ML20210D398
| ML20210D398 | |
| Person / Time | |
|---|---|
| Issue date: | 04/29/1987 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1583, NUDOCS 8705080167 | |
| Download: ML20210D398 (329) | |
Text
RCl?S/ ISP,9 l UNITED STATES O NUCLEAR REGULATORY COMMISSION OTG YAL IN THE MATTER OF: DOCKET NO:
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON THERMAL HYDRAULIC PHENOMENA O
LOCATION: IDAHO FALLS, IDAHO PAGES: 296 - 520 DATE: WEDNESDAY, APRIL 29, 1987 l .
l ACRSOFF'CECDP'I ws.apggJromACRSClia
, ACE-FEDERAL REPORTERS, INC.
OfficialReporters 444 North Capitol Street Washington, D.C. 20001
- (202)347-3700 9'OSOOO1e7 P'a429 NATIONWIDE COVERAGE f 7- SS POR
1 UNITED STATES OF AMERICA O 2 NUCLEAR REGULATORY COMMISSION 3
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON THERMAL HYDRAULIC PHENOMENA 4
5 Westbank Quality Inn 6 475 River Parkway Bonneville Room 7
Wednesday, April 29, 1987 9
The subcommittee meeting reconvened at 8:30 a.m.,
10 Mr. Carlyle Michelson, chairman, presiding.
11 ACRS MEMBERS PRESENT:
12 MR. CARLYLE MICHELSON O' 13 MR. JESSE C. EBERSOLE 14 MR. DAVID A. WARD 15 ACRS CONSULTANTS PRESENT:
16 DR. CATTON 17 DR. SCHROCK 18 DR. SULLIVAN 9
DR. TIEN 20 21 22 23
() 24 T 25 ACE FEDERAL REPORTERS, INC.
- o:-347 37m Nanonwide Coserage muk3 4 6646 1
PUBLIC NOTICE BY THE UNITED STATES NUCLEAR REGULATORY COMMISSIONERS' ADVISORY COMMITTEE ON REACTOR SAFEGUARDS WEDNESDAY, APRIL 29, 1987 I
I The contents of this stenographic transcript of the proceedings of the United States Nuclear Regulatory Commission's Advisory Committee on Reactor Safeguards (ACRS), as reported herein, is an uncorrected record of 1,
the discussions recorded at the meeting held on the above date.
No member of the ACRS Staff and no participant at
, this meeting accepts any responsibility for errors or 1
inacc'uracies of statement or data contained in this transcript.
l i
i ,
i
- l 1
. f O i I
7280 01 01 297
('%
A ,N/bc 1 PROCEEDINGS 2 MR. MICHELSON: The meeting will now come to 3 order. This is the second day of the meeting of the 4 Advisory Committee on Reactor Safeguards, Subcommittee on 5 Thermal Hydraulic Phenomena.
a 6 I'm Carl Michelson, subcommittee chairman. We 7 will continue with the meeting where we left off yesterday.
8 Before we get into the meeting, I was wondering if, for the 9 record, we could ask the staff is there any kind of nice 10 little document that we could read concerning the 11 justification for the once-through steam generator work some 12 documents show?
13 Here's the problem. Here's why we think it's the 14 problem. But, hopefully, not too long.
15 MR. BECKNER: We have our B&W plan, which is 16 NUREG 1236. It hasn't been published yet, but it should be
! 17 any day. I think we've given out drafts before; if not, 18 we'll give another draft to you.
19 MR. MICHELSON: If you will send us a draft copy 20 then, so we can look at it.
21 MR. BECKNER: It may have gone through some minor 22 changes. I think we'd better send you another copy.
23 MR. MICHELSON: We are of course going to review 24 the thermal hydraulics program. And I'm sure, in the full 25 committee, we can anticipate people asking why do we need ACE FEDERAL REPORTERS, INC.
202-347-Do Nationwide Coserage N 4336-6M6
i 7280 01 01 298
()V/bc 1 this program, and the justification, as I see it, really 2 wasn't presented here but rather the assumption that you j 3 need it, now here's how we're going to do it.
- 4 I need to go back and review the justification in 5 case we're asked. It would be helpful.
I 6 MR. BOEHNERT: Bill, what was that number again?
j 7 MR. BECKNER: NUREG 1236.
j 8 MR. EBERSOLE: I hope there will be in that j 9 document sort of a broad explanation of the general transfer
- 10 differences between the B&W design and the CE-Westinghouse 11 in the context of the primary control parameters to use on l 12 the secondary to keep a balance between flux and flow.
! 13 You know, the old plants have a level control and
! 14 they have what amounts to a reset coming from steam flow, 15 water flow. And then, periodically, they get a reset from i 16 that.
l l 17 Here, my understanding is, in this particular i
i 18 boiler, there's virtually no set at the working level when 19 it's running. I'd like to see you say something about the l
j i
20 transition. This weird way we were talking about yesterday I 21 of spraying on the fraction of the tubes and tell me why i 22 just a retention of the normal method wouldn't be equally 23 effective? Or is this simply a diminished version of the
- 24 normal mode of cooling.
( 25 In short, go into the business of super-heated i
) /\CE FEDERAL REPORTERS, INC.
202 347-37(v) Nationwide Coserage MX)-336-6M6 i I
7280 01 01 299 r
(_yjV/bc 1 designs, where the boiler is virtually dry except at the 2 vary bottom, and say something about the control problems.
3 Can you do that? Is there a problem there?
4 MR. SHOTKIN: No problem. It's just that the 5 report is being published.
6 MR. EBERSOLE: Well, just to start at this unique 7 method of shutdown cooling with the aux feedpumps, it seems 8 to start rather late in the game.
9 MR. MICHELSON: There's one other point that 10 perhaps you can clarify for me.
11 Is there a means now provided to divert main 12 feedwater to the auxilliary feedwater nozzles?
() 13 MR. CHARLTON: There is a way to do that, but the 14 pipes are so much smaller, you can't get it in.
15 MR. MICHELSON: Do you know roughly how much main 16 feedwater one could divert?
17 MR. CONDIE: The aux feed capability, at least on 18 Oconee, is 1,200 gpm for a steam generator. Some of our 19 RELAP calculations have shown up to 6,800 gpm total. That's 20 like 5,400-5,600 gpm of the main feed could go to the aux 21 feed.
22 MR. MICHELSON: Do you have a provision to divert 23 from the main feedwater sparger up to the auxilliary c 24 feedwater sparger?
) 25 MR. CONDIE: That's correct.
ACE FEDERAL REPORTERS, INC.
202-347 370) Nationwide Coserage mig-33MM6
1 7280 01 01 300 lkV/bc 1 MR. MICHELSON: Is that valving manually 2 operated?
3 MR. CONDIE: Right. That's correct. Well, I 4 don't know if it's manually. It's remotely operated.
5 MR. MICHELSON: Is it remote manual, or is it 6 automatic?
7 MR. CONDIE: It's remote manual.
8 MR. EBERSOLE: The source of pumping power for 9 that would either be steam or electric motors, depending on 10 how they designed the main feedwater pumps.
11 If it's steam, of course, it's not likely to be 12 with you. It's maybe more likely to be with you if it's
(;
1_/ 13 motor-driven.
14 Is the value of that flow established on the 15 basis that you may not have the flow available if you lose l
16 the main feeds?
I 17 MR. CHARLTON: That's why you have the aux feed.
18 MR. EBERSOLE: I know that, but they don't really l
l 19 ! count on the flow from the main feeds in a serious way, do t
i 20 '
they? Because they may not have it. You can't count on the 21 main feedpumps. They're a luxury, so to speak.
22 ! So I don't know what degree of dependence they l l 23 l place on this connection, but it better not be much.
24 MR. CONDIE: I can't answer that, really, other ew !
k_) 25 than the analysis that we did. And I don't remember the l
l ace FEDER.u REPORTERS, INC.
