ML20138F495
| ML20138F495 | |
| Person / Time | |
|---|---|
| Issue date: | 10/11/1996 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-2077, NUDOCS 9610170066 | |
| Download: ML20138F495 (135) | |
Text
._ .____ - . . . .. . _. -. _ . - - . . . . - -
Official Trcnscript cf Prscocdings O NUCLEAR REGULATORY COMMISSION
/tRS7~ 2o??
Title:
Advisory Committee on Reactor Safeguards 435th Meeting TRO4 (ACRS)
RETURN ORIGINAL l TO BJWHITE M/S T-2E26 Docket Number: (not applicable) 41s-713o THANKS!
Location: Rockville, Maryland .
l O Date: Friday, October 11,1996 l
l l
Work Order No.: NRC-871 Pages 471-564 Oi 6 961011 T-2077 PDR .g j
NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers 1323 Rhode Island Avenue, N.W.
p Washington, D.C. 20005 V 3 2) 234-443:
01r51 RS C"ce 003y- Re:a.
~
ca
. .r f n *< <
10~:10 _ #6 01 'l'e v0T2 :::ee
)
.- . - . - . . - -.- -- ._- .---- --- - - - . ~ - . . _ _ . - -.
'T l
th DI8 CLAIMER i
j PUBLIC NOTICE BY THE UNITED STATES NUCLEAR. REGULATORY COMMISSION'S ADVISORY COMMITTEE ON REACTOR SAFEGUARDS ,
l l
OCTOBER 11, 1996 l l
The contents of this transcript of the proceedings of the United States Nuclear Regulatory Commission's Advisory Committee on Reactor Safeguards on OCTOBER 11, 1996, 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 )
and edited and it may contain inaccuracies.
i o !
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVENUE. NW (202) 234 4433 WASHINGTON. D.C. 20005 (202) 234-4433 j
471 1 UNITED STATES OF AMERICA 2 NUCLEAR REGULATORY COMMISSION O 3 +++++
4 435TH MEETING S ADVISORY COMMITTEE'ON REACTOR SAFEGUARD (ACRS) 6 +++++
7 FRIDAY 8 OCTOBER 11, 1996
.9 +++++
10 ROCKVILLE, MARYLAND 11 +++++
12 The Advisory Committee met at the Nuclear 13 Regulatory Commission, Two White Flint North, Room T2B3, r
14- 11545 Rockville Pike, at 8:30 a.m., Thomas S. Kress, 15' Chairman, presiding.
16 COMMITTEE MEMBERS:
17 THOMAS S. KRESS, Chairman 18 ROBERT L. SEALE, Vice Chairman 19 GEORGE E. APOSTOLAKIS 20 JOHN J. BARTON 21- IVAN CATTON 22 MARiO H. FONTANA 123 DON W. MILLER 24 DANA A. POWERS 25 WILLIAM J. SHACK NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISt.AND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
472 1 ACRS STAFF PRESENT: ;
f 2 JOHN T. LARKINS, Executive Director 1
ROXANNE SUMMERS, Technical Secretary 3
I 4 SAM DURAISWAMY l
5 CAROL A. HARRIS 6 RICHARD P. SAVIO l
7 PAUL BOEHNERT l
8 NOEL DUDLEY 9 MADHAT M. EL-ZEFTAWY 10 MICHAEL MARKLEY l 11 AMARJIT SINGH 12 13 ALSO PRESENT:
5- 14 WAYNE HODGES 15 FAROUK ELTAWILA 16 RALPH CARUSO 17 JOE KELLY 18 DAVID BESSETTE 19 20 21 22 23 24 I
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
_mm ---
473 1 A-G-E N-D-A c3 2 AGENDA ITEM PAGE x.
3 16) Openino Remarks by the ACRS Chairman 474 4 17) Activities Associated with the NRC 476 5 Thermal Hydraulic Codes 6 17.1) Remarks by the Subcommittee Chairman 476 l
1 7 17.2) Briefing by and discussions with 479 l
8 representatives of the NRC staff 9 regarding the staff activities 1
1 10 associated with the NRC thermal l l
11 hydraulic codes 12 l 1
1 13 l r\
\'v) 14 l
1 1
15 16 17 18 19 20 21 22 23 24
/%
$j 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
474 1 P-R-O-C-E-E-D-I-N-G-S 2 (8 :34 a.m.)
3 161 OPENING REMARKS BY THE ACRS CHAIRMAN 4 CHAIRMAN KRESS: The meeting will now please 5 come to order. This is the third day of the 435th meeting 6 of the Advisory Committee on Reactor Safeguards. During 7 today's meeting, the Committee will consider the following 8 items: First, the activities associated with the NRC 9 thermal hydraulic codes. Then we'll have a report by the 10 Human Factors Subcommittee chairman maybe. And then we're 31 going to work on our reports most of the day.
12 This meeting is being conducted in accordance 13 with the provisions of the Federal Advisory Committee Act. I 14 Mr. Sam Duraiswamy is the designated federal official for 15 the initial portions of the meeting.
16 We have received no written statements or 17 requests for time to make oral statements from members of 18 the public regarding today's sessions. A transcript of 19 portions of the meeting is being kept. So it is requested 20 that the speakers use one of the microphones, identify 21 themselves, and speak with sufficient clarity and volume 22 so that they can be readily heard.
23 Before we start, I have a couple of reminders 24 for the members. Remember you have some interviewing to 25 do. So look at the schedule and see when you're supposed NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
475 1 to do that. It looks like this, Ivan.
,, -s) 2 And I also wish to remind you that you need to O 3 be back here in this room at 1:00 o' clock today for 4 mandatory ethics training, 1:00 to 2:30. So if you wish 5 it, you're in trouble. You have to do it separately some 6 other time. So please be here then.
7 MEMBER CATTON: How long is the ethics 8 training going to last?
9 CRAIRMAN KRESS: It's 1:00 to 2:30.
10 MEMBER CATTON: 1:00 to 2:30. An hour and a 11 half.
12 MR. DURAISWAMY: It's about an hour, Ivan, --
13 CHAIRMAN KRESS: Maybe it's an hour.
\' 14 MR. DURAISWAMY: -- and then a half an hour 15 questions and answers.
16 MEMBER CATTON: I see.
17 CHAIRMAN KRESS: Then we have to prepare a 18 report on it.
19 VICE CHAIRMAN SEALE: Depressurization.
20 MEMBER CATTON: Some of us need it more than 21 others, I guess.
22 CHAIRMAN KRESS: But before I turn to the 23 official agenda items, are there any members who wish to 24 make any comments or have anything they need to bring A
!d 25 forth?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
._.__.__.__.____m . _ _ _ . . . _ _ . _
3 i 476 4
1 (No response.)
. 2 CHAIRMAN KRESS: Seeing none, the first item 3 on the agenda this morning is the thermal hydraulics 4 activities. I guess, Ivan, that's yours. I'll turn the i
5 floor over to you.
, 6 -MEMBER CATTON: Thank you, Tom.
1 7 17) ACTIVITIES ASSOCIATED WITH THE d'
! 8 NRC THERMAL HYDRAULIC CODES 1
l 9 17.1) REMARKS BY THE SUBCOMMITTEE CHAIRMAN i
i 10 MEMBER CATTON: What we're going to hear about i
{ 11 this morning is the five-year plan for thermal hydraulics.
l 12 The focal point of the plan is the code. And the present i
13 plan, near as I can understand it, is to consolidate 14 existing codes, like TRAC and RELAP, into a resuscitated i
j 15 version of TRAC.
I 16 The intent is to take the best models and make
{
17 it work. It's not clear at this time that this will be 18 worth the effort. And I think they are probably going to 19 look into some of.the elements that make it a little bit 20 speculative.
21 If the TRAC-P code cannot be made more 22 flexible, then we might as well live with what we have and 23 pursue a more modest program. I say this at the front end 24 because I'd like to be able to make a statement in the bQ 25 letter about this.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
477 1 I'm pleased with whrt we have on the table. I gS 2 think it's an opportunity to do something that's long
's) 3 overdue. In my view, the thermal hydraulic part of 4 research has been revitalized. They not only have a new 5 code in planning but three very nice thermal hydraulic 6 facilities.
l 7 We finally have gotten away from i 8 high-pressure, high-temperature, tall, skinny facilities 9 that are expensive to operate and located God knows where 10 at some national lab.
11 CHAIRMAN KRESS: Now, wait a minute.
12 (Laughter.)
13 MEMBER CATTON: The new facilities model the 7
( i
's
14 boiling water reactor, a Babcock and Wilcox PWR, and the 15 AP600. These facilities combined with the relationship 16 with the French round out what to me appears to be a very 17 sound and worthwhile program. And in my view, even if the 18 plan code doesn't come to pass, the whole process has been 19 a worthwhile effort.
20 You'll find background information under Tab 21 17 in the notebook and the meeting Handout Number 17. In 22 the Handout Number 17, you'll find a report by Dhir and 23 also some comments by Zuber. I guess you all know Novak.
24 So what I did is I exorcised the minutes to get something f^s
(_) 25 put together.
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 478 1 To properly set the stage for what we're going l I'
<~s 2 to hear, I've asked to hear from one of the code customers kmsk l 3 first. I recommend you listen closely to what he has to l
4 say because, again, if he's unhappy, the code will not get 5 the use that the expense requires. The user is Mr.
6 Caruso. And he will' speak for both NRR and for Steve 7 Arndt, who'is at AEOD.
l l 8 I would like to mention one other customer as 9 long as I'm talking about customers. Some of the'results l
l 10 that have come out of the University of Maryland facility l 11 show, at least in my mind, a need for the simulation l l 12 arena.
! 13 The stratified flow in a pipe with the lower l
14 regions of the water being cold caused some rather strange 15 results to occur. I don't know whether I'm right.or 16- wrong, but it seems to me that an operator ought to see 17 what's going on in the plant and not strange behavior if 18 we want to depend on what he's going to do.
19 I think in this regard the new tool could be a 20 valuable help in ensuring that you have proper fidelity in 21 the simulators. And let's face it. The operators run the l 22 plant.
23 With that, I'll call on Mr. Caruso.
24 (Slide.)
O NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. g (202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l l
l -
479 1 17.2) BRIEFING BY AND DISCUSSIONS WITH REPRESENTATIVES OF 2 THE NRC STAFF REGARDING THE STAFF ACTIVITIES ASSOCIATED p)
'w/
3 WITH THE NRC THERMAL HYDRAULIC CODES 4 MR. CARUSO: I gave this presentation to the 5 subcommittee last month, and I used some of our high-tech 6 equipment to present it off the internet. See, I'm not 7 used to this retro equipment.
8 Unfortunately, you've let somebody muck around 9 in your computer system. And now I can't get Netscape 10 running. So now I've got to use transparencies. They're 11 just prints of the slides off the internet. So the "Next 12 slide" button here is not functional on your piece of 13 paper.
p l )
'd 14 My name is Ralph Caruso. I am the group ,
l 15 leader of the Analytical Support Group. which is a small i 16 group of people in NRR that use analytical codes to l
17 provide analytical support services to the Reactor Systems 18 and Containment Systems Branch, which are the line 19 branches responsible for reactor systems and containment 20 systems in NRR.
21 We were formed about three-four years ago to 22 start performing these functions, mainly in support of the 23 AP600 and SBWR reviews. Since then we've done a lot more 24 in other areas. And I'll be talking about them as part of
()
() 25 the presentation.
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-1433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
l
.___ ____. _ _ _ _ _ . .m .. __ _____ , _ _ _ _ _ . _ _ _ - . . _ _ . _ _ _ . _ . _ _ . _ _
i 480 l i- )
4 1
! 1. (Slide.)- I
\
1 2 MR. CARUSO: How do we use thermal hydraulic
! \_/
t 3 codes? Historically they were started in NRC as an
\
4 independent check of licensee calculational methods to l l
5 cross-check the vendor codes, the vendor calculations, the j l
6 evaluation models that were used by licensees.
7 In addition, they have been used to perform 8 sensitivity calculations. If you want to understand hou a 9 particular evaluation model might be sensitive to slight 10 changes ~in input conditions because of uncertainty of the 11 input conditions, use the thermal hydraulic codes and vary 12 the inputs slightly. And run them long enough, and you 13 get the results. And you plot them against one another, 14 and you get a sensitivity calcuation.
15 In addition, event analyses. Frequently we 16 have operating events occur at reactors during shutdown; 17 for example, the instrument technician opens the wrong 18 valve and drains down part of the reactor coolant system.
19 The question is always asked: How long would it take to 20 drain down ti. 3 actor coolant system? How close did we 21 get to core melt? Okay? Events analyses.
22 Recently, as the codes have become faster and 23 computers have become more powerful, these codes are being 24 used for simulation purposes, simulations to support
() 25 probablistic risk assessment, simulations for training, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
481 1 simulations to look at operating procedures to understand 7S 2 what the operators will see and try to develop procedures
() 3 that the operators can use during transients and 4 accidents.
5 (Slide.)
6 MR. CARUSO: I'll give you some examples of 7 the types of analyses that we get in my group. On the 8 very top are vendor comparisons. We haven't done a lot of 9 these lately, although the business seems to be picking up 10 as the vendors go to longer-live.S cores with higher 11 burn-ups, as they put in cores which are designed to 12 reduce flux at the reactor vessel wall so they are 13 creating higher peaking factors in the cores. This is the
/~N
\N -) 14 original reason why we have the codes. The need dropped 15 off for a while, but now it's picking up.
16 PRA success criteria is a big issue. This is 1
17 really exploding. We want to do risk-based regulation. l l
l 18 In order to do risk-based regulation, you've got to have a l
19 good understanding of how the machine is actually going 20 to behave. That means you need a best estimate code to do 21 that.
22 Steam generator differential pressure 23 calculations. Steam generators have been having problems.
24 And I've been called upon several times to calculate
(~x
(,) 25 differential pressures or flow patterns inside steam NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
482 1 generators. The codes I've used for this were not really 2 well-suited for that purpose. And, as I say, although 3 we've got results and we try to defend them, they're 4 really extending the range of the codes quite 5 considerably.
6 BWR shroud cracking. We had a situation here 7 where there was cracking at the bottom of the shrouds of 8 BWRs and there was some concern that if you had a steam 9 line break or a recirculation line break, the lateral 10 loads could cause the shroud and the core barrel to uplift l
11 or tip over and somehow relocate itself inside the vessel.
12 We used TRAC-P for this analysis.
13 Shutdown events, the-Wolf Creek. These are
- 14. the events analyses that I'm talking about. Wolf Creek 15' was a case where someone drained down a reactor coolant 16 system during shutdown. And we looked at how long it 17 would have taken to drain all the way down to get to a 1
18 point where the core would have been uncovered.
19 Hope Creek was a case where someone shot off 20 an RHR system. We used TRAC-P -- TRAC-P it was -- to 21 understand what the flow was like inside the Hope Creek 22 reactor vessel during this transient and how close we came 23 to actually boiling in the core.
24 Finally we had a case at Shearon Harris, where 25 an ECCS system operated in a degraded mode for about a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4 33 WASHINGTON, D.C. 20005-3701 (202) 234-4433
483
- . 1 year and a half. A question that came up was suppose we i
2 had had an accident during that time. Design basis 3 accident, slightly more severe. We used RELAP in order to 4 do those calculations.
! 5 (Slide.) j 6 MR. CARUSO: These are the types of problems 1
- 7 that I get. Now, considering these problems and '
J l 8 considering where we're going, I think I'd like to divide 9 the world into three classes of code users. Okay?
10 The first code user, the knowledgeable, 1 11 qualified, user. These are the people we've had for the 12 past 25 years running these codes. They know the details 2
13 of the codes. They have to. You cannot use these codes
)
14 intelligently unless you know what's inside them.
I 15 They understand how to put together the input 16 decks. They know how to use the various opf:ons, some of 17 which are very arcane, some of which are buried deep i.
~
18 inside the code. They also know when the code is giving 19 them good results and when it's giving them garbage.
20 That's because they use it a lot. They have to use it a 1
21 lot. If they don't use it a lot, they don't know how to a
3 22 use it. They just don't do a good job.
4 23 These people are rare. They're expensive.
4 i 24 They take a long time to develop. And you never want to 4
() 25 lose them because they're valuable. Unfortunately, as I i
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 3
1323 RHODE ISLAND AVE., N.W.
j (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
_ - . _ . . . _ . . - . . _ _ . _ _ . _ _ . . . _ . _ . ~ . _ . _ _ _ _ _ . _ . . . _ _ _ . _ _ . _ . _ . . _ _ _ . _ . _ . . . _
- 484 i.
1 said, they're rare.
