finish today if we possibly can. So the meeting NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
258 l 1 is adjourned as far as the record is concerned.
,s 2 (Whereupon, at 4:50 p.m., the meeting in the
! 's s /
'~~'
3 above-entitled matter was adjourned from the record, to 4 reconvene at 8:30 a.m., Thursday, July 24, 1997.)
5 l
6 l
7 1 i
1 I
8 9
10 l
1 11 12 1
13 )
73 l
6 l
l
\j 14 !
15 16 17 j 18 19 20 21 22 23 l
24 g-~x, 25 i% )*
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASH!NGTON, D C. 20005 3701 (202) 234-4433 l
{%
V CERTIFICATE ,
This is to certify that the attached proceedings before the United States Nuclear Regulatory Commission in the matter of:
l Name of Proceeding: 93" ACNW Docket Number: N/A ,
Place of Proceeding: SAN ANTONIO, TEXAS i were held as herein appears, and that this is the original ;
transcript thereof for the file of the United States Nuclear Regulatory Commission taken by me and, thereafter reduced to ,
typewriting by me or under the direction of the court !
! reporting company, and that the transcript is a true and ,
accurate record of the foregoing proceedings.
i ALMU A __ 0}4_ '
dGNNi PEEg / ~ ~
Official Weporter Neal R. Gross and Co., Inc.
l l
l l
l
(
i NEAL R. GROSS COURT REPORTERS AND TRANSCRil3ERS 1323 RilODEISLAND AVENLT,NW l' (202)234 4433 WASIIINGTON.D.C. 20005 (202)234-4433
\