[ :o:- W r oo Nanonwide Cmcrue W 336 tM6 L
7280 01 01 301
( V/bc 1 basis of the assumptions. Bv: they were able to show that 2 that's how much feedwater went into the aux feed system.
i 3 MR. EBERSOLE: They must put some qualifiers on 4 it.
l 5 MR. CONDIE: You're correct though. For example, 6 in Davis-Besse, they're all steam-powered except for the 7 startup pumps.
8 MR. EBERSOLE: And it persisted that way for too 9 many years.
10 MR. MICHELSON: Let's go ahead now and proceed 11 then, unless there are other questions from yesterday.
12 (No response.)
13 MR. MICHELSON: We'll proceed. Thank you.
14 MR. SIIOTKIN : Mr. Chairman, I'd just like to 15 introduce Norman Lauben, who is over here.
16 MR. MICHELSON: Excuse me. Do you have a l
17 separate section for B&W7 18 MR. SilOTKIN: B&W, G.E., Bill Beckner has that.
19 I guess our first speaker is Jim Wolf from Idaho.
20 (Slide.)
21 MR. WOLF: I'm Jim Wolf. What I'd like to talk 22 about this morning is pretty much of a continuation of the 23 discussion that Keith Condie started yesterday on the review 24 of our ISIF scaling document.
25 You may remember, in the scaling document itself, l Acu.eeonRat auroRTuRs. isc.
- , _ ~_ - --
-_ - - - - = _ _ . . _ - . _. . - _ __ -_ _- _ _
7280 01 01 302 O we did not make definite recommendations on the type of
(_jV/bc 1 2 facility that we would recommend building, or do anything 3 along that line. The main purpose of the scaling report 4 itself was to put forth the scaling laws, rank somewhat 5 objectively, albeit, the different scaling concepts. And 6 then, as a next step, apply those scaling concepts to the 7 actual facility itself.
8 (Slide.)
9 This morning, what I would like to discuss is how 10 we took the scaling concepts that were outlined in the 11 scaling study and went ahead and came up with a working 12 recommendation that we are using as the basis for the model 13 Keith discussed yesterday, to look at some of the 14 centergistic effects, and as kind of a baseline target for 15 people to start making suggestions and modifications on the 16 overall design.
! 17 Specifically, I'll be discussing what we feel are 18 the required parameters that we need to address as far as 19 coming up with a working recommendation, the relationships 20 between these different parameters, and then a discussion of I
21 the tradeoffs, the practicalities that are involved in i 22 applying the scaling laws to como up with a practical i
23 facility that can be built with the amount of money 24 available, and that will produce essentially the biggest
- () 25 bang for the buck.
ace FEDERAL. IlEl'ORTERS, INC.
}i 2em Um Ltionwide Cmeuge W D6w6
7280 01 01 303
(~%
(_jV /bc 1 Then, finally, we'll discuss the working 2 recommendation concept. And then go into a brief discussion 3 of the comments that have been made by our scaling 4 consultants that attended our review meetings, and how we 5 have incorporated their comments into our working concept.
6 (Slide.)
7 The parameters that are necessary to define any 8 type of a system model, of course, include what the working 9 fluid should be, whether it's water, freon or some other 10 fluid like that; the pressure that you require the facility 11 to operate at, whether it's what we call full pressure, 12 2,250 psi, whether it's reduced pressure, atmospheric
( 13 pressure in the case of many of the air / water experiments.
14 Another required parameter to properly scale a 15 model facility is the length ratio. In other words, how 16 does the piping length and the standard that you define your 17 length to be, compared to a full-scale plant, is it a 3/8 18 ratio, is it a half ratio?
! 19 In other words, is your piping half as long as
! 20 you would find in a full-scale plant, 3/8 as long, and so 21 on?
22 A related parameter that can also be defined, is l
23 it the diameter ratio, or the length ratio to the diameter i
i 24 ratio?
() 25 And this is comparing again your model plan to l
ACE FEDERAL REPORTERS, INC.
- o:wroo Neionwide nnaage sm3wra,
7280 01 01 304 A
(_jV/bc 1 what you would find in your full-scale reference plant.
2 We'll see some plots later on where we make quite heavy use 3 of this length ratio to diameter ratio in talking about our ,
4 linear facility or something near linear.
5 MR. WARD: The diameter of what, are you talking 6 about?
7 MR. WOLF: You're talking about your piping 8 diameter between your model and your full-scale reference 9 plant.
10 MR. WARD: So the core, the simulated core, is 11 scaled how then?
t 12 MR. WOLF: It's scaled by the diameter also.
() 13 MR. WARD: The total core or a fuel bundle? In 14 other words, when you have this .07 diameter ratio, let's 15 say.
16 MR. WOLF: That usually refers to the piping 17 itself.
i 18 MR. WARD: What ratio would you use then for the ,
19 core?
! 20 MR. WOLF: Usually, for the core, we would l i
1 21 probably speak in terms of the length to diameter ratio. l l 22 For instance, on the 10th linear scale, which has a length i
j 23 to diameter ratio of .1, you're looking at a core that is a j 24 tenth as high as a full-scale core.
() 25 In other words, it's one and a half feet and i
i t j ACE FEDERAL REeonTEas, INC. !
j :o:-m.rm Nationwide Coserage mMW(M6 f i___.____.
7280 01 01 305 A
(_jV/bc 1 high, instead of a 12 foot rod, I should say. And 6 or 8 2 inches across instead of 6 feet.
3 So you shrink all the dimensions by either the l 4 length ratio or the length to diameter ratio.
5 DR. CATTON: What do you do about the fuel?
6 MR. WOLF: The fuel, as we'll see when we ge into 7 the discussion of core power, full power versus decay heat, 8 what we tried to scale is either the full-scale power or the 9 scaled decay heat. We'll see that one of the scaling 10 distortions that we have to accept is a heat flux 11 distortion.
I 12 So what we did is we scaled the total power. One
() 13 of the things that you have to give up is scaled heat flux.
14 So we will in these facilities not scale the core to have a 15 typical heat flux.
I 16 DR. CATTON: So you wind up with the core just 17 being a source.
18 MR. WOLF: The core is strictly a heat source.
i 19 DR. CATTON: Basically, you're not scaling it.
20 MR. WOLF: The only thing we're scaling is the
{ 21 power input.
! 22 DR. CATTON: So why do you care what the diameter 23 ratio of the core is?
l 24 MR. WOLF: We like to preserve the correct flow
() 25 area, even though we don't have the right heat flux. So we ace FEnERai. IREPORTERS, INC.
- o .m-rm Nationwide cmerage s m m 6-6646
- . - , - _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ . _ . ,, r- .
. .- . _ __ _ _ _ ._ ~ _ ._ _ ._ . .~ ._._ _ __ .._.4. ___- .__._._ _.__. _ ____._.
1 i
i
} 7280 01 01 306 1 <
4
()V/bc 1 want the right flow velocities and void distribution.
) 2 DR. CATTON: How do you rationalize not caring 1 3 about the fact that you're not scaling the core correctly?
I 4 MR. WOLF: The majority, the type of transients i 5 that will be of interest if you look at, you know, the data i
6 base that is available on existing model plants, like the
) 7 semi-scale, MIST, there's been a large amount with the full l 8 length core, a large amount of data that has already been 9 gathered in that area.
i 10 DR. CATTON: They're all tall and skinny.
11 Here, part of what you're doing is to get around i
j 12 some of this criticism of being too tall and skinny; yet,
() 13 you're going to not worry about the core.
14 Somehow, that's inconsistent.
! 15 MR. WOLF As we worked through the scaling laws I
- 16 and how we apply them, that was one of the concessions. To i
I 17 go to anything otherithan a power to volume scale facility, 18 you end up with a distortion in the heat flux.
ii
- 19 DR. CATTON
- I'd like to see the rationalization l 20 that leads me to do that, rather than to give a little on i
l 21 the pressure.
l 22 MR. WOLP: Okay. I can go through that. ,
i l 23 DR. CATTON: When we discussed the pressure
! 24 yesterday, I was getting the feeling that you guys were
() 25 absolutely rigid on what you were trying -- not absolutely t l
f /\CE. FEDERAL REPonTnns, INC. ;
L 202. m .n oo Nanonwide courage soo-n w a 6 1
. . _ . . . _ _ . _ _ . _ - - _ _ _ - _ - , - _ _ .-_ - _ - m.--___ _.-_ _ _ - ..-. _ __ _ - - - . -
7280 01 01 307 m
(_,V/bc 1 rigid, but you were trying to accomplish something. Yet, 2 here, you're willing to give up a great deal with the core 3 that you won't give up on the pressure. So I can't 4 rationalize it.