2 The second class of code user is what I call-
- ,O
- . 3 the occasional user, the knowledgeable engineer, some 4 would say the unsophisticated engineer. This is the l
l 5 person who knows systems in thermal hydraulics, who knows I
l 6 the' basics of the codes, but does not run them on a
! 7 regular basis. These are the engineers in the Reactor 8 Systems Branch who get the phone call on Monday morning.
9 that: some plant had an event where they drained a system 10 down and he wants to know how long would it take to drain 11 a reactor vessel if you opened up this valve.
12 These people have a valid need to perform 13 simple thermal hydraulic analyses on a short-term basis.
14 Right now they can't because they don't have the 15 experience with the codes as they sit right-now to perform 16 these analyses.
17 Third class of users is simulated users. I 18 call them system analysts not trained in thermal 19 hydraulics. They need a simulator. Okay? They're not 20 going to build the input models. They're not going to 21 understand all the code options. They want to play "What 22 if?" games. They want to sit down in front of a 23 simulator, push the buttons, open the valves, shut the 24 pump on, shut the pump off, turn the pump off, turn the
) 25 pump-on, run these scenarios just like a simulator.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433 d
, . . _ _ . _ . _ . _ ~ _ . . _ . _ _ . _ . _ . _ _ _ _ _ . . _ _ . _ . _ _ . . _ _ . _ . . . . _ _ . _ . . _ . _
o 4 485 e
1 This is probably the largest growing group of 2 users. Okay?. For PRA practitioners, for training, O~ 3 procedural development, and then throw a management in 4 there. Some might consider them to be Class 4 users, but 5 we'll throw them in Class 3.
6 We need codes that satisfy the needs-of these 7 two groups of users because the needs are growing largest 8 in these areas.
9 (Slide.)
10 MR. CARUSO: What are our regulatory'needs?
11 I'm going to use two words in this list here, "should" and 12 "must." "Should" means it would be nice to have and I i l
13 would really like it. "Must" means it must be done. '
14 Okay?
15 First of all, the code must be robust. What 16 do I mean by " robust"? Robust means that you can run the 17 code, and you don't have to nurse it. Right now the 18 current generation of codes, Class 1 user sits there, runs 19 it for an hour, looks at the result, checks to see how far 20 the time step has decreased, adjust a factor here, adjust 21 a factor there.
22 Occasionally the code bombs. He has to 23 restart it. That's not a robust code. Robust code means 24 you start it, and it runs to completion and you get the
-O(,) 25 results. It must be defensible by the code author.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
-- - ~ - . - . - . . . - . - . - . . . - . - -
486 1 I'm not' going to defend the internals of-the l
2 code. I' don't defend the internals of my word processor.
O 3 I don't defend the internals of my Fortran compiler. I 4 don't want to defend the internals of my thermal hydraulic 5 code. Okay?
6- I've got to defend the input model that I 7 create. I've got to defend my choice of the code, but I 8 don't want to defend the intern 01s. That's why we've got 9 an Office of Research.
-10 The code should be easy to use. It should be 11 simple and easy to develop and modify plant models. Now, 12 if we had the internet here, I have a hot link here to j 13 show you what a RELAP5 input deck looks like. I'm sorry I
, )
~- 14 can't show it to you.
15 How many of you have ever seen a RELAPS input 16 deck?
17 (.o N response.)
18 MR. CARUSO: Okay. Well, for those of you who 19 haven't seen it, it's line after line, 80-column. A j 20 RELAPS input deck is actually a deck, a deck, a deck of 21 cards, computer cards, 80-column punch card images. Okay?
22 And it's number, number, number, number, number, with an 23 occasional comment, where the individual who built the 24 deck decided to put in a comment.
(- 25 It's inscrutable. If you don't know the 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
487 1 you can't make any sense out of it whatsoever. Okay?
,-~ 2 They're put together by hand. Individuals sitting down at V
3 a keypunch, nowadays a terminal, type them up one by one, 4 card by card. They have to check individually that erch l 5 pipe attaches to another pipe at the right elevation, at l 6 the right location. They have to individually input the 7 diameters of the pipe, friction factors. ,
l 8 All this stuff is done individually by hand. ,
1 1
9 It makes it very difficult to use.
10 MEMBER CATTON: Just let me interject for a l
11 moment. There are some horror stories about these decks 12 as well. Apparently I recollect I guess maybe ten years 13 ago where there were contracts in place to develop decks. l
(,,) l x,_,1 14 It turns out some have disappeared, some were unusable.
15 And what's the cost of one of these input 16 decks?
17 MR. CARUSO: If you wanted to start from 18 scratch, probably on the order of half a million to 19 three-quarters of a million dollars to build a deck for 20 one plant.
21 MEMBER CATTON: And we've actually lost some.
22 MR. CARUSO: Over the years as the codes have l 23 changed, old decks no longer run with new versions of the 24 code. So I've had a long-term project in place to take
.A
( ),
25 old decks and upgrade them to run with the newer versions l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND A'K., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
- 7. _. _ . -
! 488 1 of the code. And_we have been spending money on this.
I 2' It's very important that new versions of the
!O j 3 codes be able to read old decks. The horror story that I 4 have-is that I spent $300,000 for a series of decks. Two 5 weeks after they were delivered, a new version of the code 6 came out th'at would not run those decks. And I had to 7 spend another $30,000 to upgrade those decks.
8 VICE CHAIRMAN SEALE: Could I ask --
9 MR. CARUSO: And we don't want that to happen 10 again.
11 VICE CHAIRMAN SEALE: Are there places where 12 the older versions of the code are still available on a 13 machine so that you could run the decks in the framework 14 that they were originally intended for?
15 MR. CARUSOi Well, the problem is that if you 16 go far enough back, you get the versions of the codes that 17 don't run on current computers. You have to go find 18 yourself a --
l 19 VICE CHAIRMAN SEALE: My point is that after 20 you go through all of this agony of making the change, it 21 would then be nice to compare what you got with the old 22 code version with shat you get with the new version --
23 MR. CARUSO: Yes.
24 VICE CHAIRMAN SEALE: -- just to be sure that 25 all of that slogging around that you did didn't create NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005 3701 (202) 234-4433
489 1 another problem or two.
fm 2 MR. CARUSO: You're right. It would be nice.
1 3 Unfortunately, some of these runs take so long that it's 1
4 not worthwhile to run those comparisons. j l
5 MEMBER CATTON: What I'm to generate here is a 6 commercial for standardization. I think you've got to lay I 4
7 down standards for input / output for your codes, and then 8 you've got to stick with them. That's not been done in 1
9 the past, and you can see the result.
10 MR. CARUSO: Okay. Useful, scrutable output.
11 The output from these ceias is 132 columns wide. Is that 12 familiar to anyone? Width of standard computer output !
a 13 paper. It's line after line after line of numbers. It is (3
- V 14 also inscrutable to anyone except a Class 1 user. Okay? l 15 We have in the past five or six or maybe ten
!1 16 years developed some programs which can read those 17 outputs, then produce plots. But that software is 18 stand-alone software. It's cobbled together. It's 19 difficult to use. It's just not user-friendly.
20 The code must be platform-independent, must be 1
21 platform-independent. Okay? Then I bring this up because 1 1
22 the codes were originally designed to run on 23 supercomputers with small memories. That means that they l J
24 used a lot of clever programming techniques, byte packing,
, ~.
(,) 25 container arrays, devices to fit inside small memories, 1
il
, 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
490 l l
1 which'are just a hindrance to running these codes on 2 modern computers.
3' Code developers should not take advantage of 4 particular hardware architectures in designing the' codes.
l 5 This week I'm buying DEC alphas. Next week I may buy 6 RS6000s. The week after I may buy PC platforms. The 7 codes have to be able to run on all these different 8 platforms.
9 MEMBER CATTON: Without recompilation?-
10 MR. CARUSO: Oh, you have to recompile them. l 11' Okay? But don't -- l 12 MEMBER CATTON: Just checking.
13 MR. CARUSO: Don't assume that a word size is I
[
\
14 60 bits. Okay? That's a big problem that's occurred.
15 Codes should be fast-running. In this matter, sometimes
'16 it's easier to buy a new computer, buy a faster computer, 17 and you get the performance that you need.
18 The code should be linkable to other codes.
19 This is a problem or a feature that we have identified as 20 part of the advanced reactor reviews where containment 21 performance was linked very closely to reactor system 22 performance. And it would be very nice if you could have 23 an independent containment code talking to a reactor 24 system code, exchanging information about the boundary 25 conditions where fluids flow between the two.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4 33 t~
491 1 So far this has proven to be very difficult to implement with the current generation of codes.
,q 2
%) 3 MEMBER POWERS: If you're running a' consistent 4 thermal. hydraulic code, same code, have you considered
- 5. going to these multi-processor units, customized 6 multi-processor units?
7 MR. CARUSO: That's a possibility. I mean, 8 there would be some cases where it would be nice to have a 9 four-processor unit with CONTAIN running on one processor 10 and RELAP running on another process and having them talk l
11 to one another, a coarse grain nodalization. It's a 12 possibility.
13 MEMBER POWERS: I was thinking in terms of 14 these things that have like 1,000 processors --
15 MR. CARUSO: Ah, yes.
16 MEMBER POWERS: -- and massive parallel 17 computing. Certainly the thermal hydraulics they use for 18 our weapons programs now have gone to that kind of thing.
19 MR. CARUSO: I have mixed feelings about that.
20 On the one hand, it would be nice to be able to run stuff 21 very quickly, but, on the other hand, if Research develops l 22 a code that only runs on a 1,000-unit, 1,000-node 23 multi-processor, then that means I'm stuck with that 24 hardware. Okay? It must be platform-independent. I 25 MEMBER CATTON: That can be handled through NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
492
\
1 compilations many times.
f
,- 2 MR. CARUSO: Okay. If there are Fortran i'") 3 compilers for the 1,000-processor unit that allow me to 4 take an existing code and run it there and it runs fast, 5 then fine, I'm willing to use that. But I'm not willing 6 to let someone develop coding that is specific to one 7 particular machine architecture. That's what got us into 8 this mess in the first place. Okay?
9 The code should have the ability to range 10 various parameters. We're always going to have the Class 11 1 users around. Some of the Class 2 users might want to 12 be able to play with the code, play with the internals, 13 vary some heat transfer correlations slightly. This is
,/ -
\-) 14 very useful for --
15 MEMBER CATTON: This is important if you have 16 to make an uncertainty evaluation.
17 MR. CARUSO: Yes, exactly. The code 18 documentation must identify the limits of applicability of 19 the models and correlations as well as the uncertainty 20 associated with those.
21 And, actually, this is a subunit of something 22 on my next slide. I'll get into this a little bit more.
23 (Slide.)
24 MR. CARUSO: So what recommendations do I have (h
( ,
) 25 for code development? The first one, work closely with NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
493 i:
l' the user community. ~ Code developers should talk'to the 2 users. They have lots of good ideas. People that use i -
1 3 these codes frequently understand them better than the I
4 people who develop them.
i 5' And in this regard, Research and the rest of i
j -6 the NRC user' community have agreed to establish a working ;
i 7 group to exchange information and ideas about how we use i.
- 8 the codes and what our needs are.
, 9 We're working with our Research right now to 10 develop the details of how this is going to work, but we 11 feel that'we have established a good working relationship
- . 12 with them. We're going to continue it. So I just l'
{ 13 encourage them to keep talking to them.
\
14 Look at other. engineering software. Do not e
I 15 reinvent the wheel. There's lots of other engineering i
} 16 software that has great user interfaces. Buildings were
! 17 designed this way. Buildings are designed this way.
18 There are commercial thermal hydraulic packages. Finite l
l' 19 element software, CFD codes have great front ends and back i
l 20 ends. Allow the users to do productive work. Look at how 3
l 21 they do their jobs, t
j- 22 Document, document, document. Document, 23 document, document. Document, document, document.
j 24 MEMBER FONTANA: In English.
25 MR. CARUSO: Write it down.
j NEAL R. GROSS i
COURT REPORTERS AND TRANSCRIBERS
! 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
494 1 MEMBER FONTANA: In English.
s 2 MR. CARUSO: Okay? This has been a big i \
s
'~'/
3 problem in the past. With the internet and electronic 4 documents, there's no reason not to keep everything online 5 available to everyone, the developers and the users.
6 Write it down. Okay?
7 VICE CHAIRMAN SEALE: And if it's 8 understandable, that's even better.
9 (Laughter.)
10 MR. CARUSO: If you make it available to 11 people, they will ask questions. You'll get feedback, and 12 it will get clarified. That's why you've got to write it 13 down and make it available. It's not enough to just write q
-/ 14 it down and put it in your file cabinet. Write it down.
15 Put it up electronically on the internet. Okay?
16 That's where this suggestion comes from.
17 Gather the user suggestions and hints into an electronic 18 document that's available to all the users. Okay?
19 Provide full disclosure of all code 20 limitations, big complaint I have. Lots of code 21 developers know that there are limitations in their codes, 22 but they don't want to release these dirty little secrets.
23 Okay?
24 Be open. Tell us where the limitations are.
C 'T
' ,_,/ 25 Tell us what the problems are. The users will give you an NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
495 1 idea on how to fix it if you let people know where the 2 problems are.
3 Finally, the code developers should be' forced, 4 forced if necessary, to use their own product. Okay? One 5 regulatory analysis at least a year. What I mean by 6 " regulatory analysis" is do an actual problem. Come to 7 me. I have plenty of work to do. I'll give you a 8 problem, a comparison of a vendor calculation or an events 9~ analysis, have you use your code to perform an actual 10 analysis. Best way to figure out whether you're doing a 11 good job is to use your own product. Okay?
12 With that, I'll open myself up to any 13 questions.
14 MEMBER CATTON: Any questions for Mr. Caruso?
15 MEMBER POWERS: Have you given an estimate on 16 for a typical code that you're interested in, what your 17 things that you mandatorily require here, either your 18 recommendations would cost as far as manpower, number of 19 FTEs that it would take to do your regulatory analysis 20 each year to provide full disclosure to document, 21 document, document to gather user suggestions?
22 MR. CARUSO: I really haven't thought about it 23 independently. I've seen the estimates that the five-year 24 plan has. And they look very reasonable to accomplish
$ 25 this.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C.. 20005-3701 (202) 234-4433
. _ _ . _ ._. m . . _ _ _ __ .__ m _ ,
496 1 MEMBER POWERS: For the codes that I have some l- 2 familiarity'with, which are not thermal hydraulic codes,
)
this looks like a man-year effort to me just to meet your 3
, 4 recommendations. That's no development. That's just the l
l 5 things that you're requiring there.
- 6 MR. CARUSO
- Okay, I haven't gotten into the l
7 resource estimate for what this would take.
8 MEMBER CATTON: But, Dana, if you were to I l
t
! 9 standardize a lot of these things, I think you could find 10 that other than the core of the code itself, a lot of this 11 stuff is available.
l 12 MEMBER POWERS: I don't know what you mean, l
13 Ivan. He's saying do a regulatory analysis each year to 14' gather user suggestions. It's one of the --
l 15 MR. CARUSO: No, that isn't what I mean.
16 Excuse me.
l 17 MEMBER CATTON: This is to force the developer l 18 to have some --
l 19 MR. CARUSO: I'm sorry.
20 MEMBER CATTON: -- respect for the needs of 21 the user, which has not been common in the past.
22 MR. CARUSO: I'm sorry. I didn't understand g 23 what you were saying. This one here, " Gather the users' 24 suggestions"?
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 l
- - . _ ~ _
. . - . - -. - _.-. - - - - _ . _ - . . . . . - - ~. - . ~ . .
- 497 1 MR. CARUSO
- That really is provide an open 2 forum where users can provide comments to a bulletin O. 3 board, for example, that are always available. That's not 4 something that requires any ongoing --
5 MEMBER CATTON: A good example of this --
6
'l MEMBER POWERS: It does.
l 1
7 MEMBER CATTON: -- is LYNX. If you look onto ,
8 the internet for LYNX, what you'll find is that all over 9 the world people are developing things for this new code.
10 And they have just such an electronic document. You can 1
11 go into it and find the trouble somebody has with this, 12 that, or the other, --
i 13 MR. CARUSO: Right.
1 1 14 MEMBER CATTON: -- clues as to what to do to 15 circumvent a particular kind of problem. It's really a I 16 very valuable asset, particularly if it's indexed.
17 MEMBER POWERS: But, Ivan, .I can't' imagine --
18 MEMBER CATTON: And it doesn't cost.
19 MEMBER POWERS: -- anything that would be more 20 effete than to provide a forum for user comments and then 21 ignore it.
22 MEMBER CATTON: Well, that happens, too, you 23 see.
24 MEMBER POWERS: But if you're going to provide 25 a forum for user feedback and then turn around and make NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
]
498 i
! 1 use of.it, they. start to chewing up man-hours. And, boy, l 2. user interfaces chew up man-hours. It chews up man-hours 3 like you will not believe.
4 MEMBER CATTON: But if you don't do it right,.
- 5 nobody will use it. Why bother? And I think that's one 6 end of'the --
7 MEMBER POWERS: I don't disagree with you. I f 8 mean, it's not the concept that I'm disagreeing with. It 9 is recognize that there is a resource requirement here and 9
10 that you've got to factor that in.
11 MEMBER CATTON: Oh, that's right. I wouldn't l
} l 12 disagree. My feeling is that some of the things that Mr.
- 1
- 13 Caruso has talked about are probably more important'than 'I
- r I
14 the enhancement of the code itself.
1 J
l 15 MEMBER POWERS: -Could well be true, s
16 MEMBER CATTON: And if you --
! 17 MEMBER POWERS: But understand that it's also 18 as manpower-intensive as --
19 MEMBER CATTON: That's probably right, 20 probably right.
21 MR. CARUSO: I'll give you --
22 MEMBER CATTON: We have 20 years of history of 23 what happens if you don't.
24 MEMBER POWERS: Well, I have 20 years of
( 25 history of what'happens if you say, "And do this in your NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
499 1 spare time."
l p]
i 2 MEMBER CATTON: I guess I'm at a university.
3 MR. CARUSO: I guess I would make the I
4 4 observation that right now with the current user interface 1
1 5 and the current code structures, I think we spend a .
l 6 tremendous amount of time working around them and doing a !
7 lot of things that are not productive.
8 And I realize that if you build a fancy GUI, !
l l
9 the code users are going to have to -- or the code l l
10 developers are going to spend time supporting the users 11 more time than they otherwise would. But right now the l 12 code users are spending a lot of their time typing in 13 80-column card images and tracing down typographical
( ' '
)
14 errors in massive numbers that are literally 15 incomprehensible.
16 MR. HODGES: This is Wayne Hodges. I'm the 17 Director of Systems Technology in the Office of Research.
18 I'd like to make a couple of comments if I 19 could.
20 MEMBER CATTON: Sure.
21 MR. HODGES: First off, I'm not going to try 22 to defend what's happened over many years. I can't. But 23 Dr. Eltawila and I have been in our current jobs for a 24 little over two years now. And during that two-year time p
, (_) 25 we have been listening very closely to the users. That NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISt.AND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
500 f
1 has been a major input into a thermal hydraulic research 2 plant.
3 Most of the issues that Ralph.Caruso has j 4 mentioned today, we are already planning, we have in the 3
5 planning or we are already doing. So when he says,
. 6 " Document. Document," that's in our plan as to how to do i
7 that. It'doesn't just mean also what gets at a user. It 8 means the code itself needs.to be documented.
9 There's a lot of things that go into that. We i
10 have given it a lot of thought to that. We are working 1
, 11 very closely with the users to see what they need. I am 12 and have for the last two years required that the people l 13 in my division run the codes that they're using.
14 I think that's the main thing that Ralph is 15 talking about, that they are familiar with the code and
~
16 the problems and the kinds of problems that users have.
i.
4 j 17 That-is a requirement I've had for the last two years. )
3 18 It's not just one analysis per year.
19 It's to be using the codes. I think we're 20 going a long way towards satisfying the kinds of things 21 that he's listed there. I don't disagree with the points 22 he's listed, but I don't want the impression to be left l 23 that we're not doing them. We are doing them.
24 What you will see is a group who is actually
() 25 performing the kinds of things that he's already talking NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
. (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
501 1 about. l
,f 3 2 MEMBER CATTON: I don't think he said you l
\ I v
3 weren't.
1 4 MR. CARUSO: I would support Wayne in that. I 5 Wayne has been very responsive. And I think we are well 6 on the way to the development of some very useful tools 7 here.
8 MEMBER CATTON: Very good. Any other 9 questions for Mr. Caruso?
10 (No response.)
11 MEMBER CATTON: Seeing none, I'd just like to 12 mention that Mr. Caruso came in here special for this 13 meeting. He was supposed to be on a vacation. And I O) t
'v 14 appreciate it very much.
15 MEMBER POWERS: Talking to us is better than 16 any vacation; right? Being up here is better than any 17 vacation.
18 MEMBER CATTON: He'll feel so good when he 19 walks out of here.
20 (Slide.)
21 DR. SLTAWILA: Good morning. My name is 22 Farouk Eltawila. I'm the Chief of the Reactor and Plant 23 System Branch.
24 I'm going to give the presentation about our
(,./ 25 five-year plan. And Joe Kelly and Dave Bessette 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
502 1 sitting on the staff table here, and they will be helping 2 me responding to any question you might have.
3 Before I.go through my presentation, I would 4 like to say that I'm extremely disappointed with Professor 5 Catton's opening remark by indicating that the effort is 6 not worthwhile and speculative --
7 MEMBER CATTON: Excuse me. The comment about 8 not worthwhile was if certain things can't be done, it's 9 not worthwhile. I also said I'm very pleased with what 10 you have on the table.
11 DR. ELTAWILA: Okay, i 12 MEMBER CATTON: But, again, if modularity and 13 flexibility are not attainable, then that's a different O
14 picture.
15 DR. ELTAWILA: If the Committee has to take 16 into consideration whether to keep the status quo as is 17 and to spend 50 percent of our budget in maintaining the 18 existing core'or try to improve the capability into a 19 single code, that's the most important thing. And I'm 20 going to go through that in my presentation. And I would 21 like you to pay attention to that. I thought that we 22 should have solved this issue at the subcommittee level.
23 The other thing that I think everybody refused 24 to acknowledge, that the Office of Research is the biggest 25 user of the codes. We have been developing the input deck NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4 433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
. _ _ _ = - _ . _ - _ _ . _ _ . . _ . ~ . . . _ _ . . _ . . . _ - . _ _ _ _ . _ . _ _ . _
503 i 1 for the AP600, for the SBWR. We have been running 2 analysis for the SBWR and for the AP600. So we are user, O 3 and we are talking to the user.
f
- 4 The third point, documentation, documentation, i
j 5 documentation, already exists. If you open the internet, j 6 you can open RELAPS home page. It has a documentation l
- 7 more than any other code that I'm aware of right now. So i
e a we are using the technology. We improve the technology.
{ 9 So, with that, I'd like to go over my i
- , 10 presentation.
- 11 (Slide.)
12 DR. ELTAWILA: As I indicated, Joe Kelly and
}
! 13 Dave Bessette are here to help me. This plan has been 3
14 developed in response to Chairman Jackson's request that
=I j 15 the staff develop a five-year research plan to_ demonstrate i
i 16 how we are going to be maintaining capability to provide j 17 the Commission with the tools needed to address thermal j
j 18 hydraulic issues in the future.
19 So I outline my presentation here to give you 20 overall goals of the five years' plan to meet Chairman 21 Jackson's requests. I'm going to discuss with you the 1
22 current' situation on the thermal hydraulic code with some j 23 historical perspective and define the goals that we need 24 the code to achieve in order to have a modern tool.
() 25 Our approach is consolidation of the existing NEAL R. GFU3SS COURT REPORTERS AND TRANSCRIBER 3 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
504 1 code into a single code. I'm going to discuss briefly if 7s 2 I'm allowed the thermal hydraulic test program and our
( \
i /
3 buildup of in-house capability. And I will complete, have 4 some concluding remark.
5 (Slide.)
6 DR. ELTAWILA: The overall goal of the 7 research plan as defined in the paper, the Commission 8 paper, that you have is to maintain a core competency in 9 the thermal hydraulic, reactor physics, and plant 10 transient codes to support the regulator and to have the 11 capability to be interacting with the people in the 12 international community.
13 (Slide.)
/ \
- 1
\/ 14 DR. ELTAWILA: In order to do that, we plan to 15 improve and maintain reactor safety codes -- and you can 16 see that going through these yellow boxes here -- to 17 consist of code consolidation, code improvement, 18 uncertainty quantification, and code assessment, and a 19 graphical user interface. Incidentally, we have started 20 the development of a graphical user interface. The first 21 prototype will be available in June of 1997.
22 We plan to perform fundamental and the small 23 experiment to improve the two-phase flow formulation. And 24 that you can see here. We'll be performing these through
/%
) 25 an experimental program to provide the needed model to 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 505 1
1 improve the code. ;
I 2 We are planning to maintain experimental 3 facilities, the three experimental facilities that 4
4 Professor Catton talked about, so we can run an integral 5 facility or combine -- integral tests or combine tests to 6- provide the information needed for code assessment, i
7 And since the staff is going to be an integral 8 part of all this process where they ure going to be 9 working with the contractor or the professor that they are 10 running these experiments, that will lead automatically l 1
[ 11 into a buildup of in-house capabilities.
12 (Slide.)
13 DR. ELTAWILA: Just to put things in the right 14 perspective here, if you look at' Fiscal Year 1997 budget, 15 we found that we are spending 49 percent of the budget on 16 maintenance of the code. l l
17 I can't do anything about it. We have four l
18 codes. We have to maintain them. We have to keep them up 4
19 to date. I cannot touch that. So if you look at the 20 rest, since I'm going to focus most of my presentation on 21 the experiment and on the code consolidation, I just want 22 to show you that we have a balanced approach here.
23 If you subtract the 49, if you look at that we 24 have close to 30 percent of our budget or more is going
() 25 toward experimental, fundamental experiment and integral NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2M33 WASHINGTON, D C. 20005-3701 (202) 234-4433
,.-_ _ _ _ . , . __ _ __ _ _ _ _ - ~ , _ - _ , .
506
'l experiment, to improve the knowledge in thermal 2 hydraulics,'and about 10 percent may be going for CFD and 3 the consolidation of the code.
4 (Slide.)
5 DR. ELTAWILA: Currently, just for those 6 members-who are not aware of the different codes that we 7 have, we are maintaining four codes right now. We are 8 maintaining the RELAPS code for a small-break LOCA in the 9 transient; the TRAC-PF code for large-break LOCA; the 10 TRAC-BF for BWR LOCA transient; and the RAMONA code, which 11' couples thermal hydraulic and neutronic codes.
12 When you come down to it, you will find that 13 the RELAPS code and the TRAC-PF_ code duplicate each other 14 in most all aspects of it. Same thing here between 15 TRAC-BF and RAMONA. They duplicate each other. ;
16 These duplications are not necessarily at this 17 time. It was intended in the task to have the labs 18 compete with each other so you improve things. As it 19 happened, as the budget went down, both codes' budgets 20 went down, _and we did improve any of them.
21 As you heard before, these codes had been
'22 developed in the early '70s. And we have been trying to 23 convert these codes into the new work station environment.
24 However, that was done in a very ad hoc
() 25 fashion, which resulted that the code developed or put a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE.. N.W.
(202) 234 4 433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
507 1 lot of creative ideas in this conversion. And these 7- 2 creative ideas are causing us problems right now that
(
x_)
3 every time we exercise the code in a different kind of 4 application, we find that the code is not robust, is 5 failing, and we are spending a lot of resources trying to 6 fix that.
7 In the past also the error corrections that 8 were discovered, if they never went to the root cause of 9 the problem, they tried to fix them, just the problem and 10 not the root cause itself.
11 (Slide.)
12 DR. ELTAWILA: If you look back at the 13 accident at TMI and what happened after that and we were
/N i ks 14 not able to give the code to analyze the accident, it was 15 recognized at that time that the transient itself has so 16 many phenomena, much more than what was modeled in the l l
17 code at this time. I 18 So as a result of that, it was identified 19 that, for example, some of the phenomena, that they were 20 not modeled by core heat up, pressurizer discharge, 21 secondary side voiding and refilling, and the interacting 22 operator actions.
23 All of these things were not modeled in the 24 code, which caused the code to fail. And, as a result of
(_ 25 the TMI accident, a flurry of national and international NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
508 1 programs started for code development and experimental 2 programs.
,'~'!
3 What is the result of that extensive work that 4 happened in the '80s were the codes that we have right 5 now. And there were a lot of models and correlations that 6 they did not have an adequate database for them. So there 7 are a lot of ad hoc models in the code.
8 Even with these ad hoc models and the 9 limitation in the numerical techniques, these codes are 10 capable of describing the governing phenomena and general 11 system behavior of a large variety of accident scenarios.
12 (Slide.)
13 DR. ELTAWILA: In conjunction with the EY 14 national and international cooperation, OECD, Organization ;
15 for international Economic Cooperation and Development, in 16 the past 10 to 15 years has been running international I 17 standard problems. These standard problems were intended 18 to compare the code prediction of experimental data.
19 In July of this year, a final draft report is 1
20 issued by OECD about the lessons learned from these past i 1
21 ten years of history of ISP. The report concludes that 22 experienced users are able to get the relevant phenomena, l 23 even in cases when the complex scenario is involved in the 24 transient.
p)
(, 25 If you look at a single version of the codes, NEAL R. GROSS CO'5T REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
509 1 you find that -- and we are talking about all codes, not fs 2 NRC codes only -- a wide ru..ge of results are obtained 3 using the same versions of code. That highlights all the 4 models that are available in these codes and the user 5 effect.
6 MEMBER FONTANA: I don't quite understand the 7 second bullet, " Experienced Users Are Able to Get the 8 Relevant Phenomena." You mean the phenomena was in the 9 code and it's hard to use the models properly?
10 DR. ELTAWILA: I think that they have a lot of 11 things that they understand about the code itself to make l 12 it, to force it to produce some of these results.
13 MEMBER FONTANA: Okay.
r^% ]
(' -)
14 DR. ELTAWILA: So there was some good news and 15 some bad news. But, again, this report that was issued 16 this year still identified a lot of deficiencies. We 17 still are unable to predict the break flow, stratification I l
18 in hot and cold leg, mass distributio.as in the primary and !
l 19 the secondary sides, mixing in the downcomer, steam l 20 condensation, pressurizer behavior, core uncovery and l J
21 heat-up, CCFL in steam generator plena, and low-pressure 22 period.
23 This was based on old versions of the code, !
24 not the one that we have developed for AP600 because that
(s
(_,/ 25 issue has been resolved right now.
1 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 510 1 CHAIRMAN KRESS: As I look at those 7-s 2 deficiencies, most of them are because you're talking V 3 about a 1-D lumped parameter code.
4 DR. ELTAWILA: Most of them. Yes, you're 5 right about that. That's correct. Well, break flow --
6 CHAIRMAN KRESS: Break flow ycu could probably 7 do something.
l 8 DR. ELTAWILA: Yes. Most of the other areas 9 are definitely a two-dimensional phenomenon, multi-D 10 phenomenon.
11 (Slide.)
12 DR. ELTAWILA: The other challenge to our ,
l l
13 existing code in the U.S. was the Lasalle incident. When I
\
' -) 14 we tried to analyze the incident itself, the code failed 15 to predict the instability issue.
16 A follow-up activity of that indicated that 17 the existing codes right now which have inherent numerical 18 damping will result in making the code more stable than 19 the extra plant. So these artificial viscosity terms and 20 things like that are causing the code to be more stable.
21 So we were unable to predict this phenomenon.
22 The report that was issued by Brookhaven 23 National Laboratory as a result of that assessment 14 indicated that a computer simulation with 3-D neutronic (3 25 capability is indispensable for understanding the local
( ,/
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
511 1 out-of-phase oscillation analyses. So this is another
,, 3 2 limitation of a 1-D code.
3 (Slide.)
4 DR. ELTAWILA: In addition to this limitation, 5 we find that there are other code limitations. For 6 example, as Ralph indicated, the computer speed has 7 significantly increased. The cost of computers themselves 8 went down drastically. I think you can come by a work 9 station very cheaply nowadays.
10 But if you look at the actual cost that we are 11 paying to perform this analysis, it is very high. And the 12 main reason for that, the complexity of the code, the 13 duration of the transient, -- we're dealing with
14 transients that now take over 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />, you know -- and 15 difficulty in interpreting the results. The codes are 16 very complex, and the output is not very friendly. So it 17 takes a long time to try to interpret the results.
18 As I indicated, there are limitations in the 19 physical model. And there are some models that are being 20 used outside their database. And there are limitations in 21 the numerical method itself.
22 (Slide.)
23 DR. ELTAWILA: So what do we do about this 24 situation? I think what we decided to do is to look at
(_) 25 how we can make our codes state of the art. So we need 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
i 512 1 have a code that's state of the art that is capable of 2 performing accident and transient analysis for existing O 3 plants as well as advanced lightwater reactor.