5 MR. WOLF: The concensus that came out of, you 6 know, our consultants meetings on this was that the heat 7 transfer, especially the core heat transfer, is better 8 understood than the so-called hydrodynamics of the system.
9 And if you had to trade off onc versus the other 10 as far as the facility design, you gained more by preserving 11 the large volume type facility at the expense of, you know, 12 having correctly-scaled heat flux and a correctly scaled
() 13 core.
14 MR. MICHELSON: Could you quickly refresh my 15 memory as to the purpose of this facility?
16 MR. WOLF: Well, there are several purposes. The 17 long-range purpose is we might say to provide exploratory 18 research, to examine new areas, for example, of operator 19 action, and to identify, you know, potential problems before 20 they reach the point where we might say that they become 21 safety issues, and address specific problems as they come 22 up.
23 MR. MICHELSON: Don't you have to try to properly 24 scale the core for some of those studies?
() 25 I assume what the problem will be is you still ACE FuoERAE REPORTERS, INC.
202 347-3701) Nationwide Coseraec M ill-33M616 1.
. - - _ - . - . -- - _ - . - . . ~ . _ - _ . - . . - - _ - _ - - . - . - - - - -
' 7280 01 01 308 h1/bc 1 worry about core damage if you do the wrong operator !
! 2 actions. I i
i 3 MR. WOLF: One of the provisions that will l i
i 4 probably be included in the actual hardware design of this 1 i
j 5 facility, we're looking at the feasibility of extending the ,
6 thermal hydraulic aspect of, you know, the intact core j 7 versus something that is replaceable by a degraded core 8 geometry, i
i 9 You can't go from intact core geometry, you know, i
! 10 to a rubble bed. But we're looking at the possibility of 1
j 11 putting a rubble bed core in there in the facility. You i
- 12 know, how super heat is generated in the core, and how you ,
13 would look at fission transport, and things like that.
14 But, for severe core damage, we can't damage the l
l, 15 core.
16 DR. CATTON: Don't you think you'd be better off i ,
- 17 looking at rubble core as a separate effect? Because you're l
18 still going to be somewhat tall and skinny. If you get >
1 19 recirculation or whatever in a rubble bed, it's certainly l
20 going to give you different results than a quasi, one-
) 21 dimensional flow-through, i
22 The scaling is going to be different. But I l t' 23 think it's a separate question.
l 24 MR. WOLF: Yes. This is one of the thoughts for 25 the facility down the road, not one of the primary i
t i ~
1 Acn.FenEnAi. IturonTnns, INC.
, Oo: m 17m Nanonwide cmerage m33uu6
7280 01 01 _
309 O</8c 1 oe3ectivee.
2 DR. CATTON: That's why the front end of the 3 scaling is so important because these things tend to live 4 for a long time.
5 MR. WOLF That's correct, but let me walk 6 through the way that we applied the scaling laws and how the 7 compromises that we had to make in the areas of power, 8 pressure, and things like that, to give you an idea of what 4 9 you give up and what you gain by various compromises that
]
1 10 you have to make.
j i, 11 (Slide.)
12 Now, I don't intend to go through this entire
!O i
13 chere of geremeter reeioe in eny dete11. Thie 1e in the i 14 scaling report. Basically, what it does is it shows the 1
15 relationships for a number of different scaling parameters, l
16 such as length, diameter, velocity, power to volume, 17 gravity, heat flux for the various different conditions that 18 you can see here, that you will find throughout an integral l 19 transient.
! 20 This will be either for single-phase forced 21 circulation, single-phase natural circulation, and then a 22 transition to two phased natural circulation.
23 Now, what I would like to do next is show a 24 series of plots that essentially take these individual O 2s geremetere eed g1et them for e nem8er of cenditiene thee, l
ACE FnonRAL REPORTERS, INC.
- o: m -3 W Nadonwide Coserage MU6 M44
7280 01 01 310
()V/bc 1 hopefully, will shed some light on this entire subject.
2 Before I show the overall plot with the myriad 3 curves on there, let me just talk about the axis that we 4 have.
5 (Slide.)
6 As you see, plotted on the X-axis, we have a 7 number of the different ratios that can be computed from the 8 results of our scaling report -- the length ratio, the 9 velocity or time ratio and the heat flux ratio.
10 Now, we have plotted those either against the 11 power required from the facility, either full power, full-12 scaled power, or what we ere looking at in the area of decay 13 power of 5 percent.
14 Now, what we have actually shown here are some of 15 the computer parameters that.you can calculate from these 16 various parameter ratios, such as, for instance, we plotted 17 system volume, for instance, as a function of the length 18 ratio or the area ratio, and power.
19 We plotted the number of steam generator tubes 20 that would be required for a particular set of conditions, 21 and that's just basically the amount of heat transfer area 22 that you need to reject your total core power.
23 We've also plotted the length to diameter ratio 24 that we have. And, as you can see, this is a very, very
() 25 complicated plot. But you can use this type of a plot to Acn. FEDERAL REPORTERS, INC 202-347 3700 Nationwide Coserage M O-) tt. .(M6 h
_ . _ . ____ _ _~ _ _ .. . . _ _ _ _ _ _ . _ _ _ _ . _ . _ _ _ _ _ _ _ _ _ . . _ _ _ _ . .
t 4
i 1
- 7280 01 01 311 kV/bc 1 start making practical assumptions on the type of facility
- 2 that you have the resources to build.
l 3 Let me just walk through what some.of these lines l
l 4 are. What these are, they're lines of constant parameters.
l 5 Here we have the L over D ratio equal 1 for everything along i
- 6 this line here. And this essentially is the linear scale i
7 facility, the 10th linear scale, which means that the length
, 8 ratio compared to the full-scale plan is .1, is right here.
i j 9 So that for an L over D of 1 and a length ratio 10 of .1, you end up right here on this plot. As you can see, i
l 11 we then march across for the L over D ratios at 2, 3, 5, 7.5 l 12 and on down here to 10, and so on.
( 13 We've also plotted lines of constant steam l 14 generator tube numbers, or the amount of heat transfer
! 15 needed to reject your core power.
I ;
! 16 We have 1,500 tubes, 1,000 tubes, 250 tubes, and i
l 17 on down to 100 tubes to the few that we have in our current
{ 18 facilities.
l 19 We've also plotted lines of constant volume, 500 i
20 cubic feet; 250 cubic feet, 100 cubic feet, and so on.
21 So what you can do is you can use a plot like l 22 this to essentially hone in on the type of facility you want 1 23 by making assumptions that, well, we want as close to true i
l 24 metrical similitude as possible.
( 25 In that case, you would probably want to be down t
i
! ACE. FEDERAL REPORTERS. INC, mm.nm m,nwe cmerne umw6
] _- ____ __ _
7280 01 01 312
, ( V/bc 1 here on your L over D ratio of 1, 1.5 or maybe 2. If you're 2 worried about your velocity or time ratios -- in other 3 words, what distortion are you willing to accept in your 4 velocity ratios, or how fast your transient runs -- you can 5 look at what effect the changes in that parameter have in 6 the type of design you want.
7 You see down here for the velocity or time ratios 8 of 1, you're pretty much limited to your power to volume 9 scale facilities. The same thing on your heat flux. For a 10 permanent scale heat flux of 1, you are way over here on 11 this edge with a length to diameter ratio of somewhere 12 approaching those that we have in semi-scale.
( 13 MR. SULLIVAN: Is there any reason why you can't 14 put semi-scale on that?