4 It must have capabilities suited to address l
S safety significance issues that are posing challenges to i 6 existing code. So it must have a coupled containment and 7 thermal hydraulic capability. It must have coupled 8 thermal. hydraulic and neutronic capabilities.
9 MEMBER FONTANA: Multi-spatial, too?
10 Neutronics would be at this point? l 11 DR. ELTAWILA: We would like to have a 3-D j 12 neutronic as well as point kinetics, too, to check the 13 results of the code.
14 (Slide.)
15 DR._ELTAWILA: Again, we want the code to-be 16 easy to modify to incorporate new models to address a 17 specific issue of interest. It must have online user 18 assistance to be able to report errors, to modify the 19 code, and it must have a graphical user interface, must be 20 easy to use, run fast, and accurate and robust.
21 It must have smaller incremental costs in the 22 future for performing analysis and maintenance. We cannot 23 keep spending a few million dollars every year on code 24 maintenance alone. And we have reached the conclusion 25 that we do not need four codes to do the work that we 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
.-_ -_ , , - =-.
513 1 required to do.
,- 2 We believe that a single code will be much N'~'I 3 better for us, the first thing that we will have all the 4 staff trained to use one code. So that will save us in 5 having different staff members trained in using different 6 codes. And there are codes available through our 7 international cooperation that we can use them and run 8 them to understand the physical behavior and as a backup 9 of our codes.
10 CHAIRMAN KRESS: Farouk, do you see any need 11 for a code like this to be linkable to a fission product 12 and chemistry transport code, like VICTORIA, for example, 13 or is that going beyond the need for --
\ s# 14 DR. ELTAWILA: Ideally that's the situation 15 that you want to do, that you want to be able to link the 16 new code with a code like SCDAP and eventually VICTORIA 17 and so on. But, as things happen, right now the SCDAP 18 code is completely intermingled with the RELAP5 code.
19 So it would not be a simple link. So we might 20 have to develop a complete package, SCDAP package, that 21 will be able to be interlinked with the code. But 22 definitely ideally that's the situation that you want to 23 have.
24 So what is the approach that we would like to (3
(,,) 25 use to reach this objective here? We are going to use an NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
514 1 approach that's almost identical to the CSAU methodology.
,- - - 2 So our approach right now -- I'm talking about
\J 3 the code area -- is one to grade and upgrade the existing 4 capability into a single code. And it will consist of 5 three elements. I'm going to go over them in details.
6 (Slide.)
7 DR. ELTAWILA: The first element, to define 8 the code requirement capability. You go then through the 9 consolidation phase. And then you go through assessment l 10 and parameter ranging. In parallel with that, we are l 11 going to start pilot studies to explore advanced concepts 1
12 that could lead to improved codes. j 1
13 (Slide.)
O
's #
14 DR. ELTAWILA: So if you direct your attention l 15 to the figure here, the first element of the program is i 16 Step One here of Element 1, to give the phenomenon 17 identification and ranking table.
18 So what we need to do is to identify all of 19 the scenarios that we want the code to identify, whether 20 it's large-break and small-break LOCA, station blackout, 21 steam generator tube rupture, ATWS, and all the scenarios 22 that we want the code to use.
23 We identify the plant design that we would be 24 looking at. We are going to be looking at operating PWRs
(%
(_) 25 and BWRs, and we are going to be looking at advanced water NEAL R. GROSS COURT REPORTERS AND TRANSCRIBE 9S 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
l
.] 515 1 reactors, like AP600. Then we are going to either modify i
ew 2 or look at the existing parts.
f V'
j- 3 We have maybe about 50 percent of PIRTs, 100 4 percent, of course, for ALWR, but we don't have enough for 5 boiling water reactor and not as much for PWR. So we will l' 6 have to develop a phenomena identification and ranking 4
7 table for these scenarios for some of the operating i
8 plants. And for that we will come up with the 9 high-ranking phenomena'that need to be in the code.
t l
l 10 (Slide.)
! 11 DR. ELTAWILA: In Step Two, as I indicated, we i
12 are going to implement a model data structure and a
13 architecture in the TRAC-P code. We are not going into.
l [\-
- 14 that step blindly. I think we have an adequate i
4- 15 understanding right now that the TRAC-P is amenable to i 16 give us the tool that we want.
i 17 So we feel that with the modernization effort
[
, 18 that's been sponsored by NRC and CAPL at LANL. We will 19 have at least a modern structure here of using Fortran 90.
20 In conjunction with that, we are going to be l_ 21 identifying options and requirements for modularity. It 1
4 J22 might not be the Cadillac of modularity, but we will have 23 a certain level of modularity in the code to satisfy our 24 needs.
25 Our need is to be able to modify the code, if 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
516 1 you want to add a new model that you can add it to the f- 2 code. And by the end of this phase, we would be able or
() 3 have the code requirement. And we are confident that we 4 would be able to incorporate all this into a TRAC ++, which 5 will have a modern architecture and would have a large o degree of modularity in it. )
1 7 (Slide.)
1 8 DR. ELTAWILA: In Step Three of Element 1, we I l
9 are going to look at the documentation of the TRAC-P, the 10 TRAC-B, the RAMONA, and RAMONA. And we would be looking 11 at the code manuals, the user guides, the programmer 12 guide, any developmental assessments that were done to the i
13 code, and looking at the model and correlation.
/T l
l
\ l l
's / 14 From these and using the PIRT, which 1 15 identified the high-ranking phenomena, we will identify 16 and select the model that needed to be incorporated into 17 the consolidated code. We're going to do that in four l
18 steps here.
19 (Slide.)
20 DR. ELTAWILA: In Step One, we are going to 21 recover the BWR TRAC-B capabilities. So we are going to 22 add all the BWR-specific mc.dels to the base TRAC-P code.
23 We know that there are mode.ls that are duplicative in both 24 versions of the code.
O
( ,/ 25 So we are going to use a peer review process NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N,W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
517 1 to assess the parameter range over which each model, the s 2 competing models, are overlapping. And we will choose the
( )
3 one that covers the total parameter range for both BWR and 4 PWR. If we don't have a single model that covers the 5 parameter range, we'll incorporate what models in the 6 code.
l 7 Do you have a question? l 8 CRAIRMAN KRESS: No. I see it on your slide 9 now.
I 10 (Slide.)
11 DR. ELTAWILA: The second step will be to 12 recover the 3-D reactor kinetic capability in RAMONA. And 13 we'll do that again through the peer review process. l
,rs
\M 14 We're going to look at the NESTLE model in the TRAC-B and l
15 NEM model in the TRAC-B. l 16 And we'll decide which capability is suited 17 for our need and yours. And then we'll incorporate that 18 into the modern architecture TRAC-B. And we'll select the 19 appropriate neutronic to thermal hydraulic interfaces.
20 (Slide.)
21 DR. ELTAWILA: What we were going to do is to 22 recover the RELAPS capabilities and incorporate it here.
23 And we want to be assured that adequacy of the typical 24 RELAP5 calculation can be done. That will include AP600,
/~
k_,h
) 25 small-break LOCA, and operating transient.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
518 1 Again, we're going to use the PIRT to identify f w., 2 a model that's needed to be upgraded to retain the RELAPS 3 capability. We have made a lot of investment in RELAPS in 4 the past two or three years to handle AP600. We want to 5 see how this capability can be recovered in the TRAC code.
6 We will assess competing models again to see 7 the parameter range to cover BWR, PWR, and ALWR, and, 8 again, which will use the peer review process. In Element 9 4 itself, we know that there are some models that they are 10 deficient in the existing codes.
11 So what we would like to look at, the 12 higher-ranking phenomena here. And if we know that there 13 are deficient models that can be improved, we run the O
Y- 14 experimental program and develop this model. And we'll 15 incorporate it in that code structure. This will take 16 care of the consolidation phase.
17 (Slide.)
18 DR. ELTAWILA: These are some of the potential 19 areas that we think that upgrade would be needed. We have 20 the constitutive model; for example, interphase drag; ;
I 21 AP600 specific components, like the thermal front in CMT; l
22 and some numeric issues to assure that the code is robust 23 at low-pressure conditions.
24 MEMBER CATTON: What about stratification in g)
(, 25 cold leg / hot leg? I don't see that there.
I NEAL R. GROSS I CoVRT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234 4433
519 1 DR. ELTAWILA: This is an example, yes, es 2 MEMBER CATTON: Okay.
< 3 DR. ELTAWILA: Definitely that's an issue that 4 we will be discussing, considering definitely.
5 (Slide.)
6 DR. ELTAWILA: I think Step Three is exactly 7 identical to the CSAU methodology. In Step One we have to 8 perform the assessment itself. And for that we have to 9 establish an assessment matrix that will cover all the 10 range for the plants that we will be analyzing: ALWR, 11 BWR, and PWR.
12 Since we will have a new code, we will have to l
13 define nodalization for the nuclear power plant and I,,')
\'l 14 convert the input deck. And then we will compare the 15 calculation versus -- this should be SET, the separate 16 effect test, using the same nodalization as used in the 17 nuclear power plant and documentary results.
18 (Slide.)
19 DR. ELTAWILA: We'll do the same thing with 20 the integral test calculation. If the results here are 21 acceptable, we'll go decide. The noding that we have 22 chosen is the optimum noding for this code. And if the 23 answer is if we need any code noding changed, we'll go 24 through the GUI here, which will help us modify the input
(^)8
(_ 25 deck and to come up with the nodalization, go through 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
.______._._..__-._..._._____.__.___.m._._ _ -
520 l
i _1- process until we know that we have reached the optimum l 2- nodalization.
j O 3 And then we'11 go through the CSAU methodology l
4- and by using the guidance provided in the CSAU methodology i
l 5 to address the scalability of the model and to quantify 1
3 6 uncertainty for all PIRT high-ranked phenomena.
1 7 The result will be included as part of the j 4
8 code documentation. Individual model uncertainties will !
9 be available for use in future CSAU efforts. And data 10 used in uncertainty quantification will be electronically
- . 11 archived in a format that expedites its result when the I 1
12 model is improved.
)
- 13 (Slide.)
1 14 DR. ELTAWILA: So much has been said about the l 4
. 15 improved user interface. We already started that effort.
- 16 We know that the codes are not user-friendly in either 17 preparing the input or looking at the output.
1 J.
l 18 So we have some effort to improve these j 19 capabilities or the input in preparation of the input 1 20 deck. We are going to use a component-based, 21 hardware-oriented GUI that will allow us, for example, to 22 look at an elbow by combining the pipes' schedule and bend i
- 23 radiuses, and so on to be able to build that model.
! 24 It would be icon-driven. So it would be much 25 easier. And there would be an automatic user guideline NEAL R. GROSS 2
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 3
- - - - , , - - + r - , ,, , - , . , . . - , - - , , = . - - - , - - - - - , , - - - -
__ _ . _ . _ _ _ _ _ ____._.___.-___._.__..__._._.__...m.__..
.__.m ___ _ _
521 1 with hyperlinks to the code manual so they can assess the 2 code during the part of the input deck where in the manual 3 this would be discussed.
4' It will have online library of input models so 5 ene user can choose some of these input models or can make 6 some changes of these. And it will be the GUI should be 7 easy to translate and modify existing input models. We 8 have started that activity already.
9 (Slide.)
10 DR. ELTAWILA: In the area of output, we want 11 to be able to get plant schematic from the input deck,.be 12 able to look at the point and click on any variable so we 13 can see how it's performing during the transient or 14 cross-processing. And we will have to have a plot option 15' of time history, profile, and so on.
16 What I discussed thus far is our effort here 17 to go through the code consolidation, code improvement,
)
18 assessment, uncertainty quantification to be able to come 19 at single hydraulic codes in about four to five years from 20 now. And in conjunction with that, we are developing the 21 user interfaces.
22 (Slide.)
23 DR. ELTAWILA: The success of the code itself 24 will depend on the amount of thermal hydraulic improvement 25 in the model and improvement into two-phase flow NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
n . . - _ _ _ - . .
522 1 formulation and the constitutive relationship. SoLif we I 2 take everything that we have right now and put it in the 3 code, we might end up with the same kind'of code that we l 4 have. So we really need to do some pilot study in certain 5 areas that then help us in improvement in the constitutive I 6 model and applicability of advanced modeling and numerics.
7 So we are planning to start a small-scale 8 experimental program to develop the database needed to 9 develop and improve the two-phase formulation. And we are 10 going to look at some of the results from these and the l Ell one that we feel that can have a potential benefit on 12 improving the code robustness and accuracy and so on.
13- We'll test it first. We'll peer review it O 14 before we incorporate it into.the code. So you can see j
15 there is element here of an experimental program to 4 16 provide the data needed to improve the codes.
17 (Slide.)
18 DR. ELTAWILA: It's the same thing that I said 19 before but put it in a time line here so you would be able 20 to see the plan that we are proceeding with. The TRAC-P 21 database restructuring is ongoing right now and will be 22 finished early in Fiscal Year 1998.
23 In parallel with that, we are starting 24 interacting with the user community. And we are going to
() 25 interact with the TRAC-P people and some of the experts 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 )
i
)
523 i 1 the area of software. We'll try to identify option and 2 define requirements for modular structure. That will be a 3 very critical decision that we will reach that position.
4 Once we define that and we -- I'm confident i-5 right now that we'll be able to use to a lot of extent the 6 TRAC-P structure. We will implement the modular structure 7 sometime in Fiscal Year '98 '99.
8 MEMBER CATTON: What is database 9 restructuring?
10 DR. ELTAWILA: Joe? ,
j 11 MR. KELLY: This is Joe Kelly from Research.
12 As Ralph Caruso mentioned during his and 13 Farouk reiterated, the codes'were developed in the '70s.
! O 14 And the architecture of the codes, especially as to how j 15 they stored variables, was aimed at large mainframes that 16 had very small memory, like typically the CDC 7600.
l 17 So, for example, in the TRAC code, there's 18 something called a container array. And that's how they 19 used dynamic -- well, it's a type of homespun dynamic 20 memory allocation to overcome the limitations of Fortran 21 77.
22 And with that container array, there's a lot 23 of equivalencing' going on using artificial pointers and 24 also the bid packing.
f 25 ACTING CHAIRMAN CATTON: Okay.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
._.__......_.._..__.__-_.__.._..-_.__._._._._..._m..._ -..._ _ - . _ . m ..
524 1 MR. KELLY: And these things make the code a 2 nightmare for maintenance. So the first step in our 3 program'is to get rid of all that stuff and bring it up to 4 a more modern coding standard.
5 ACTING CHAIRMAN CATTON: Now, I thought you 6 had something underway now to address whether or not-7 TRAC-P could be appropriately restructured to make it j 8 modular.
9 DR.-ELTAWILA: I think the database-10 restructuring is going on right now. And Joe and --
11 ACTING CHAIRMAN CATTON: I'm not referring to-
- 12. database restructuring now. I'm referring to the 13 facility.
Oi 14- DR. ELTAWILA: No. The modularity.
15 ACTING CHAIRMAN CATTON: Modularity.
16 DR. ELTAWILA: Yes. Joe has been discussing 17 that issue right now. Do you want to say anything about 18 it, Joe,-with the --
19 ACTING CHAIRMAN CATTON: If it's the same as 20 the -- you don't know any more now than you did at the 21 subcommittee meeting about --
22 DR. ELTAWILA: Well --
23 ACTING CHAIRMAN CATTON: Okay. I think that's 24 crucial.
25 DR. ELTAWILA: No.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
j (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 m -y qw-a - ---4 q y yyMwy e -- -g : 1
i 1
525 l l
1 MR. KELLY: This is Joe Kelly again. l l
,r~3 2 And, yes, we do. We have had a meeting with l t i V
3 the people from Los Alamos. We've also talked to, for 4 example, John Mahafee at Penn State, who was one of the 5 developers. And we have a much better understanding of 6 what's in the TRAC code now and what may need to be done l 7 in the future.
l 8 Also, we have looked at the linkage between j 9 TRAC and MELCOR which was done. In that case they used a 10 two-dimensional vessel model that was developed external l l
11 to TRAC. And they used it actually as a module within the l 12 TRAC code exactly in the type way that we want to be able l 13 to do in the future.
( )
k 14 So we have a much better faith that we'll be 15 able to put models directly into the TRAC code in the 16 future.
17 DR. ELTAWILA: The actual code is MELPROG.
18 It's not MELCOR.
19 MR. KELLY: Sorry.
20 ACTING CHAIRMAN CATTON: I guess it would be 21 very helpful to me to have seen some more description of 22 what you're going to do to make TRAC modular. This is 23 what I referred to in my opening statements. I think this 24 is a real pinch point in the process.
/^\
( ,) 25 DR. ELTAWILA: Absolutely. We agree.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
- . . . - - - . - . . . . . . . . . . . . . . . - . - ..~ .. - . . ~ . _ .. . . ~ . . ..