15 MR. WOLF: Semi-scale is down here on this axis.
16 Later on, I've got a plot that shows what the L over D for 17 semi-scale is like, 48 to 1, the L over D ratio.
18 MIST is right here. You might notice that I've 19 shown the University of Maryland facility. I'll make a 20 couple more comments about that later.
l 21 ! This plot, as you might notice, appears for full l
22 1 pressure water. If you look at the pressure Maryland 23 operates at, it doesn't really belong on here. It was 24 l included just to show the volume comparison between the
() 25 University of Maryland facility and, for instance, MIST not f
- ACE. FEDERAL REvonTEns INc.
, 202-347 3700 Nationwide Ctnerage N o-33MM6 l
-- . . . _ . _ - - _ - _- - . - - . - - . - . . , - _ - _ - - . . - _ - . - _ , - _ - , , , - - . _ . - , _ . . , . _ _ _ , _ _ , . - - , - - _ p._ _ . - - _ . - .
7280 01 01 313 kV/bc 1 semi-scale.
2 MR. WARD: The volume is the total volume of the 3 primary system?
4 MR. WOLF: Total primary system volume, that's 5 correct.
6 (Slide.)
7 Now what I would like to do is walk through an 8 example of how you go down through the various system 9 parameters that you need to specify in order to have a sense 10 of the facility.
11 These turn out to be very similar to what wa 12 finally ended up with as our working recommendation. So, g
(_/ 13 the first parameter was that we would have as our working 14 fluid water. And it would be at full pressure.
15 The reasoning behind this is, if you remember the i
16 list of parameter ratios that we showed, for property ratios 17 that are not 1, you undergo a discontinuity between the 18 single phase and two-phase.
19 You undergo a continuity in your transition from 20 single phase to two-phase for all those different 21 parameters. If you stick with a property ratio, or a plant i 22 pressure to model pressure of 1 -- in other words, if your 23 scale facility is capable of reproducing the exact pressure 24 that you would see in your reference plant, those property
,/- y
\_/ 25 ratios are 1. And you do not see the discontinuity in 1 Acii FErmnai. RiiPourtins Ixc.
k :o:w roo Nanonmde coserace N u t t6m6
7280 01 01 314 7
V/bc 1 parameters that you have to produce pressure.
2 To pick a system volume, practicality indicates 3 that you would probably want something less than 250 cubic 4 feet; 250 cubic feet is roughly on the order of LOFT size 5 volume. Anything bigger than that, you probably could not 6 afford to build or have a building big enough to build.
7 Number of steam generator tubes. Probably 8 somewhere on the order of 500 tubes. Here, again, that's 9 somewhat dictated by the size and the core power that you 10 operate at. More than 500 tubes, your steam generator gets 11 to be the driving component in the cost of the overall 12 facility.
13 Your length to diameter ratio. Probably 14 somewhere less than 2. If you get much greater than 2, you 15 start to see very large distortions, or you start to deviate 16 from the idea of symmetrical similitude. So you would 17 probably be down on the lower end of our length diameter 18 ratio.
19 DR. CATTON: Do you have a criterion?
20 MR. WOLF: No. This is somewhat arbitrary. As I 21 walk through that plot that I showed previously, where we 22 essentially go in and we'll block out everything that is 23 greater than 250 cubic feet, and so on, you'll see that for 24 L over D ratios of 2, we have a reasonable velocity and time
( 25 ,
ratio.
l ACE.FEnERAt. REPORTERS, INC.
- L
- o:w.Mm Nanonwide Coserage sm-BM M,
7280 01 01 315 lkV/bc 1 And, you know, that is somewhat arbitrary -- 2, 2 1.5, you know, something on there.
3 DR. CATTON: Is the velocity time ratio important 4 to you?
5 MR. WOLF: Yes, because for one thing, the time 6 ratio tells you how much faster than real time the transient 7 will proceed. Your velocity ratio is important because it 8 relates -- the velocity in your model to what it should be 9 in the plant.
10 If you start having too low velocities, you start 11 having entrainment problems and you'll probably have trouble 12 developing your flow regimes like you should be prepared to.
) 13 So you'd like to have that as close to one as you 14 could.
I 15 ! MR. WARD: Does an L over D ratio less than 2 put F
16 you way over at the left end of that plot?
17 MR. WOLF: That's right. An L over D ratio of 1 18 would be an exact xerox reproduction of your reference 19 plant. You shrink the length and diameter ratios the same 20 amount.
i 1
21 l Here we're saying that we don't. Your length to 22 diameter ratio is 2. So you're a little bit tall and 23 skinny, but not really bad.
24 DR. CATTON: I still don't know how you judge
(~T i l C) 25 : bad.
i ACE FEDERAI REpoRTEns, INc.
[ :o:Wron Nanenaide Gncrape Smd %M46
7280 01 01 316 lkV/bc 1 MR. WOLF: It is somewhat subjective.
2 DR. CATTON: It should be judged based on the 3 tendency to multi-dimensional effects. And I don't even 4 hear that word.
5 MR. WOLF: We'll get to that.
6 DR. CATTON: If we're going to get to it, that's 7 all right.
J 8 MR. WOLF: If you look at the results of the 9 scaling study that Keith showed for the different facility 10 concepts, you know, other than our current power volume 11 scale facilities, they came out reproducing the phenomenon i 12 the best overall.
G
. (_/ 13 They reduced height; full pressure water l 14 facilities were second. As far as, you know, the number of
, 15 phenomena and how well they did reproduce the phenomena.
l 16 One of the assumptions that we have made for this i
17 example is that the core power will be full-scale core
( 18 power.
19 Practical limits on core power indicate that you l
20 would have a very difficult time supplying more than 10 1
21 megawatts to an experimental facility for the use of its 22 core power.
I
! 23 The only experimental facility that I'm aware of 24 that did have an excess of 10 megawatts of core power was l ( 25 some of the tests run down in G.E. and San Jose, and I think i
l I /\CE FEDERAL REPORTERS, INC.
j 202-347-3NC Nationwide Coserage 800-336-6646 j
7280 01 01 317 V/bc 1 they had 17 megawatts available, but they melted cables and 2 they blacked out half the city.
3 The existing facilities, such as ROSA, LOBI, MIST 4 all have power available less than 10 megawatts. To kind of 5 calibrate the cost -- and by 10 megawatts, that means 10 6 megawatts at the test site.
7 If you get in power ranges above 10 megawatts, 8 you start to have to increase in any facility available your 9 supply of power into the facility itself. And that's where 10 costs really just go out of site.
11 As a calibration point, Idaho Falls is currently 12 building a new generating station on the Snake River with 13 lowhead hydro or lowhead turbines. For 20 megawatts, the 14 cost is $50 million to supply that additional power.
15 So we're pretty much limited just 16 practicalitywise to less than 10 megawatts.
17 And as a base plant, reference plant, we chose 18 the Oconee plant for a lot of the same reasons that Tom 19 Charlton discussed yesterday. We have a good thermal 20 hydraulic model available.
21 We have good plant data and we have good
, 22 engineering data on what the plant itself looks like.
! 23 (Slide.)
! 24 What I would like to do now is to take those lO 25 verieme peremetere eed these essumptiens thee were mede end i
ACE FEDERAL. REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6M6 I
7280 Ol'01 318 1 ()V/bc 1 walk through how we use a chart like this to essentially 2 hone in on the acceptable region, or the region we would 3 feel comfortable with as far as a scale facility.
4 If we make the assumption that we will use full 5 pressure water, the next variable that we want to look at is 6 the primary system volume.
7 We said that we would probably, just from a 8 practical point of view, limit it to less than 200 cubic
. 9 feet. So what you see if you come down this line of 10 constant volume, everything above that line is eliminated.
11 Everything with a system volume above 200 feet is 12 eliminated.
( 13 ,
So what you've seen is you've already, by 14 specifying the maximum volume, you've already put some 15 constraints on the combinations of core power, length ratio, 16 core power, heat flux ratios that you could have.
i 17 (Slide.)
18 The next criterion was the number of steam 19 generator tubes available. We said that we would probably 20 want to limit ourselves to 500 steam generator tubes or l
21 less, and you see that if you ignore everything above the l
22 line, constant number of steam generator tubes, you would l
23 exclude this volume and this volume.
l l
24 So you can see we're shrinking the available
( 25 plant space, you might call it, by making some of these so-l l ACE FEDERAL REPORTERS, INC.
l 202.M7-3700 Nationwide Coserage 80)-336-(M6
7280 01 01 319
( Y/bc 1 called trade-offs, where you're trading off a number of 2 tubes for volume size.
3 (Slide.)
i
. 4 The next compromise that we made was we said, 5 okay, we want to keep as close to geometrical similitude as 6 possible. That is, we want to be down on this end of the L i 7 over D curve.
8 And looking at the types of length ratios and 4
9 velocity ratios that you could have, albeit somewhat i
l 10 arbitrary, we will not consider anything with an L over D 1 11 ratio of greater than 2.
12 If we go greater tha 2, we're tending to get back
(~/T s, 13 towards the types of facilities that we have had in the 14 past.
15 So, by sticking with an L over D ratio of 2, we 16 have essentially eliminated this entire bottom portion of i 17 the curve and we are honing in on the type of facility that 18 we have available to us.
l t
! 19 (Slide.)
20 The final parameter that we looked at was core i
I 21 power. And we said, for full scale core power, we want to l 22 limit it to 10 megawatts or less for full-scale power.
l 23 If you come across on the power curve at 10 i
l 24 megawatts, you see that the only portion of the curve that
- ( 25 is left is this dark, shaded portion down here. What that
{
ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide cmcrage soo-336-6M6 l
t
7280 01 01 320
((~%,,7/bc 1 gives you is essentially a very, very small length length 2 ratio. It gives you very low velocities or time ratios, 3 roughly .3 to maybe .4.
4 That indicates that your transient will run 40-50 5 percent faster than real time. Your heat flux ratios are 6 drastically distorted compared to what you would like if you 7 had ideal scaling.
8 You're looking at flux ratios somewhere on the 9 order of roughly 3-3.1 times what you would like it to be.
10 DR. TIEN: Based on your criteria here that shows 11 the University of Maryland facility actuall is quite good.
12 It comes very close. It's not an exact curve.
13 MR. WOLF: You have to remember that the Maryland 14 facility, that these curves were generated assuming full 15 pressure water, but volumewise, it does fit down. It's a 16 large-volume facility compared to MIST and Semiscale.
17 Now, that's an example of the type of procedure 18 you can go through in developing what we call a working 19 recommendation.
20 (Slide.)
21 And what we wanted to provide in this 22 recommendation was a basis for which we could go ahead and 23 do additional analysis. In other words, do our TRAC 24 analysis, start looking at some of the synergistic effects
( 25 that the scaling report didn't address.
ACE IIEDERAI. REPORTERS, INC.
202-347 3700 Nationwide nr.erage m33wa6
7280 01 01 321
( 7/bc 1 Remember, the scaling report only addressed 2 individual local phenomena, and did not say anything about 3 how these phenomena might interact with each other.
4 We wanted to have a basis for a critique of the 5 design. You need a base point to start discussions with 6 We feel that this working recommendation would propose a 7 type of facility that we could have, both the NRC and the 1
8 industry start critiquing some of the assumptions we made, 9 like Professor Catton mentioned.
10 The L over D ratio of 2 is somewhat objective.
11 Maybe we need to go to 1.5, maybe 2.5 would be sufficient.
12 We need a base facility that we can start doing
( 13 some more detailed cost estimation from. You have to start 14 doing some amount of fairly detailed -- how much piping you 15 need, how many steam generator tubes you need -- before you 16 can really start nailing down the exact cost of the 17 facility.
18 And we need a base facility so that once these 19 other items have been addressed, you can start a conceptual 20 design model.
21 (Slide.)
22 Now, the basis for the working recommendation 23 itself falls into several categories: The objectives of the 24 facility itself, where we want to provide code model
() 25 development, code assessment data, transient identification, 1 ACE FEDERAL REPORTERS, INC.
202447-3700 Nationwide Coserage MO- LIMM6
= , _ .
7280 01 01 322 r
(ms\1/bc 1 response to plant events, and so on and so forth.
2 The concept should have the ability to simulate 3 important phenomena; that definitely includes the three-4 dimensional phenomena that we have not to date gotten out of 5 our existing integral type facilities.
6 The basis is also looking at what data we do have 7 from available facilities. We do have a certain amount of 8 data that is good data, and we do have a lot of data 9 available. But there's no sense in going back and building 10 another facility that will reproduce that same data.
11 We want to correct weaknesses in past or current 12 facilities. Maybe not so much weaknesses, but areas where
( 13 we currently do not have data.
14 In other words, one of the primary areas besides, 15 you know, looking at multi-dimensional effects, is another 16 scaling point.
17 All of our codes have been developed from 18 experimental plant data that have been powered volume scale l 19 facilities. The tall, skinny.
20 This would give us a new point to compare the 21 codes. You have to look at the available dollars. Dr.
22 Shotkin addressed that yesterday in his slide, looking at 23 the budgets. You also have to address the areas of greatest
( 24 code activity.
) 25 These all have to be factored into the type of ace FEoERAE IREeonTEas. INC.
202-347-3700 Nanonwide emerage M n L 33MM6
7280 01 01 323 O
( j7 /bc 1 design that is made.
2 (Slide.)
3 To jump ahead for a second, the proposed working 4 concept that we made to NRC is basically a facility that is 5 full-pressure water, has a length ratio of .375. In other 6 words, this is a 3/8 size facility. It's a big facility.
7 I It has a length to diameter ratio of 2, as you 8 can see, we are deviating from our geometrical similitude a 9 little bit. But, in the next slide, we look at our facility 10 parameter plot again. We'll show you why.
11 And we also recommend that it be scaled to decay i
12 power rather than full-scale power. There are some very, 13 very good reasons why you buy much more with scaled power 14 than one with full power. And of course the concept will be 15 based on the Oconee plant.
16 (Slide.)
17 If you take that plot that I walked through 18 earlier and impose the recommended conditions that we have,
, 19 you'll see that we end up right there at the circle. As far 20 as the type of facility, instead of being over here on the 1 21 full power curve, we have moved over to the decay heat 22 curve.
23 For this type of facility, decay heat turns out
- 24 to be 2.7 megawatts at 5 percent decay heat. With this
() 25 particular design, you see that with an L over D ratio of 2, ACE-FEoERat. ItEl'ORTERS, INC.
l 202 34747m Nationwide Coserage 8m-3 E M46
7280 01 01 324
( V/bc ~1 we end up with a length ratio of not quite .4, a velocity 2 ratio of about . 6, and our heat flux ratio is about 1.6, 3 somewtere in that area.
4 You might remember, the so-called 10th linear 5 scale facility that we had talked about earlier, that 6 facility would be on this plot right here. If you look at 7 the 10 megawatt full power line, it would be right there.
8 MR. WARD: Jim, could I ask you a question about 9 the heat flux ratio?
10 That could be adjusted. I mean, the ratio you're 11 showing there is really for the given heater design, I 12 guess, the given size of the heater?
( 13 MR. WOLF: The heat flux ratio here?
14 MR. WARD: Yes.
15 MR. WOLF: That's the scaled. The heat flux
- 16 ratio comes right out of the scaling laws.
l l 17 (Slide.)
l
! 18 Once you define your heat input, your length i
j 19 ratio, your core delta t that you want, the heat flux is
! 20 essentially defined for you for a particular L over D ratio.
i
- 21 So what you end up with -- l I
22 MR. WARD: Okay.
l 23 MR. WOLF: Okay. ,
1
! 24 (Slide.)
( 25 One of the criteria, one of the assumptions that e
AcEEEDERAL REPORTERS, INC.