526
'1- ACTING CHAIRMAN CATTON: And in doing it way 2 out in '97, that's a year into your program. I'd like to 3 hear a really positive statement about the capability of 4 TRAC-P to be made flexible and modular.
5 MR. HODGES: Ivan, we're two weeks into Fiscal 6 Year '97 already.
7 ACTING CHAIRMAN CATTON: Say it again.
8 MR. HODGES: We are already two weeks into 9 Fiscal Year '97.
10 ACTING CHAIRMAN CATTON: Okay.
11 MR. HODGES. Those are fiscal years.
12 ACTING CHAIRMAN CATTON: Okay. Okay.
13 (Laughter.)
14 DR. ELTAWILA: Really, from the discussion I 1
15 had with the staff, I think we are going to be achieving a-16 high degree of modularity but not the puritan definition 17- of modularity. And I think for our purpose, that would be i 18 good. j 19 I think as long as we'll be able to make 20 modification to the code or to add new model in the code 21 and make it easier for the ucer and robust, I think that 22 will go a long way to achieve this code.
23 And if you recall, at this stage here, once 24 you have the consolidated code, as we start to develop a 25 mode; for example, the piping model that you talk about, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 j
527 1 to look at the thermal stratification and pipe and so on, 2 that can be built as a special model, standalone, sit on 3 the side, and be linked. The CMT model, you can build it 4 on.the side, have it on equations and so on. And you link 5 it on the side.
6 So although the TRAC model architecture is j I
7' going to be that a lot of the other improvement will be in j 8 a modular form. And it will be linked to the code in 9 gradual steps.
10 Anybody want to say anything? Please feel 11 free.
l 12 MR. KELLY: Just to clarify a little, the 13 point of the yellow box that says " Identify Options and 14 Define Requirements for Modular Structure," during that 15 period of time, what we will be doing is testing various 16 approaches to linking modules with the TRAC code to 3 17 determine which is the best approach.
18 Once we know what the best approach is, then 19 we will write the requirements and then have it j i
20 impler ;d in TRAC.
l 21 ACTING CHAIRMAN CATTON: Okay. So this is not 22 ceally a question of whether or not you can-do it. It's !
l 23 you believe you can do it, and now what remains is exactly 24 what is it that you're going to do. i 25 MR. KELLY: Exactly. What kind of information NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE. N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
528 1 has to be passed from the module to TRAC and vice versa,
,-s, 2 how does it have to fit into the matrix structure, etc.
)
U 3 And we have enough examples that we're sure we 4 can do it. And it's a question of what's the best way to 5 get the best amount of results for the effort.
6 ACTING CHAIRMAN CATTON: That didn't come 7 through with what you were saying.
8 Go ahead.
9 DR. ELTAWILA: That's why he's the expert.
10 ACTING CHAIRMAN CATTON: I hear him. I know.
11 (Laughter.)
12 MEMBER MILLER: I had a question on that. Are 13 you ready to finish this overhead? I had a question. I'm (3,
'\ s 14 an I&C guy. I'm going to ask a thermal hydraulics 15 question here. On the --
16 ACTING CHAIRMAN CATTON: You don't have to 17 apologize.
18 MEMBER MILLER: Well, you ask I&C questions 19 all the time.
20 I had a question on the 3-D instability code 21 for the BWR. Is that going to be able to handle the 22 in-phase and out-of-phase instabilities with the thermal 23 hydraulics and neutronics and all of that type of --
24 DR. ELTAWILA: Yes. If we have the 3-D and
(.
( ,) 25 your front capability would be able to deal with in-phase NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) TJ4 '433
529 1 and out-of-phase. oscillation, yes.
2 MEMBER MILLER: Are codes of that type not 3 available today?
4 DR. ELTAWILA: It is available. Believe it or 5 not, the TRAC-P BWR has that capability. And the TRAC-P-6 but a DOE version has that capability. And RELAPS has a 7 3-D neutronic capability, but, again, it's DOE versions.
'8 So the capabilities exist, but we have not 9 adopted an NRC version of the code as of yet. It's RAMONA 10 that has that capability.
11 .MR. HODGES: And, in fact, we actually did 12 calculations like that maybe eight-ten years ago, but they 13 were very expensive and very time-consuming. So you don't 14 do a lot of them.
15 MEMBER MILLER: I'm asking the question I 16 guess from two different perspectives. I have a graduate 17 student who would like to have one right now for another 18 application.
19 But, two,'I would have thought GE or others 20 would have had those kinds of codes some time ago. And 21 you say that GE and others have had those, but they're 22 very expensive to use?
23 MR. HODGES: And, in fact, the NRC has had 24 them, but they're also expensive to use. And so you use 25 them very sparingly.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON D.C. 20005-3701 (202) 234-4433
530 1 MEMBER MILLSR: So when GE assured us that
-~ 2 Lasalle-type things would not occur, they ran those codes V 3 some years ago with that in mind; right? Is that --
4 DR. ELTAWILA: If I understand, the situation 5 with the BWR instability issue, it is not something that 6 you can predict it, that the code will not be able to 7 predict that behavior. But you cannot identify a region 8 that you can if you operate outside this region, you will 9 not get into that situation.
10 MEMBER MILLER: Right. Well, GE always said 11 if you operate outside a certain region, it couldn't 12 occur.
13 MR. HODGES: It's very sensitive to minor s /
N' 14 changes. And so to be able to predict all the conditions 15 under which it might occur is difficult because just, for 16 example, a change in the buildup of the xenon over a 17 period of time can make a difference between stable and 18 unstable. That's happened in real plants.
19 So very minor changes that you might not have 20 included in your initial analysis can make a difference.
21 MR. CARUSO: There are a lot of degrees of 22 freedom in these calculations, in these stability 23 calculations. And each one of the calculations takes a 24 long time to do. So it's exploring an N-dimensional (3
! ,/ 25 universe, which is very large.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
531 1 MEMBER MILLER: Also, I'm led to understand by 2 colleagues from the university sector we have a better 3 understanding now of the neutronics-thermal hydraulics 4 interactions that we did a dozen years ago.
5 MR. CARUSO: They seem to do a very good job 6 of calculating actual events when they occur because they 7 know what the actual boundary conditions are to do the 8 calculations. And they say, "Oh, yes. It should have 9 gone unstable there."
10 But if you say, "Will it go unstable in this 11 other condition?" hmmm.
1 12 MEMBER MILLER: We still don't know that, j 13 MR. CARUSO: We're not sure.
14 ACTING CHAIRMAN CATTON: Part of it gets back 15 -- I mean, we understand the interaction because it's just 16 through phase change, but what happens is that the code 17 takes so long to run. You don't properly exercise all the 18 possibilities. So you miss one. And that happens to be 19 the one that causes the problem in one of the plants.
20 The whole BWR issue is a -- I mean, it's a 21- commercial for fast-running code. It's a commercial for a 22 need for flexibility. You can't put a program together 23 and meet all the requirements of Mr. Caruso and do these 24 things. You just can't.
25 You have to have the capability for somebody NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
532
-1 to come in and say, " Gee, for this particular kind of 2 problem, I know that these things are relatively O 3 unimportant. Tune them out somehow or put in a simple 4 model to represent them. Then exercise all of the 5 parameters."
6 And when the BWR Lasalle instability incident 7 took place, the TRAC-BWR that was the staff's tool that 8 was available through Idaho, they just couldn't get it to 9 run.
10 They couldn't get the job done. And I think, 11 in part, it was because it took so long to run. Now, GE
! 12 could. They made their system run. But, again, to save 13 money, they limited what they looked at.
14 MR. CARUSO: Right.
15 ACTING CHAIRMAN CATTON: They didn't 'nclude 16 the whole plant. It depends on the level that I included.
17 If I include it at a level where I can go in and actually 18 see a velocity distribution in a pipe, I ain't ever' going 19' to get it done because it's just too much computation.
20 I have to back off and say, " Gee, for a pipe, 21 in this particular case, all I need is a rho V2 times Co 22 for pressure drop. That's enough. I have to have the 23 flexibility to go in and do that so that I can vary where 24 the details are captured."
) 25 If you can't do this, you don't get these 4
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
533 1- problems run. And that's one of the reasons that Wolfgang 2 at Brookhaven -- he had a tool like this. So he was able O 3 to go in and actually do the whole problem. And the 4 reason that it's so important that the staff can do this 5 is because at the time GE said, "You're wrong." But 6 Wolfaang is kind of like an elephant.
7 Why don't you go ahead?
8 MEMBER MILLER: Okay.
9 MR. HODGES: Just as a question, his plant 10 analyzer was one diesel. '.t would not have handled the 11 problem he was referring to. It would have handled --
12 MEMBER MILLER: No. You're right.
13 MR. HODGES: It would have handled the 14 Lasalle.
'15 MEMBER MILLER: Sure.
16 MR. HODGES: But the out-of-phase oscillations 17 he was referring to. But we did go through RAMONA about 18 eight or ten years ago.
19 ACTING CHAIRMAN CATTON: See, and, again, 20 RAMONA has relatively unsophisticated thermal hydraulics. j 21 MR. HODGES: Yes.
22 ACTING CHAIRMAN CATTON: So it turns out you 23 could reduce the complexity in the thermal hydraulics, 24 speed up the code. Then you're able to look at many 25 things.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
534 1 MR. HODGES: But what you're saying, Ivan, is f~ 2 we need modularity. And we agree. And that's why you've i
-l .
3 got to have it there.
4 ACTING CHAIRMAN CATTON: Absolutely, 5 absolutely. And that's why I really want to hear strong 6 statements about the capability of TRAC-P to accomplish 7 this.
8 DR. EI.TANILA : Did you hear that statement, 9 Professor Catton? We'll have Joe state it again.
10 ACTING CHAIRMAN CATTON: And again and again 11 and again.
12 DR. ELTAWILA: If Ralph Caruso can say 13 " Documentation, documentation, documentation," I would say i/ 14 "It is modular. It can be modular. And we will succeed 15 in doing that."
16 MR. HODGES: And I think it's important that 17 you give us an opportunity to do that. My concern with 18 your earlier statement -- I think it's the same one that 19 Farouk had -- is if you start out with a negative 20 statement on what we're trying to do, you're prejudging 21 it. You could kill that program before we ever get 22 started. And that's a concern I have because I think it's 23 important to do this.
24 ACTING CHAIRMAN CATTON: I just want to be s
t/
1, 25 sure you address the issue.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHPJGTON, D.C. 20005-3701 (202) 234-4433 ll
535 1 DR. ELTAWILA: We are not ignoring it.
j- s 2 ACTING CHAIRMAN CATTON: Okay.
\] 3 DR. ELTAWILA: And it was one of our 4 requirements from day one on this issue.
5 MEMBER MILLER: Farouk, can I make one more --
6 I'm sorry.
7 ACTING CHAIRMAN CATTON: Sure. Go ahead.
8 MEMBER MILLER: I have one more question.
9 Just out of curiosity, what kind of computer platform is 10 going to be required to operate this code?
11 DR. ELTAWILA: As Ralph indicated, it would be 12 platform-independent so that the code, you would be able 13 to put it on any work station or PC or something like f )
\/ 14 that. So that's the way that we are planning to develop 15 it.
16 MEMBER POWERS: Farouk, let met ask you a 17 question about not the chart that you've got up there but 18 the chart that you will put up in the year 2000 if your 19 computer system works.
20 Have you thought about going to a massively 21 parallel system? Once you have this code in a reasonably 22 stabilized form, you're still doing things a lot with it.
23 And the possibility of having a custom massively parallel 24 platform for this computer that would allow you to do some f^s,
( ,/ 25 of these uncertainty analyses and parameter variations NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (2r.) 2344433
536 .
l i
1 that right now you're time-constrained? j
,s 2 DR. ELTAWILA: I think yes. The answer is
( ) <
~#
3 yes. And we have ongoing efforts right now with Le Brock, 1
4 Incorporation, I think, a small company. And sl:e had her 5 Ph.D. on parallel processing. And she has a patent 6 pending on this issue. So independent of this effort, 7 she's starting looking at the issue of parallel i'
8 processing.
9 And, again, I did not bring enough viewgraphs.
10 One of our requirements that the code, too, will be able i
11 to be amenable for parallelization in the future. So we 1 1
3 12 have that effort ongoing. And definitely we are taking it '
13 into consideration. ]
r~^N 3 e
) l i/ 14 MEMBER POWERS: It sounds like the right level 15 of effort. It's a futuristic thing. I have seen this J 16 done for band structure calculations that used to be in 17 this mode of it takes forever to do them and whatnot. And j f
18 once they had a stabilized code, they had a custom-built !
19 massively parallel system for them.
20 And it's mind-bogg'ing. They can do band 21 structure calculations in the twinkling of an eye. And so 22 for your C' ass 2, 3 users, suddenly this becomes a routine 23 tool. They -- 1 24 ACTING CHAIRMAN CATTON: To a certain extent,
(_,) 25 some of that is available now. Sun actually has a j NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4 33
537 1 parallel maker compiler, but that may not be the best way 7- 2 to do it. So I'm a little curious about how --
k.)3 3 MEMBER POWERS: One of those mysteries.
4 ACTING CHAIRMAN CATTON: You know, it's got to 5 be Research.
6 (Laughter.)
i 7 DR. ELTAWILA: We think we can live forever. !
l 8 ACTING CHAIRMAN CATTON: Well, and you have ,
9 broad shoulders.
l 10 You know, you can start at the front end to l 11 say, "I want to design this to be parallel" in a certain i l
12 sense. And you can take parts of the problem and 13 deliberately say, "This goes to a CPU" and so forth.
- i
\/ 14 And you can definitely say that this gives you 15 fantastic performance on a relatively weak computer. On 16 the other hand, it's not going to meet the requirements 17 that Mr. Caruso laid down. If you adhere to those 18 requirements, which means he's going to take an 19 off-the-shelf system somewhere and you're going to run 20 this problem, then you write your program a different way.
21 If you're going to say that I'm going to do 22 some of the compilation myself to set this thing up and 23 I'm going to use Sun work stations with 12 CPUs, then you 24 do something else if you want to optimize it. I'm not
,/ N
() 25 quite sure where they're headed yet.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234 4433
538 l 1 MR. CARUSO: I guess I would say, once again, j 7\ 2 I have mixed feelings about this. I would like a code --
b 3 ACTING CHAIRMAN CATTON: I do, too. l l
4 MR. CARUSO: I would like a code that I could 5 feed to a special purpose compiler that can handle a Sun 1
I 6 with 4A processors which would be able to use all those l
7 processors to run the code faster. That's good. ,
l 8 But I don't want somebody to write a code that 9 can only run on a purpose-built machine. I mean, the 10 government procurement process is such that I can't afford l 11 that.
12 ACTING CHAIRMAN CATTON: See, that says that 13 the kind of parallel processing " hey're going to do is ia t ) .
k/ 14 more mediocre. It's you let the system compiler decide i l
i 15 what to do. That way you can feed this to any machine. )
16 And if you happen to have a better machine, it 3 l
?
17 will perform better. But you never really achieve the 18 true potential of the multiple CPUs. l 19 MEMBER POWERS: Well, truthfully I am thinking 20 in terms of Research being its biggest customer for its 21 own product here and --
22 ACTING CHAIRMAN CATTON: Well, they certainly 23 could do that in the back room and --
24 MEMBER POWERS: -- and the concern that you O
(,) 25 have that it's very difficult to do uncertainty analyses NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND A'd., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
539 1 on these codes when it takes so much time to run them.
fS 2 But, again, I'm very definitely thinking in terms of not L) 3 what's up on the screen but the years that have fallen off 4 the edge out there.
5 I mean, you've got to have a stable code to go 6 to these massively parallel systems. It's just a thought 7 in the future. And it sounds like you're pursuing it at 8 about the -- maybe even more aggressively than I would.
9 ACTING CHAIRMAN CATTON: My understanding is 10 that an1 they laid out the architecture for MELCOR, not 11 MELP..OG but MELCOR, that their original intent was to have 12 the capability to go to multiple CPUs and direct how it 13 would work. I think you ought to take a look at that.
(n 14 DR. ELTAWILA: Okay.
15 ACTING CHAIRMAN CATTON: Now, it never 16 happened because the parallel processing didn't come along 17 fast enough to be implemented in the right way.
18 DR. ELTAWILA: Okay.
19 MR. CARUSO: I mean, one of the problems we 20 have right now is just getting the optimizing compilers to 21 optimize these codes --
22 ACTING CHAIRMAN CATTON: Yes.
23 t MR. CARUSO: -- because the codes are so 24 convoluted and you have such clever programmers working to rm k 25 make these things run fast on the CDC 7600 that when you NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
540 1 feed it to an RS6000 compiler, it can't figure out what f
f- 2 exactly is he doing here in optimizing. And we would
\' ,)
3 prefer that they write very, very simple code and let the 4 compilers optimize it, rather than let the programmers 5 optimize it.
6 ACTING CHAIRMAN CATTON: That's a commercial 7 for standardization.
8 MEMBER POWERS: Does NRC have a programming 9 standard?
10 DR. ELTAWILA: Yes. I 11 MR. CARUSO: I believe we have official 12 documents which describe standards.
13 ACTING CHAIRMAN CATTON: That's new. In the )
/~'T I 14 past there was not a standard.
15 DR. ELTAWILA: It's not a standard. There is I i
16 a NUREG/CR report about software documentation --
17 MR. CARUSO: Yes, yes.
18 DR. ELTAWILA: -- and development, yes. l 19 MR. CARUSO: That's right.
20 ACTING CHAIRMAN CATTON: But one of the 21 problems in the past was that programmers could do 22 whatever the hell they wanted. And I'm hoping that -- as 23 a matter of fact, we'll ask for some sort of a 24 scandardization.
7
(,) 25 MR. HODGES: The past is gone, Ivan.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433
541 1 (Laughter.)
2 MEMBER POWERS: History is bunk; right?
7s 3 ACTING CHAIRMAN CATTON: You'll have to 4 forgive my unease about all of this because the past 20 5 years shows that words are not what matters. What matters 6 is what winds up being on paper as guidance to the people 7 who are going to do the job.
8 MEMBER SHACK: Then please give us an 9 opportunity to show that we have changed.
10 (Laughter.)
11 ACTING CHAIRMAN CATTON: Okay.
12 DR. ELTAWILA: Okay. I think you can see 13 there are a lot of decision points here. For example,
)
14 when we start talking about BWR consolidation, we will 15 have to make a decision about the model that would be 16 selected. The model would be selected here for 3-D 17 neutronics. They are the last five model sections. So 18 there are a lot of, again, data determination about 19 modularity.
20 So there are a lot of pinch points, as you 21 call them, Professor Catton, that we are planning to 22 review progress and we continue to involve the peer review 23 process in identifying what's needed to be added into 24 these codes.
25 (Slide.)
NEAL R. GROSS COURT R2 PORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
542 1
1 DR. ELTAWILA: The next item on my discussion I 2 does complete the' consolidation and the improvement effort l 3 cf our codes. And, as Professor Catton indicated, this.is 4 a major activity. i 5 We would like to hear that the process is j 6 acceptable, I think. And, again, I would repeat. I think l 7 we would be able to achieve modularity during the TRAC-P: l 1
8 structure. And we will keep you' informed on a regular 9 basis about progress in this area. So that completes our 10 discussion here.
11 One item that usually is left is that our j 12 current capability is not complete, we suspect. When you 13 look at some of the issues that came out of these OECD l 14 lessons learned from international standard problems, j 15 downcomer mixing, the mixing in the plena of the steam 16 generator certification and cold legs and so on, all of 17 these phenomena cannot be dealt with with the current 18 two-phase flow modeling in the system code that we have.
19 So we are going to tap into what is 20 commercially available in the CFD codes. And these codes 21 are now being supported very extensively by the chemical
- 22. and the petroleura industries. So there are a ?ot of codes 23 that we can tap into them.
24 However, in order to address the 25 nuclear-specific kind of modeling feature in these codes, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
543 1 we have to spend some resources. So we are going to 2 initiate some pilot study and capitalize on what's been
- ,--)
~
3 developed by others 4 We are not going to reinvent the wheel to look 5 at -- CFD extends their capability to deal with some of 6 the issues that we are interested in. And we are going to 7 evaluate the results from these studies. And we'll decide 8 whether we'll keep it as a separate effect tool that we 9 can use or in the future if we need to link it to some 10 other codes or something like that.
11 So this would be going about four percent of l 12 our budget is devoted to CFD capability.
13 (Slide.)
f}
'/ 14 DR. ELTAWILA: If we look at the other element 15 of our research program --
16 ACTING CHAIRMAN CATTON: Before you leave 17 that, as part of your cooperative program with the French, 18 are you going to have access to some of the work that they l
19 have done in the CFD arena?
20 DR. ELTAWILA: Absolutely. We were preparing 21 a letter, if you recall, that we had --
22 ACTING CHAIRMAN CATTON: I understand that, 23 but I didn't recall that this was a part of it.
24 DR. ELTAWILA: No. The TRIO code that the n
(,,I 25 French, the TRIO code --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 2344433
-544 1 ACTING CHAIRMAN CATTON: Just a minute. In 2- nuclear engineering, in design several years ago there was
_O 3 a series of papers about their program. And it'had to do 4 more with the fast reactor than others. They have done 5 stratification, the effect of stratification, on 6 turbulence. They really have a ve.;y extensive program in 7 the CFD arena.
8 DR. ELTAWILA: That's correct.
9 ACTING CHAIRMAN CATTON: Now the question: Is 10 that part of your agreement --
11 DR. ELTAWILA: Well --
12 ACTING CHAIRMAN CATTON: -- or is your 13 agreement just with Grenoble? q
$. 14 DR. ELTAWILA: Well, the TRIO code is in
- 15. Grenoble. It might have subcontracting somewhere else, 16' but --
17 ACTING CHAIRMAN CATTON: CATHARE has nothing 18 to do with this other stuff.
19 DR. ELTAWILA: No.
20 ACTING CHAIRMAN CATTON: Then you don't --
21 DR. ELTAWILA: The other stuff you're talking 22 about may be at the universities and things like --
23 ACTING CHAIRMAN CATTON: Well, I'm not sure 24 where it is, but if you look in nuclear engineering and 25 design, a few years ago they had a whole series of papers NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE. N.W.
(202) 234 4 433 WASHINGTON, D C. 20005-3701 (202) 234 4433
545 1 describing the French program. And it's not in the f3 2 nuclear safety CATHARE kind of program. It was separate.
i
'u 6 3 You might take a look.
4 DR. ELTAWILA: Okay. We'll take a look, yes.
5 ACTING CHAIRMAN CATTON: I think they've done 6 a better job at a lot of these things than most.
7 DR. ELTAWILA: Okay. I'd like to just briefly 8 go over the experimental program. And the goals of the 9 experimental program are to advance the state of the art 10 in two-phase flow modeling and to obtain the detailed data 11 from fundamental experiments.
12 (Slide.)
13 DR. ELTAWILA: We will try to get information x- 14 to provide additional data for thermal hydraulic code 15 validation, whether it is coming from separate effects 16 tests or combined and integral tests that can be run at 17 the University of Maryland, at Purdue, or OSU.
18 We will perform independent confirmatory tests <
I 19 on an applicant design. For example, when you get an
]
20 application like the AP600 and we need to perform studies i
21 or if we get a future application, we would be able to use ]
22 this experimental facility to look at the application or i
23 the submittal that we have. ,
24 Since the staff is going to be heavily
/s !
(,,) 25 involved in the development of the test matrix and we 1 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
546 1 would be sending.them to work with the professor that is 2 running this facility, we would be able to maintain our 3 skills here in a running experiment and develop our skills 4 in interpreting the data coming out of the experiment.
5 At the same time, it will help graduate 6- students at these universities to be able to understand 7 thermal hydraulics to replace the aging thermal hydraulic 8 --
, 9 (Laughter.)
10 ACTING CHAIRMAN CATTON: Is he looking at me?
11 MEMBER POWERS: He looked directly at you, 12 directly at you. And so did everybody else in the room, i
13 by the way. I 14 DR. ELTAWILA: Again, in order to obtain
-15 information from international organizations, we have to 16 have our own program to be able to exchange with them.
17 They are not going to give us anything for free. So I 18 think that's another reason why we should maintain the 19 experimental facilities.
20 ACTING CHAIRMAN CATTON: I attended a meeting 21 with Farouk and others in Grenoble. The group at Grenoble 22 has been in this business for a long time, and they have 23 some excellent facilities.
24 But is this the extent of your involvement 25 with the foreign people? I mean, the English at Harwell NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
] 547
! I have an exceptionally good two-phase flow and thermal 2 hydraulics. group, people like Jeff Hewitt.
4 1
O 3 DR. ELTAWILA: Jeff Hewitt, yes.
4 ACTING CHAIRMAN CATTON: There are groups at. .
j 5 various places in Germany. There is the Japanese. Is 4
i 6 this going to be a truly international program or just 7 U.S.-French? i l
j 8 DR.'ELTAWILA: It is a truly international i
- 9 program. It just happens we were in Japan before that and j 10 discussed it. I know Wayne is going to England. One of l-1
- 11 the agenda on his items is to know the details of their I i
- 12 program and see how we will cooperate in that.
4 13 Hopefully in the future -- I know that the l O 14 Koreans are heavily involved in building facilities right 15 now. And we will discuss cooperation with them'. So we're
[ 16 doing it slowly, but it just happened that this time was 17 the French term.
18 MEMBER POWERS: As long as you're sidetracked, 19 Farouk, I know that you spent a salubrious three or four 20 days in the wilds of Russia looking at thermal hydraulic 21 facilities there in a particularly enjoyable expedition.
22 I wondered if --
23 DR. ELTAWILA: The swim in the river; right?
24 MEMBER POWERS: 'With the mosquito fests going 25 around. I wondered if they had figured in your plans at NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
. . . _ . . _ - ~ . __ .. _ _ _ _ _ _ _ _ ._ _ _ . - . . . _ _ _ . . _ _ _ _ .
548 1 all or you hope never to see that place again.
2 DR. ELTAWILA: I will never say~anything on
' the record, 3
4 (Laughter.)
5 DR. ELTAWILA: Ask me later. Actually, our )
i 7 6 cooperation with the Russians is very limited in this
]
7 area. They have facilities, but I think we would not be I
8 able to get too much information out of them. So I think j' l
j 9 we need to discuss it. They are coming here late in
! 10 October.
! 11 But most of their facilities, I don't think i 12 they are modern facilities. And looking at them right
! 13- now, I don't see to compare them to the ROSA facility or I
'v 14 the BETHSY facility in France is --
l 15 MEMBER POWERS: There's no comparison.
i 16 DR. ELTAWILA: There's no comparison.
- 17 ACTING CHAIRMAN CATTON
- But, on the other l
18 hand, the Russians have done very good basic work. So in l 19 the separate effects arena, it could be that they have i
j 20 some valuable things to offer. And I would agree with you i
! 21 as far as facilities. It's not much. But separate
- 22 effects I think they can contribute.
23 DR. ELTAWILA: Okay. That's one of the
! 24 actions that Tom King has.to discuss with them. You know,
() 25 he's our coordinator on the Russian activities. But 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
+ - . n - , - , , m , - - - --e ,m -
.e
549
- i. 1 think the message that -- as you know, the Chairman is 2 encouraging international cooperation. And we are moving 3 in that direction definitely.
1 1
1 4 As Professor Catton indicated, we have three 5 facilities, and we are going to start a separate effects n
6 test program. The OSU facility, which is built to support i
7 the AP600 certification, we have the facility as run right
,' 8 now for NRC tests that was running DBE and beyond design 9 basis accident tests. It is 129 volume, one-quarter 10 height, and can run 400 psi, up to 400 psi.
11 The former was built or scale model of the f 12 SBWR, but we are going to use it for generic BWR purpose.
! 13 And it's 1/400 volume scale, one-quarter height, and 150 f' N 14 psi.
15 The University of Maryland, it's the B&W 16 Group. And we're using it extensively for mixing the i 17 study. And you can see.the volume and height ratio over l 18 there.
- 19 As I indicated, we are going to start some
, 20 separate effects test program. We are going to try to
~
21 take advantage of some non-intrusive instrumentation
! 22 that's been developed, either in this field or in another 23 field. And we'll try to get some other information. Just j 24 a simple example here on a 2-phase flow measurement, void
[ 25 fraction, two-phase density.
i NEAL R. GROSS COURT REPORTERS AND TR*,NSCRIBERS
(
1323 RHODE ISLAND AVL , N.W.
(202) 234-4433 WASHINGTON, D.C. 20001 3701 (202) 234 4 33
{
<--=p- p y y,j - -v- - --<-y-
550 1 We are planning to hold a workshop in Santa 2 Barbara, California in March 1997 to look at advancement b,ey 3 in thermal hydraulic instrumentation to see what is going 4 on in the world and whether we need to develop additional 5 instrumentation or not.
6 ACTING CHAIRMAN CATTON: The interesting thing l
7 about the French relationship is that their facility is a j j
8 three-loop Westinghouse PWR. And it's probably one of the I
9 best facilities I think I've ever seen for that sort of l
10 thing. l 1
11 DR. ELTAWILA: That was in --
i 12 ACTING CHAIRMAN CATTON: And it's almost the l 13 same scale as ROSA, isn't it?
(~% l
'- 14 DR. ELTAWILA: No, it's much --
1 15 ACTING CHAIRMAN CATTON: It's pretty damn big.
16 DR. ELTAWILA: -- smaller than ROSA. ROSA is 17 1/30tn. BETHSY is 1/100th.
18 MR. BESSETTE: That's right. g 19 ACTING CHAIRMAN CATTON: The 1/100th?
20 DR. ELTAWILA: Yes.
21 ACTING CHAIRMAN CATTON: But it's really a 22 good facility. And I think it goes well with the group 23 that you've got.
24 DR. ELTAWILA: Yes.
(3
's ,) 25 MEMBER POWERS: When you look at 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
551 1 non-intrusive instrumentation for your experimental 2 program, is there any interest in looking at them for 3 actual reactor applications as well or is that quite a
, 4 fantasy?
5 DR ELTAWILA: No. That would be a side 6 benefit, but, as you know, by regulation or whatever you 7 call it, by charter, NRC is not supposed to develop 8 anything for the industry. So if anything comes out as a 9 benefit for in-plant measurement, I think that that would 10 be even better.
11 (Slide.)
12 DR. ELTAWILA: I did not -- actually, j
13 Professor Ishii and Professor Jose Reyes in the 14 subcommittee provided a very extensive test plan, some of' 15 the phenomena that we might be interested in looking at.
16 So here, you know, some of the ohenomena that 17 we.would like to study experimentally at OSU and PUMA. We ,
l 18 would like to look at single-phase, two-phase natural 19 cirmilation, the critical flow. This is one of the 20 lessons learned from the small break LOCA ISP that is 21 conducted by CSNI break flow valve, critical flow. These 22 are very difficult to model right now and predict. ,
1 l 23 We noticed that when you activate ADS 1 24 through 3 and you have a break in AP600, that you get 25 multiple choke locations. So we want to study the effect NEAL R. GFM)SS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
1 552 1 1 of multiple choke flow on the behavior, thermal hydraulic
() 2 3
behavior, of the plant.
We would like to look at counter-current flow l
l 4 limits. The effect of the discharge _of the pressurizer in !
5 the AP600 test. facility, we would like to study this 6 phenomenon further.
7 Downcomer mixing, that's a very important 8 issue. And, again, we can run some of these tests at 9 Maryland or at OSU. Flow stability and the heat transfer, !
10 that's one of the advantages of the tall chimney in the l 11 formal facility. So we will be able to look at the l 12 stability issue. I 13 Liquid entrainment and phase separation in 14 tees, one of the_ design features of AP600 that you want to 15 active ADS for so we can bring the pressure down. So the 16 separation of phases at the tee junction is a very 17 important phenomenon. Do we like to look at that in 18 detail?
19 As Professor Catton indicated, thermal i 1
20 stratification in horizontal pipes and, for example, as 21 you mentioned, in the formal meeting, the effect of 22 separation pool condensation, and you get stratified j 1
23 later, which can affect containment pressure.
24 So there is a lot of phenomena that would be
() 25 worthwhile to study experimentally. Whether we will NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE. N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4 433
553 1 succeed in putting models in the code in the immediate
., - .s 2 future, that remains to be caen. But we need to study 3 them first. l 4 (Slide.)
5 DR. ELTAWILA: Okay. The n-house 6 capabilities. As I indicated before, the program is being 7 directed from here. There is no doubt about it. I have a 8 lot of young engineers that are eager to get their hands 9 on moving forward to start developing the code and looking 10 at the different options. And I would like to get them 11 moving so we don't continue discouraging them.
12 We are going to be involved in developing the 13 code. We develop models in-house. We are going to work (3
e i
\/ 14 very closely with our contractor. And nothing is going to 15 go into the code unti.' the staff here understands what it 16 means and will be put by the staff. So we will be 17 developing models. We incorporate the models in the code.