202 347 3700 Nanonwide cmerage mL336wa6
) ,
j 7280 01 01 325 V/bc 1 you have to make when you develop the recommendation that
' really doesn't show up on the parameter plot, is what to use 2
i 3 as your working fluid and your pressure.
f 4 And to review the justification for the full l 5 pressure water as opposed to another working fluid, or a 4
6 lower pressure, is that full pressure water minimizes the
) 7 analysis effort to draw inferences from the model data to 8 the plant data.
l 9 In other words, one of the objectives of the l 10 scaling study was to look at facilities that would minimize
$ 11 this effect. And with full pressure water, you do as close .
! 12 as possible. You minimize the difficulty in relating the inU 13 results through the code to your actual reference plant.
i j 14 It also provides the best simulation of our local l 15 phenomena as shown by the scaling study; and it also l 16 provides an undistorted transition from the so-called single i
17 phase flow to our two-phase operation.
18 (Slide.)
i 19 And just to show uhe same type of plot that Keith 1 '
l 20 showed yesterday, we're talking about our reduced height, i
l 21 full pressure water facility here as opposed to the subset I 22 of that, our linear scale facility.
J 23 And some of the other reduced type full pressure l
24 freon, and our current volume, full height, full pressure.
25 One of the reasons that we discounted freon was 1
ace FEDERAL REvonTEns, INC.
I 202 M ruo Nationwide Cmcrage 8m 33Mpi
7280 01 01 326 O
Nmjl/bc 1 the difficulty in essentially drawing inferences from plant 2 data to code data, or model data to scale plant data. The 3 codes just are not set up to handle freon, even though it 4 had a low overall numerical score.
5 I don't think that anyone would really advocate a 6 work horse integral facility to be based on freon.
7 (Slide.)
8 To just kind of put the recommended facility in 9 perspective with our past experimental facilities, what 10 we've done here is shown the L over D ratios for some of the 11 current and past facilities.
12 DR. CATTON: Just a quick question. I'm one of
() 13 the ones whose advocated freon.
14 (Laughter.)
15 DR. CATTON: I probably did it in total ignorance 16 of the amount of tuning that's been done to the codes. But, i 17 a lot of the correlations that we think are in the codes i
18 come from nondimensional readings, and this kind of thing.
i 19 If they, indeed, were used in the codes, then there should i 20 be no problem to switch to freon.
21 Is the reason you can't switch because of all the 22 tuning to fit the various facilities, constants associated i 23 with water, have been introduced?
24 MR. WOLP: Yes, there are no freon properties in 25 the code.
I Acu FunERAi. RuronTEns. INC.
l
- o:-m.nm Nanonwide Gncrage .w 336 ud6
7280 01 01 327
()7/bc 1 DR. CATTON: As a matter of fact, when I ran an 2 experiment at UCLA, Dupont gave me 10 drums of freon for 3 nothing.
4 But, am I correct that that's the reason that you 5 can't use freon, is because it's been tuned too much to 6 water?
7 MR. WOLP: Maybe even more important than that 8 is, if you suddenly switch to code predictions based on 9 freon, there's probably a couple of reasons.
10 You have to go through a very large code 11 assessment program before you can have a lot of confidence.
12 DR. CATTON: You shouldn't have to if the codes 13 are first principle. Like people argue, you shouldn't have 14 to do that. If you do have to do it, then these first 15 principles don't exist in the codes. Something is 16 squirrely.
17 MR. WOLF: Are they really first principle versus 18 specific water correlations.
19 DR. CATTON: You see, RELAP-5 was taken and 20 converted to a liquid metal code. It was done pretty 21 quickly as far as I know. If that can be Cone, why can't 22 you just take the liquid metal properties out and stuff in 23 freon properties?
24 You may not be able to do that with TRAC, but
() 25 you've already demonstrated that you can do that with RELAP-ACE.Friotinai. REronTiins, INC.
202-347 3710 Nationwide Coserage Mn33MM6
-_-._--_.._-_.m,-, _ _ . --._ -____ -. .-.. .- _ _.~. , _ , . . - - -
7280 01 01 328 V/bc 1 5.
2 MR. WOLF: We've done that with RELAP-5 in the 3 ATHENA code.
4 DR. CATTON: That's right.
5 MR. WOLF: I think it's more a perception of 6 maybe not the mechanics of doing it, but the amount of 7 effort you would have to go through to validate it.
8 Right now, they're still having discussions on 9 the applicability of the code to water systems and making 10 sure that the water-based correlation or the correlations 11 that are currently in the code are being used over the 12 correct range of conditions for which the correlations were
( 13 developed.
14 You probably just magnify that problem by an 15 order of magnitude with freon.
16 DR. CATTON: I can hardly wait to hear Vic Radson 17 at the next meeting describing ATHENA.
18 (Laughter.)
19 DR. SCHROCK: Could I add a comment, Jim?
20 I think it's a fact that there are a very large 21 number of imperical correlations in dimensional form, 22 applicable only to water in all of the codes.
, 23 And so reconstructing those correlations to be 24 applicable to froon does, in fact, represent a considerable
- ( 25 task. It's not that it can't be done, but the original data 1
Acti-Fimiinai. Riti>onTiins. INC.
! 2n2447 3700 N.nionwide cmerage m M w uan
7280 01 01 329 O
\m 7/bc 1 base would have to be reexamined in order to see how to 2 apply it to freon.
3 The original correlations were developed in 4 dimensional form rather than dimensionless form.
5 DR. TIEN: Could I comment on this?
6 In your proposed working concept, the main idea 7 is to really have an L over R ratio come down to 1.5, or 8 whatever.
! 9 MR. WOLF As opposed to 1, that's correct.
10 DR. TIEN: Now, if you assume right now it's at 11 3A scale, did you try to have an even smaller scale so that 12 you even reduce your L over D even lower?
() 13 That's expensive.
I 14 MR. WOLF: What we looked at was an L over D of 1 15 j with the length ratio itself at 1/10. This was the so-l 16 called 1/10 linear scale facility.
17 The next slide talks about some of the problems 18 with that particular concept. Once you get down on that 19 very f ar end, especially with length ratios that you can j 20 supply power for, there is a real problem in building that 21 type of facility.
22 DR. TIEN: Realizing that building any facility ll
- 23 would be very expensive, where would be the major cost in 24 terms of some of the parameters you mentioned?
( 25 MR. WOLF: The major cost -- one of the major 4 Acit FrinitRAL IttivoRTi Rs, INc.
w w.3no Natninwide Gncrage W 3 % (4M 1
- - - ~ - - - - , - - - . _ _ . - - , , _ . . - , - . ~ - ,. - - - -- , - _ - --.--- - -..-.- -_- - ..,- ,_ , , ,_ _ . , _ ., - , - - .
- _ _ - +- 4 * = 4 7280 01 01 330 A
(_j7/bc 1 costs is supplying power capability. If you have to go l
2 [ above 10 megawatts, the cost becomes prohibitive. So you're 3 limited to 10 megawatts or less.
4 The other biggest cost is probably in the steam 5 generator with the number of tubes that you have fabricating 6 the steam generator.
7 DR. TIEN: But you limit that to 5007 8 MR. WOLP: We limited that to 500 tubes, yes.
9 MR. SULLIVAN: You're talking about -- I don't 10 know -- the semi-scale point. I assume that when Keith 11 presented the presentation yesterday, it had semi-scale and 12 semi-scale actual.
13 What you do is you look at this stuff and you 14 say, okay, I think I'll change the scaling in the hot leg 15 pipes or the cold leg pipes.
16 So you change the size of the pipes because you 17 fool like you get by with it. And that's the actual that 18 you showed yesterday. Right?
l 19 MR. CONDIE: Right.
20 MR. SULLIVAN: Are you guys going to do that to 21 this?
22 MR. WOLP: This one we don't have to. We can 23 have an exact L over D ratio of 2 throughout the whole 24 system. We don't have to go in and increase the pipe size.
( 25 MR. SULLIVAN: You might want to do it because Act: Fimitnai. Ritronriins, INC.