18 We are going to do assessments here. We are 19 going to do reactor safety analysis here. We started 20 already. At least seven members of the staff, each one of 21 them is running one of the AP600 calculations.
22 We are going to be involved in the thermal 23 hydraulic experiments. And we'll send people to watch 24 these experiments and run them and plan the experiment p
(_,/ 25 itself. We have to interact very closely with the code NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RtiODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
554 1 developer to know their needs. And maybe they can gy 2 together decide about what experiment needs to be done.
O 3 We are going to continue our international 4 cooperation in the area of code development; for example, 5 the CAMP, OECD international center problem. As 6 indicated, we are cooperating with CEA and JAERI and look 7 forward to cooperate with other countries who have some 8 hydraulic research programs.
9 We are going to continue with training of the 10 staff to run and interpret code results. For example, we 11 give them training on RELAPS code. And we're going to 12 give them training next month about running the TRAC code.
13 All the staff expressed interest in running
'/ 14 Fortran 90, a 77 on Fortran 90. So we'll offer them that 15 opportunity. And we'll continue to recruit and hire 16 needed skills.
17 (Slide.)
18 DR. ELTAWILA: In conclusion, we believe that 19 the proposed research plan will reduce the code 20 maintenance cost and improve performance. We believe that 21 the TRAC-P code will be able to give us the modularity 22 that we're looking for.
23 We believe that the plan that we put forward 24 to the Commission maintains a high level of research r~s
() 25 expertise in-house and selected contractor 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
555 1 consultants.
2 There is a balanced approach in the research 7S O 3 plan. It's not only code development. It has code 4 development elements, integral and combined experiments, 5 and separate effect experiments, model development and 6 improvement, and code assessment.
7 The plan as structured right now will help 8 retain the international cooperation because as we develop 9 the GUI and as we develop the code, we will require that a 10 converter is being developed along the way so people that 11 invest in the input deck will be able to convert their 12 input deck.
13 It will not be 100 percent conversion. I
('_ \
14 think there will be some labor associated with it. But at l
15 least we will make it much simpler to convert input deck l l
16 so they don't lose all the investment that they had made 17 in the inpet deck.
18 MEMBER Shl.OK: Are you --
19 DR. ELTAWILA: We are desperately seeking your 20 approval. Sorry.
21 MEMBER SHACK: Are you going to be able to do 22 that and make it easier to create new decks?
23 MEMBER FONTANA: Using the GUI? Yes.
24 MR. CARUSO: Yes. Yes.
n k,) 25 DR. ELTAWILA: This is one of the requirements NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (2C2) 234-4433
556 1 of the GUI. As I was saying, I think we waited. And 2 Professor Catton is correct. We waited a long time to
/ \
3 reach these points here. And we have the Chairman on our 4 side. And she really wants to help us establish 5 capability.
6 And I would like to start. And we can have a 7 review process as many times as you want it, but we have 8 to start. The money ic in the budget. I cannot start 9 until we have some positive encouragement from the 10 Committee here that we are on the right track.
11 And the details, we can discuss them along the l
12 way, but I really think we have a very good program that l
13 will build the capability here and will make us recognize I fh U 14 again as international not leader, maybe as equal, in this 15 area. I'm not going to choose the word " leader" anymore. l 1
1 16 Once we receive your comments, which I'm sure 17 is going to be favorable, --
18 (Laughter.)
19 DR. ELTAWILA: -- we plan to finalize the 20 program and submit it to the Commission for approval.
21 ACTING CHAIRMAN CATTON: Thank you, Farouk.
22 DR. ELTAWILA: Thank you very much.
23 ACTING CHAIRMAN CATTON: Thank you.
24 MEMBER BARTON: You almost want to stand up q
Q 25 and applaud after that.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234 4833
1 557 1 (Laughter.)
l l
(q
"/
2 MEMBER SHACK: We actually finished the t 3 thermal hydraulics presentation on time.
4 MEMBER BARTON: On schedule, on time.
I 5 ACTING CHAIRMAN CATTON: It's ten minutes 6 over, and I'm going to use another --
7 MEMBER SHACK: No, no. 10:30.
8 DR. ELTAWILA: 10:30.
9 ACTING CHAIRMAN CATTON: Okay. Okay. Then j 10 you've left me five minutes. .
11 DR. ELTAWILA: I watched myself very carefully 12 when it comes --
13 MEMBER POWERS: Dr. Eltawila, can I --
(~%
14 MEMBER SHACK: You're usually not the problem.
15 DR. ELTAWILA: I think --
16 MEMBER POWERS: Dr. Eltawila, can I ask a j
17 question about your last couple of points? We are a :
8 J
18 roadblock here apparently to getting on with this program. j j
19 How much of a roadblock are we?
20 DR. ELTAWILA: We would definitely like to 1 21 hear from you in favor. It is important to hear from the {
22 ACRS. It is an important program. It's moving in the !
23 direction that we want to see. We don't want to waste the 24 taxpayer money.
,m
(,) 25 So I believe that if we get any negative NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
558 l
1 feedback from the ACRS, it will set us back a few months
,s 2 to try to develop another plan. And I really don't think N ,
3 we're going to come with any better plan in our ACRS. The l i
4 plan is comprehensive, and it's doable.
5 So, to answer your question in short, it is a
]
]
6 --
7 MEMBER POWERS: What I am trying to understand i 8 is the urgency of a response and some feeling for the 9 completeness of that response. Are you looking for an 10 immediate blanket endorsement or are you looking for a 11 " Gee, it looks good, and we'd like to look at it further" j 12 kind of response?
13 DR. ELTAWILA: No. I think we --
( b.
f
~Lf 'l 14 MR. HODGES: Clearly we're in the start of the 15 program. There's a lot yet to be done. There are a lot j 16 of decisions that have to be made along the way. We need l 17 your input. We need your comments and your criticism.
18 We're not trying to imply we don't want criticism. Yet, 19 we need it. And we seek that as well.
20 I think what we have been trying to say is if 21 the letter is too negative, then it can affect funding.
22 It can kill the program before we ever get started.
23 That's what we're very sensitive about.
24 It we can have an agreement that as we go
,/~N
(_) 25 along we shara what we're doing, we get your feedback, 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
l 559 1 get your comments, and.we try to go forward and you think 2 that~we're generally on the right track, a letter like 3 that to me would be a success. And I would be very happy t
i-l I
4 with'something like that.
L 5 MEMBER POWERS: Can you give me a-feel for the 6 urgency?
7 ACTING CHAIRMAN CATTON: We're into the fiscal.
8 year. .
l 9 MR. HODGES: I would think that -- we're 10 already started in the plan. The Commission has a plan in 11 their hands already I think they're looking to use to kind 12 of give a flavor of: Do you think we're on the right 13 track? So that if you could do it within a few weeks, I 14 think that would be very useful.
15 DR. ELTAWILA: Another element in the urgency, 16 Dr. Powers, is that seriously we' hired six young engineers 17 that are involved right now. And they came to work on 18 code development and running experiments.
19 And they are just on board right now getting l
20 very impatient, which was very hard to come up with with 21 good people. And it would be harder to try to maintain 22 them if.we don't give them creative work to do here, l
23 MEMBER POWERS: I'm familiar with that ,
24 problem.
25 ACTING CHAIRMAN CATTON: I think it's a very 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
-. ,.- ,~, . . - - -
- 560 e l' exciting opportunity, but what I want to be sure of is l 2 that we don't fall into the same pit that we fell in i
O 3 before.
4' i' Twenty years of experience show that a single 5 code for all users is a mistake. Now,.there may be ways 6 to do it, and I think there are. But this means that when l i 7 you start putting together a code that-has these O characteristics of modularity and flexibility, if you have 9 a single code for all purposes, it has to be able to i
2 10 incorporate in a separate model.
}
, 11 And I'm not absolutely convinced that TRAC-P, j 12 at least from what I've heard, is the tool because I hear 4
13 a hint of reservation in statements about modularity, i
14 My feeling is this program ought to go
- j. 15 forward, no matter what, and then be redirected. If TRAC h
16 turns out to be a headache, then you do whatever the hell 17 you have to do to fix it. But it should go forward.
18 I think this hodgepodge of codes with strange 19 --
and I've tried to get into the internals of some of 20 these for my own purposes. And some of the programming 21 you run into is bizarre.
22 People build logic into it. And then they 23 build logic on top of the logic. And then that logic was 24 wrong. Some more is rolled in on the top of it. And you I 25 can't figure out what the hell the code is doing. The NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
561 1 answer happens to look right, but you have no idea why. l 2 I think all of this has to be cleaned up. I (O 3 think that, if nothing else happens, you had a really j 4 clean code using standard programming techniques with some 5 sort of standard kind of input / output interfaces with some 6 kind of a standard GUI. I think that would go a long way 7 and would be a very valuable contribution to where we're 8 at. )
l 9 'On t.op of that, if you can generate this 10 modular. tool as well, -- I mean, I don't know what to say 11 -- I would be very pleased.
12 DR. ELTAWILA: I would be the first one to go'
]
13 to my management and say, "It cannot be done. I will 14 stop." I have no -- no. I mean, it's --
15 ACTING CHAIRMAN CATTON: No, no. I don't say 16 stop because a lot of these things are long overdue. They 17 should have been done a long time ago.
18 DR. ELTAWILA: No. I appreciate that.
l 19 ACTING CHAIRMAN CATTON: And they haven't.
20 And I think they need to be done. This combined with the 21 three facilities and your arrangement with the French, I 22 think thermal hydraulics right now looks more exciting
! 23 than it has in the past ten years. And I'm really l
- 24 pleased.
25 DR. ELTAWILA: I will not say any word beyond 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
-_ _ , - _e, , _
.__m _ . ~ _ . _ _ ._ _ __._.-_. - ._ _ __ _ _ . _ . _ _ _ __. _ _ _ _ _ _ _ . . . . _ _ _ _ _ . _ _ .
! 562 i
- 1 that.
2 (Laughter.)
3 MR. HODGES: We are trying very hard, Ivan, to I
4 do the kind of things you're talking about. And, as I i-5 said earlier, I cannot necessarily defend the past. I
] <
6 probably had some part in effective decision --
j 7 ACTING CHAIRMAN CATTON: I understand that, i.
! 8 One of my colleagues up in front of us said something i
9 about what happens to you if you don't pay attention to i
- 10 history.
11 MR. HODGES: Oh, absolutely.
12 ACTING CHAIRMAN CATTON: And I want to be sure i l
13 that we pay attention to history.
4 L- 14 MR. HODGES: And a lot of what we're doing is 15 based upon lessons we've learned from the past. And 16 that's why we're doing what we're doing.
17 But, in addition, at the subcommittee meeting, i
18' I think Novak Zuber made a suggestion that we put together I 19 a lessons learned from what has happened. And I've asked 20 Lou Shrucken to start doing it. He's already started 21 working on that.
22 I think it's useful. It's not going to be a 23 quick and simple job because a lot of times you've got to 24 go back and talk to the people who you don't necessarily 25 have funding for what you're trying to get from chem.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
563 1 And so we'll have to do it in bits and pieces.
2 But we're going to do what we can to put together those 7--
~
3 lessons learned as well.
4 ACTING CHAIRMAN CATTON: Good.
5 MR. HODGES: We're listening to what you say.
6 We are very sensitive to that. And we're trying to 7 develop a tool that is useful to to NRR, to AEOD, and to 8 us.
9 And I have recognized all along that we have 1 l
10 had codes that have the Band-Aid fixes, and we need to fix )
11 that. Before we can do a lot on that, we first had to get 12 through our AP600 crisis and get a tool that would analyze j l
13 that. i
,y '
)
i 14 We've done that now. Now we're in a position 15 where we can start focusing on what we need to do for the 16 future. And what you're hearing is the early stages of 17 our planning -- and I'm sure when we come and talk to you 18 five years from now, the actual path we have taken will be 19 quite a bit different from what we laid out. But it's the 20 best we can do at this point in time.
21 ACTING CHAIRMAN CATTON: Oxay. Thank you very 22 much.
23 And I would turn the meeting back to Tom, but 24 he's not here. So I'll turn it back to Bill Shack. And
(_) 25 why don't you give us a break?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
564 1 MEMBER SHACK: It's time for a break.
,y, 2 ACTING CHAIRMAN CATTON: Thank you very much.
,)
3 MEMBER SHACK: We'll come back at a quarter 4 of.
5 (Whereupon, the foregoing matter was concluded 6 at 10:36 a.m.)
7 8
9 10 11 12 l 13 t a
'n/
14 15 16 17 18 19 20 21 i
23 24 Q,,) 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
O CERTIFICATE This is to certify that the attached proceedings before the United States Nuclear Regulatory Commission in the matter of:
Name of Proceeding: 435TH ACRS Docket Number:- N/A Place of Proceeding: ROCKVILLE, MARYLAND were held as herein appears, and that this is the original transcript thereof for the file of the United States Nuclear Regulatory Commission taken by me and, thereafter reduced to O 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.
A bW//2 .
DORBETT RINER Official Reporter Neal R. Gross and Co., Inc.
O NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVENUE, NW (202) 234-4433 WASHINGTON. D.C. 20005 (202) 234-4433
.- .- >;, n - =_
A INTRODUCTORY STATEMENT BY THE ACRS CHAIRMAN j 435TH ACRS MEETING, OCTOBER 9-12, 1996 THE MEETING WILL NOW COME TO ORDER. THIS IS THE THIRD DAY OF THE 435TH MEETING OF THE ADVISORY COMMITTEE ON REACTOR SAFEGUARDS. DURING TODAY'S MEETING, THE COMMITTEE WILL CONSIDER l THE FOLLOWING:
(1) ACTIVITIES ASSOCIATED WITH THE NRC THERMAL HYDRAULIC j CODES f (2) REPORT BY THE HUMAN FACTORS SUBCOMMITTEE CHAIRMAN (3) PREPARATION OF REPORTS PROVI NS HE FED V CO I E ACT l
1 j MR. SAM DURAISWAMY IS THE DESIGNATED FEDERAL OFFICIAL FOR THE INITIAL PORTION OF THE MEETING.
WE HAVE RECEIVED NO WRITTEN STATEMENTS OR REQUESTS FOR TIME TO MAKE ORAL STATEMENTS FROM MEMBERS OF THE PUBLIC REGARDING TODAY'S SESSIONS. A TRANSCRIPT OF PORTIONS OF THE MEETING IS BEING KEPT, AND IT IS REQUESTED THAT THE SPEAKERS USE ONE OF THE MICROPHONES, IDENTIFY THEMSELVES AND SPEAK WITH SUFFICIENT CLARITY AND VOLUME SO THAT THEY CAN BE READILY HEARD.
l 0
8 O
Thermal-Hydraulic Computer Codes-A Regulator's Perspective i
l l
)
l l
Ralph Caruso l U.S. Nuclear Regulatory Commission l Office of Nuclear Reactor Regulation j Analytical Support Group October 11,1996 Nen slide O
i-
@ow are the T/H Codes Used?
o Cross checks of vendor codes and calculations (Historical use) 1 lI o Sensitivity Calculations o Event Analyses J
1 o Simulations 1
a i
i Next slide i
I i
O O
1 i
GTypical Analysis Requests o Vendor Comparisons
- o PRA Success Criteria l
l o Steam Generator Differential Pressure Calculations j
- o BWR Shroud Cracking 4
o Shutdown events (Wolf Creek, Hope Creek) o "What-if" Scenarios (Shearon Harris)
'4 l Next slide '
O l l
l l
1 i
O
t i C. Thermal-Hydraulic Code Users 4
- 1. Knowledgeable," Qualified" users O Knows code details O Understands input deck details and can prepare complex / unusual models i O Knows how and when to use options
! O Knows when codes are providing incorrect / inconsistent results in the most challenging cases l
- 2. Occasional code user O Knows systems and thermal hydraulics 0 Knows overall basis for codes O Builds simple models O Can identify code difficulties during non-complicated transients
- 3. System analyst (not trained in T/H)
- O Needs a Simulator
- O Does not build or modify models O Consults with more experienced users to verify code output O Just a user of a plant model a PRA practitioner
- e Training a Procedure development E Management Next slide l
U J
4
$egulatory Needs for a T/H Code i
- 1. The code must be robust
, 2. The code must be defensible by the code author i i
- 3. The code should be easy to use:
O Simple to develop and modify plant models O Useful, scrutable output l
J
- 4. The code must be platform independent
- 5. The code should be fast mnning
- 6. The code should be linkable to other codes
- 7. The code should have the ability to range various parameters (e.g., heat transfer coeffecients) 1
- 8. The code documentation must identify the limits of applicability of the models and correlations, as well as the uncertainty associated with those models and correlations.
Next slide O
ORecommendations for Code Development
- 1. Work closely with the user community. They have a lot of valuable experience to share.
- 2. Look at other engineering software to see how they interface with their users.
O Finite element / stress calculation programs O CAD /CAE O CommercialT/H packages
- 3. DOCUMl!NT, DOCUMENT,IX)CUMENT!
- 4. Gather user suggestions and hints into one electronic document that is available to all users. Keep it up to date, and require that code developers review it regularly.
- 5. Provide full disclosure of all code limitations and problems in one electronic document that is available to all code users. Code developers should be required to review this document frequently.
- 6. Code developers should be required to perform at least one regulatory analysis with their product each year.
i O
O
. O O O 3
I
- Thermal-Hydraulics Research Plan Presented to The Advisory Committee on Reactor Safeguards October 11,1996 Farouk Eltawila Joseph Kelly David Bessette U.S. Nuclear Regulatory Commission US NRC l'
Outline e Overall Goals
- e Current Situation .
e Historical Perspectives e Goals and Features of Thermal-Hydraulic Code e Approach for T/H Code Consolidation I
e Thermal-Hydraulic Test Program e In-House Capabilities ,
e Concluding Remarks US NRC
O O O Overall Goals i
e Maintaining Core Competency in Thermal-Hydraulics, Reactor Physics, and Plant Transient Codes to Support Regulatory Decisions and Continuous International Interactions improve and Maintain Reactor Safety Codes Small Scale Experiments to improve Two-Phase Flow Formulation and Constitutive Relations Maintain Experimental Facilities and Conduct Experiments for Code Validation and Better Understanding of Thermal-Hydraulic Phenomena I
Develop and Maintain in-House Capabilities US NRC i
' ~
O O O NRC T/H RESEARCH PROGRAM OVERVIEW OF PROCESS USER PILOT INTERFACE STUDIES CODE ASSESSMENT NRC T/H CODE CODE AND CONSOLIDATION IMPROVEMENT 4 CODEAND .
UNCERTAINTY GUI QUANTIFICATION SEPARATE INTEGRAL NRC DATA FUNDMENTE EXPERIMENTS FFECTS W FACILIW W BANK TESTS TESTS I
O l
1 0
Y 3 0 1 2 %
4 0
0 F 3 1 3 2 1 _
0 %
0 % % % % % 0 Y 3 3
0 1 3
2 4 0 F 1 2 1 S
T N
E _
M E 9 %
R 9 % % % % % 0 Y
I 3 0 1 2 4 0 _
U 3 2 F 3 1 1 _
Q _
E _
R E
C R 8 %
U 9 % % % % % 0 O T F
9 4
0 1
5 1
2 2 4 0 S 1 E _
R O
H C
R A 7 9 % % % % %
E % 0 S Y 4 9 0 1
5 1
2 2 4 0 E F 1 R
C I
L U M A R R E O D D F Y O R H- C E
, P L A E ,
,Y A S E S R L ST M R C U T E CN R A I NC D I I NA M E A H L A AU) O OT R T N HRI MTR ET R G E CT U LSU E
C O O C ESG A R I
T N MA& C /NHE N P C A T I U L N N E AE ST ,&
E DDR Y ,S A U T O ,U H YTT T F N CS T P DNN N E
CC E UEE I
A EI E TMM M M VR OEI T DSI A S I R L RTR E S O E D
O RMH PU C MN R GO P L XS I
P S E E P
X D
F A
T O
C I ( A UPEA E C T
- .i Today's Situation e Currently NRC is Maintaining Three Thermal Hydraulic Codes (RELAP5, TRAC-BWR, TRAC-PWR) and One Coupled 3-D Neutronic and Simplified Thermal Hydraulic Code (RAMONA) e These Codes Were Developed in the 1970's For Main Frame Computers e Past Efforts to Convert the Existing Codes to New Computer Environments Were Not Comprehensive Leading to " Fixes" That Continue to Cause Problems e Code Errors / Corrections Have Often Used " Band Aid" Approach to Fix Symptoms, Not Root Cause.
US NRC
O O O Historical Perspectives l e TMI-2 Accident Highlighted Code Deficiencies: More Physical Phenomena Were involved in the Transient E.G., Core Heat-up, Pressurizer Discharge, Secondary Side Voiding and Filling, Interacting Operator Actions e These Phenomena Were Challenging to the Codes and Users e A Flurry of International Test and Code Development Programs Were Carried Out i i
e Despite the Somewhat Ad-Hoc Physical Models, and Limitations in Numerical Methods, T/H Codes Are Capable of Describing the Governing Phenomena and General System Behavior for Large Variety of Accident Scenarios .
i i
US NRC i
O O O '!t l
Lessons Learned From OECD/CSXIISP on Small Break LOCA r
. o OECD/CSNI International Standard Problems Highlight Remaining Deficiencies in Code Predictions e Experienced Users Are Able to Get the Relevant Phenomena Even in Cases When Complex Scenarios Are involved e A Wide Range of Results Are Obtained Using the Same Code Version (User Effect) e Code Deficiencies include Prediction of: Break Flow; Stratification in Cold Leg and Hot Leg; Mass Distributions in the Primary and Secondary Sides; Mixing in the Downcomer; Steam Condensation; Pressurizer Behavior; Core Uncovery and Heat-up; CCFL in SG Plena; Low Pressure Period US NRC
O O O i
Historical Perspectives (Continued) t e Assessment of the Lasalle-2 Instability incident Indicated That Codes With Inherent Numerical t Damping Renders the Predictions to Be More Stable Than the Plant Implicit Time Integration With First Order Spatial
- Differencing e Computer Simulations With Three-Dimensional Neutron Kinetics Are Indispensable for Local, Out-of-Phase Oscillation Analyses US NRC
- - - - -9
l i
Additional Code Limitations !
.- i o Computing Speed Has Significantly increased and the Cost of Computing Has Decreased. However, ,
Actual Costs Have Significantly Increased Due to:
Complexity of the Codes ;
Duration of Transients Difficulty of Interpreting the Results e Limitations in Physical Models; Use of Models Outside Data Base e Limitations of Numerical Models US NRC ,
O O. O Goals of XRC Thermal-Hydraulic Code Development e A State-of-the-Art Thermal-Hydraulic Code Capable of Performing Accident and Transient Analysis for Operating Light Water Reactors As Well As Advanced Passive Reactors e Must Have Capabilities Suited to Address Safety Significant Issues That Pose Challenge to Current Codes That Require:
Coupled RCS Thermal-Hydraulics and Containment Feedback Coupled RCS Thermal-Hydraulics and Neutron Kinetics
- US NRC
, .}
O O O Goals of XRC Thermal-Hydraulic Code Development (Continued) e Must Be Easy to Modify to incorporate New Models to Address Specific Issue of Interest e Must Have on Line User Assistance, Error Reporting and Modifications, and Graphical Users Interface e Must Be Easy to Use, Run Fast, Accurate, and Robust e Smaller incremental Cost in the Future for Performing l Analysis, Maintenance and Upgrades e One Code is Preferab e, So All Staff Can Be Knowledgeable Users. Adopt CATHARE As a Back-Up US NRC i
l O O O l l l
l l
l Approach e Integrate / Upgrade Existing Capabilities into a Single l Code Element One Code Capability Requirements Elemerit Two Consolidation Element Three Assessment & Parameters Ranging e Start Pilot Studies to Explore Advanced Concepts That Could Lead to improved Code US NRC i
Stsp Ons St:pTwo StspThrss b Identify Scenarios implement Modem Data TRAC-P TRAC-B RAMONA RELAPS E L&S BLOCA, SBO, Structure & Architecture l l l l
$ SGTR, ATWS InTRAC-P jg 3 1 r l I Og if 'f Review Documentation:
o .!=
! Identify Plant Designs identify Option and Requirements for Code Manuals, User Guides, Programmers Guide, Developmental PWRs' BWRs' ALWR
, g. Modularity Assessment, Model & Correlation ve C 1 f 1 f i r E Identify & Rank Develop Code Select Models &
$ Requirement
-+ Constitutive Correlations Phenomena (PIRT) for Consolidated Code
(,
,8 1 f yj Develop Needed Models if , Step One
$g TRAC ++ & Modem [ StepTwo jg l Step Four. Architecture StepThree o " }
if Establish Step One Assessment Matrix p StepThree improved User l l
% 1r Interface l $ Define Nodalization for
- .e NPP & Convert input
p Decks l } 1 r e i r ,e l $ Compare Calculation vs SET Data IET Data Compare Calculation vs g IETS using NPP Nodalizations + + IETS using NPP Nodalizations Base Base g Document Document
- a. I I Io 1r E
m
. E Noding Yes g StepTw Change
< I 1r No Determine Code &
g Experiment Accuracy Bias & Uncertainty - I i 3 r Determine "'
Effect & Scale
'i
.m O O O i
Code Upgrade Plan Element One--Step One: PIRT e Use Existing PIRTs, Develoa New PIRTs To Judge the importance o" Phenomena or Processes for Accident Scenarios of Interest for Different NPP Designs (PWRs, BWRs, ALWR)
Element One--Step Two: TRAC-P Modemization e implement Modern Data Structure & Code Architecture, to Permit Portability Across Variety of Platforms, increase Maintainability & Extensibility, and Enable Future Improvements US NRC
O O O Code Upgrade Plan (Continued)
Element One--Step Two: TRAC-P Modernization e Identify Options and Define Requirements for Code Structure / Modularity, Numerical Schemes, ETC. ,
e Determine the Best Approach for implementation Element One--Step Three: Models Selection e Select Models for Consolidated Code e Use PIRT, Identify Models That Need Upgrading to Retain RELAP5 Capability US NRC
_- j
- O O O Code Upgrade Plan (Continued)
Element Two--Step One: TRAC-B Model Selection o Add BWR-Specific Models To Base TRAC-P Code e Assess Competing Models Over Parameter Range important in the PIRTs (BWR & PWR) e Use Peer Review to Select Appropriate Models Element Two-Step Two: 3-D Reactor Kinetics Model e Select 3-D Reactor Kinetics Model to Recover 1 RAMONA Capability in Consolidated Code Peer Review Existing Neutronics Capabilities in i
TRAC-P (NESTLE) and TRAC-B(NEM)
Select / Develop Neutronic & T/H Interface US NRC
, O O O Code Upgrade Plan (Continued)
Element Two--Step Three: RELAP5 Model Selection I
e Assure Adequacy for Typical RELAP5 Applications, including AP600 SBLOCA & Operating Transients e Use PIRT, Identify Models That Need Upgrading to Retain RELAP5 Capability e Assess Competing Models Over Parameter Range Important in the PIRTs (BWR, PWR & ALWR) e Use Peer Review to Select Appropriate Models Element Two-Step Four: Upgrade TRAC-P Models Improved Models for AP600 SBLOCA & Operating Transients US NRC ;
O O O Code Upgrade Plan (Continued) e Potential Areas for Improvement:
e Constitutive Models ?
E.G., Interphase Drag in Rod Bundles, Condensation on Stratified Interface.
e AP600 Specific Components ?
E.G., Thermal Front in CMT e Numeric E.G., Robustness for Lew Pressure Conditions US NRC
O O O Code Upgrade Plan (Continued) i Element Three-Step One: Assessment e Peer Reviewed Assessment Matrix Based on Existing and New PIRTs (PWRs, BWRs, and ALWR) e Integral Effects Tests For Each Design & Scenario of Interest e Separate Effects Tests High Ranking Phenomena to Be Assessed Over Parameter Range Where Considered to Be Important e Automated Assessment & Comparison for Each Code Version US NRC
O O O Code Upgrade Plan (Continued) l Element Tbree-Step Two: Scalability & Uncertainty e Use the Guidance Provided in the CSAU l
Methodology to Address Scalability and to Quantify Uncertainty for All PIRT High Ranked Phenomena.
l Results Will Be included As Part of Code Documentation Individual Model Uncertainties Will Be Available for Use in Future CSAU Efforts Data Used in Uncertainty Quantification Will Be Electronically Archived in a Format That Expedites Its Reuse When a Model is improved US NRC
O O O Code Upgrade Plan (Continued)
Element Three--Step Three: Improved User /nterface e input:
Component Based & Hardware Oriented
-E.G. For Elbow ==> Pipe Schedule & Bend Radius Icon Driven Automatic User Guidelines & Hyperlinks to Manuals On-Line Library of Input Models Easy Translation & Modification of Existing Models US NRC
O O O
[
Code Upgrade Plan (Continued)
Element Three--Step Three: Improved UserInterface e Output i
Plant Schematic From Input Point & Click Variab e Selection i Plot Options: ;
-Time Histories, Profiles, Vector & Contour Plots i
US NRC
O O o Pilot Studies j l
e Provide Evolutionary improvements to Consolidated I Code Improvements to Constitutive Models Applicability of Advanced Modeling & Numerical Methods e Experimental Program to Develop Data Base to Develop / Improve Two-Phase and Constitutive Models Used in Two-Fluid Formulation e Promising Techniques Undergo Peer Review & ,
Testing in a Developmental Version Before .
Incorporation in T/H Code l
US NRC
~ r O O O FY97 FY98 FY99 FYOO FYO1 TRAC-P: Data Base Restructuring Code Architecture identify Options and Implement Define Requirements -+ Modular for Modular Structure Structure Determine Best Approach .9
+-e Model Selection ]g BWR Model Consolidation Implementation Testin9 d s_
Select Implement h 3-D Kin. 3-D Kinetics Instability & ATWS TRAC Assessment for O Relap 5 Capability SBLOCA & AP600 Model Model SET Assess &
E Selection Implementation Uncert. Quant. O O
l-Improved Experimental Program & Pilot Studies Models & i Numerics implement Upgrades i
O O O Computational Fluid Dynamics (CFD}
e Computation of Multiphase Flows Based on Current Approach Results in a Less-Than-Satisfactory
- Situation in Certain Applications e CFD Developments Are Being Pursued (E.G..,
Chemical Industry) to Compute Flow Regimes and Interfacial Configuration in Flows e Supplemental Work Related to Nuclear Applications e Initiate Pilot Projects to Capitalize on Advances in CFD for Multiphase Flows in Complex Geometry .
e Evaluate Results for Applications to Plant Transient Applications US NRC :
i
o o o j Goals of XRC Thermal-Hydraulic Test Program ]
I e Advance the State of the Art in Two-Phase Flow Modeling (Detail Data from Fundamental Experiments) o Provide Additional Data for Thermal-Hydraulic Code
- Validation (Separate, Combined, and Integral Effects) e Perform Independent Confirmatory Tests of an Applicant Design e Develop and Maintain Technical Skills e Quid Pro Quo to Obtain Data From International Experimental Programs i i
US NRC i
O O O Approach and Implementation Strategy e Maintain APEX Facility at Oregon State University PUMA Facility at Purdue University, and THECA Program at University of Maryland e Facilities Will Be Made Available to Other Researchers on a Pre-Established Schedule e Continue International Cooperation, E.G. , ROSA (JAERI, Japan), BETHSY ( CEA, France) e Details Associated With Experiments Will Be Developed and Peer Reviewed e Small-Scale Fundamental Experiments for Instrumentation Development US NRC ;
g ** 't O O O Approach and Implementation Strategy (Continued)
Preliminary List of Phenomena to Be Experimentally Examined:
e Single-Phase and Two-Phase Natural Circulation e Critical Flow (Break, Valves), Multiple Choked Flow l Behavior !
o Counter-Current Flow Limit (CCFL) e Downcomer Mixing e Flow Stability and Heat Transfer e Liquid Entrainment & Phase Separation in Tees !
e Thermal Stratification (Horizontal Pipes, Effect on ;
Suppression Pool) !
US NRC
o O o In-House Capabilities 1
Develop and Maintain Expertise in Thermal-Hydraulics and Reactor Safety Analysis e Reactor Safety Code Development e Reactor Safety Analysis e Thermal-Hydraulic Experiments e Continue International Interactions on Codes and Data (CAMP, OECD, CEA, JAERI) e Continue Training of Staff to Run and Interpret Code Results e Recruit and Hire in Needed Skills US NRC
O o O Concluding Remarks e The Proposed Research Plan Will Reduce Code Maintenance Cost and improve Performance e Maintain High Level of Research Expertise In-House and Selected Contractors / Consultants e Balanced Approach to Develop Data for Model Development, Improvement, and Assessment e Retain investment in Existing Code Decks e We Are Seeking the ACRS' Comments and Suggestion on the Plan e Plan to Be Finalized After Input From ACRS, Commission, Strategic Assessment US NRC
_ _ _ _ _ _ _ _ _ _ _ _ _ . = _ _ _ _
_____ ___ _ __- _________ _ _ _ ___________________ _