3 202 uwm Nanonwide nneraee xm unsan
4 i 7280 01 01 '
331
( 7/bc 1 you want to decrease the length and get the volume, so the l 2 velocities are correct. There are some things you want to 3 adjust locally.
4 MR. WOLF: There may be things that we can do to l
5 increase the velocity in the loop piping, for instance.
6 That's a consideration.
- 7 MR. SULLIVAN
- Are you going to do that?
8 MR. WOLF: We haven't decided that.
l 9 MR. CONDIE: Maybe I could answer that. On thoso 10 facilities that have very small diamotor ratios, such as 11 semi-scale, or very largo longth to diameter ratios, you can 12 increase the diameter in certain portions, like the hot log,
(} 13 without making appreciable changes in the volumes.
14 If you got a facility liko we're talking hero, 15 that's short and fat and approaching 250 cubic feet, you 16 realizo if you increase the diamotor of the pipe that's 17 already six inches in diamotor, that's what this one would 18 be -- 6-1/4 inches or something -- increase that by some 19 factor and then you really start to distort the volume.
20 When we'ro looking at this facility, you'll soo 21 wo're not looking -- you can't got heat transfer scale.
22 We've said we're going to use the data we've got from semi-23 scalo.
24 So we are very concerned about the mass
() 25 distribution and things like that. So it would probably be Aa.17tintiltai Iliti>oitTiins, INC.
- o .u?.17m Na:nmwide carne sn.1% ua6
' 7280 01 01 332
()//bc 1 fairly undistorted. It would be only in very limited places
- 2 that we would make some actual change to the ideal scale 3 dimensions.
4 MR. WOLF: I you look at semi-scale and MIST, for I 5 instance, with their L over Ds of 41 and 28, a little chango 6 in the hot leg piping diameter doesn't affect the overall 7 system volume.
8 But, where we're talking something like 6 or 7 J
! 9 inch hot log piping already, it would have major effects.
]
10 DR. CATTON: You indicated earlier when I asked 11 the question about pressure you were primarily interested in
! 12 the hydrodynamics effects. This was your argument.
j () 13 MR. WOLF That was a quote from Peter Grif fith 14 at our scaling meeting. ,
i 15 DR. CATTON: There are about seven or eight l
16 forces that one plays around with when you try to scale 17 hydrodynamic effects. You get things like Froud numbers, 18 and so forth.
! 19 I don't see any of that. Did you do that sort of
! 20 thing?
i 21 MR. CONDIE: That falls out of the scaling laws l
22 as they are developed. If you go back to the scaling i
23 report, you can see -- ;
24 DR. CATTON: I was looking at this table.
j i () 25 MR. CONDIE: You take those and you preserve from I
l AcitFrini:1<Ai. Riii>oitTitits, INC i ( :C W.Pm N.mnwide Gneuge km 14 Nati
7280 01 01 333
()7/bc 1 the original analysis. It comes up with those dimensionless 2 numbers and it shows which ones are preserved. And from the 3 preservation of those, then these actual relations are 4 developed.
5 DR. CATTON: I'll find that in the report?
6 MR. WOLF: Yes.
7 DR. CATTON: Good.
8 DR. TIEN: I think, in your scaling report, you 9 were concentrating on a different local phenomenon if you 10 proposed this new working concept, you know. This facility.
11 You mentioned you need additional analysis.
12 My feeling is perhaps you actually need some 13 analysis even now to justify this proposer working concept 14 before it gets too far. You want to calculate the overall 15 phenomena, not just the local phenomena, like what you have 16 done before.
17 MR. WOLF: One of the things that we're doing 18 right now is we're using the best took which we have 19 availabic, which is the TRAC code, to try and look at a 20 couple of offects.
21 We will not actually start the conceptual design 22 work until we've reached agreement with all the interested 23 parties that, yes, this is the best set of compromises that 24 could be made.
() 25 They say this is a working concept and is by no Act!.Fl!DliRAI. Ill!POR HIRS, INC.
202.m.nm Nanonwide Cmcrage Ned%fMi
l l
7280 01 01 334 O
s_s//bc 1 means cast in concrete right now. But we need a starting
! 2 point for further discussior.s on the facility design.
3 But, if you look at the L over D ratios of past i
4 facilities, you see that even LOFT tended toward the tall, 5 skinny facility type, with ROSA-IV probably being as close t 6 to our proposed facility as any at 6.9; whereas, the
! 7 facility we have proposed has the L over D ratio essentially 8 of 2.
J
, 9 (Slide.)
10 To get back to your question, Dr. Tien, did we j
11 look at the other end of smaller L over D ratios, yes, we 2
12 did. With pure linear scaling and a maximum available core 13 power of 10 megawatts for full-scale power, you are i
14 limited -- you are very limited on the type of facility that j 15 you can build. The 10th linear scale essentially shrinks 16 all of your plant dimensions by a factor of 10; instead of i
17 having a 12 foot tall core, you've got a foot and a half 18 tall core.
19 Your steam generators are five foot tall instead l
20 of 50. Full-scale power for the 10th linear scale is 21 roughly 10 megawatts. You're putting it into something 22 that's a foot tall and 6 inches across.
23 And I really wouldn't want to be around there 24 when they turned the power on. And that's determined by
( 25 your heat flux ratio of 2.16. You are seeing subcooled ace FEDERAL. IlEPORTERS, INC.
202 m-rm Nationwide cmerage un6 tu,
7280 01 01 335
( 7/bc 1 nuclear boiling in the core with this type of a heat flux 2 ratio.
3 The time ratios are very distorted - .32. Since 4 the steam generators are so short that within any type of 1
5 reasonable sized linear scaling, there's questions whether 6 you've got enough driving head for your natural circulation.
7 The lengths of piping involved are really 8 probably too short for development of any kind of reasonable 9 flow regimes, or for any type of phenomenon to develop.
10 The components are too closely coupled for any 11 kind of instrumentation at all to be installed; plus, it is 12 very difficult with these very short vessels and steam
( 13 generators to make any kind of normal type of liquid level 14 measurements, using differential pressure cells.
15 The difference in liquid heads are so small that 16 they'd be practically impossible to measure. So those kind 17 of considerations kind of pushed us the other direction, 18 away from strictly linear scaling.
19 And as a compromise, we settled somewhere around 20 2 and that's flexible.
21 (Slide.)
22 Just briefly, to discuss some of the comments 23 that our consultants have made, the 2.0 max is probably our 24 interpretation of the comments that were made at our review
() 25 meetings. But, overall, the thought was that bigger is ACli Frintinat Riii>onTrins, INC.
j 202m Mm Nahonwi& Cmeage Wi tMM6
)72800101 336
()//bc 1 better. You're better off with the biggest facility that 2 you can afford to build, even if you have to make the l
3 sacrifice of full-scale power for decay heat.
4 Remember...
5 (Slide.)
i 6 ...the plot that we showed that you could afford 7 a much bigger facility if you can turn that 10 megawatts of f
8 available power into 5 percent decay heat than if you could 9 if you were having to go full-scale power.
10 As you can see here, we've got a much larger 1
11 facility with 5 percent decay heat than with the same 12 parameters holding constant if we'd have come down here and 2
( 13 said, okay, we want our maximum available power. It would 14 be similar. We'd be somewhere down here on our volume.
! 15 Roughly, somewhere on the order of 100 cubic feet.
16 Another consideration that you need to make when i 17 considering the capabilities of the facility is what type of l 18 support facilities do you already have in place so that you 19 do not have to go back from scratch and totally rebuild.
20 (Slide.)
21 At our water reactor research test facilities 22 here, we do have available to the building three megawatts 23 of available core power without additional upgrades to the i 24 power distribution system.
25 If you take that three megawatts of availabla 1
Acit.17iinititAi. Ri:PonTitits, INC.
j l 202447mm Nationwide rmerage m HMM6
I l
1
- 7280 01 01 337 l 1 /'T
\ss1/bc 1 power, you can look at certain options that are then open to i
l 2 you. You can have a length ratio of .2 with a longth to 3 diameter ratio of .l.
4 In other words, your strictly linear type scale.
5 You could have a longth ratio of .3 with a length to 6 diameter ratio of 1.5, or you could have a length ratio of 1 7 .375 with a longth to diameter ratio of 2.
8 And this is our recommended concept there with 9 that 3 megawatts of available core power simulating decay I 10 heat rather than full-scale power.
i 11 DR. CATTON: This is full pressure, isn't it?
12 MR. WOLF: This is full pressure.
( 13 DR. CATTON: Did you generato any tables liko
- 14 this with, rather than full pressure, a thousand psi?
15 MR. WOLF: Did we? No.
16 DR. CATTON: Fivo hundred psi? You didn't?
I 17 MR. WOLF: No. We could. Thoro's no reason why l 18 we couldn't do that.
i j 19 DR. CATTON: Because I understand the tradooff
- 20 you've made betwoon bigger and power. But it sooms to me 21 you've got a throo paramotor problem. You've got bigger 22 power and pressure. And anything you give up on any one of 23 them, you lose a bit. ,
I 24 So it's kind of a surface that you nood to
) 25 optimizo rather than a two-paramotor problem.
! Aci!-17tint!RAI. Riii>onTiins. INC.
i I 20214 moo moonmae unn ve m tti, ta, I
7280 01 01 338
( //bc 1 MR. WOLF Those were easy to do because if you 2 assume the full pressure, all those parameter ratios are f
, 3 one. But, yes.
4 DR. CATTON: You can give on some of the j
5 parameters to gain somewhere else. I certainly would like 6 to soo that.
7 (Slide.)
8 MR. WOLF What I have here is a little more 9 detail on what the various paramotors for this working 10 concept to turn out to be.
11 As you 800, for right now, it's full pressure, 12 full pressure secondarios, primary system volumes on the 13 order of 150 cubic foot.
l 14 In other words, we're talking about again a very i 15 largo facility compared to the Semi-scale and MIST facility i
j 16 that we have. Ilot log diamotor piping is 6.75 inches. Cold j 17 log, 5-1/4. Our heator rod longth is now 54 inches as 18 opposed to 1-1/2.
J ,
l 19 That corresponding reduces our heat flux by the i 20 name amount. Our vessel insido diamotor is 32 inchos as i
- 21 opposed to rou0hly 6 inches. Our downcomor gap is 1.8 j
22 inchon.
23 So it's a reasonably sized downcomor gap. We can
! 24 develop the type of phenomena in the downcomer which we I( 25 think will exist. Our number of steam generator tubos is A n 171!Dl!RAl. RITOR I'l!RS, INC.
- o:M.Uon Nanonwale Cmerage m 3 % 9 ,u, 1
1 1
l 7280 01 01 339
( //bc 1 546. The tubo longth is 19 foot.
2 DR. CATTON: The downcomer gap width, did you 1
3 take a look at what kinds of things have boon speculated to 4 occur in there and see whether that 1.8 inches is off?
5 MR. WOLF: Right now, no, we haven't. This 1.8 6 falls directly out of the L over D of 2.
i 7 DR. CATTON: Because you might want to oversize I 8 the downcomer.
l
- 9 MR. WOLF
- Yes, thoro's a lot of options thero.
10 DR. CATTON: I think some of those things are a 11 good idea.
12 MR. WOLF Soo, those are details that you'd nood I
() 13 to be aware of when you actually sit down and start tho -
L 14 draftsman designing the conceptual design.
15 floy, we maybo want to oversize the downcomer.
16 The rest, our heat flux ratio, is now 1.6 instead of over 3 17 and you soo that, with this type of geometry, we've got i 18 reasonable velocity or timo ratios as compared, you know, to 19 strictly the linear type facility.
i 20 MR. SULLIVAN: Lot me toll you what bothers mo i
21 about some of this. If you go back to Keith's presentation
! 22 of yesterday, at Somiscale, they captured 64 percent of the t l 23 major phenomena. Now we've moved down to 38 porcent.
24 So we're making the problem worso. We're gaining l
() 25 in something, but we're making it worso. If you look at the Acti 171 :niinai. Riii>on nins. INC. ,
, 201 tUSm Nationwale Com,qe am )M ,4m I
~ . . _
7280 01 01 340
( //bc 1 time ratios, the thing is going to be about twice as long as 2 the transient will be in plant.
3 MR. WOLF It runs faster.
4 MR. CONDIE: Forty percent faster.
5 MR. SULLIVAN: Yes, it's 40 percent fastor. So 6 what you're doing is all the phenomena that are important 7 are going to be much fastor and you're going to be able to 8 got through them quickor.
i 9 So, even though you look like you're doing well 10 on a code comparison, it may not be, becauso you've run 11 through the phenomena so fast, you're going to be off in 12 flow and flow regimos, particularly, are really important.
( 13 So you're going to have a distortion in the flow I
14 regimos.
15 MR. CONDIE: liarold, I disagroo with that. If ,
16 you look at the local analysis, the horizontal flow regimo l 17 map, distortions are actually better than in the Semiscale 18 or in the full height.
19 That's one area we gained in this, is in the 20 horizontal flow regimo map. Neither one of them do very :
i 21 well in the vertical flow regimo transitions. So the areas 22 that this does good in are a real complomont to those areas 23 that the other one did not do as good int ovon though the 24 ovorall porcontago of phenomena captured is not as great /
( 25 MR. SULLIVAN: You've looked at all the phenomena Arit-17lti)itari. Itifi>oa ritas. INC, q 2e m um Nanon%Ie rmcrage 8m3mto,
7280 01 01 341
()//bc 1 that we are really interested in now, and you think this
- 2 facility does a much better job.
3 MR. CONDIE: Considering the data that's 4 available; if you didn't have any heat transfer data ,
i 5 available to you right now, I would say no, you don't want l l
l 6 this.
i 7 But, you look at the concerns that we have in 8 multi-dimensional offects and heat loss effects and the fact 9 that most of our concerns are with long-term transients, 10 with multiple failures and operator interaction, whero we're 11 out in timo a little bit, whero our experience with 12 Semiscale with its heat lossos has caused us to be able to
() 13 not simulate those transients.
i 14 This complomonts those right down the lino.
15 MR. SULLIVAN: And the steam generators you feel J
16 the samo way about?
17 MR. CONDIE That's correct. Sure, I do.
18 MR. SULLIVAN: I really don't in the steam 19 generators because that in one of the problems that we're
.i 20 having problems with leaking up now, particularly the once-t 21 through steam generator.
22 And you've really distorted the steam gonorators.
23 MR. CONDIE: Just in terms of longth.
24 MR. SULLIVAN: flut that leak and the heat 25 transfer in the major thing that wo don't understand well ACli I71il)liRAI. Riti>oR i t Rs. INC.
- o: wxm N.uionode Oncrue s m ut,t a n
l 7280 01 01 342 b
s_ 1/bc 1 now.
2 MR. BECKNER: Harold, recognize that we're doing 3 power to volume separate effects tests in that area. Wo 4 recognize that's a problem, but we're doing separate effects 5 testing.
6 MR. SULLIVAN: But all the ones, all the concerns 7 that MIST is bringing up now are things like steam 8 generator, the driving head. And those are the things you 9 distorted.
10 MR. BECKNER: But we are doing a separate offects 11 test at full power in the MIST generator to address those 12 phenomena. Okay?
) 13 We realize that is a problom. We agree with you.
4 14 We've got plans to address those phenomena; in offect, power 15 to volumo type scaling, which gives you the heat transfer, 16 which you may miss in this type of facility.
I 17 So it's a tradeoff. And we're trading off 18 separato offects for integral facilities, and so forth.
l 19 MR. SULLIVAN: I realize that. But, one of the 20 things that I bollove is that the separato effects ought to 21 try to got the modols right in the codos and the integral
- 22 tests wore to demonstrate that all of this works well 23 together.
- 24 You can't model steam generator exactly with tho
( 25 computer codes. So you nood something to go back and make I
Acit l'i:i)iinAi Itifi>oirii as. INC.
- , e mxm LoonsuJe Omeuge w 1 % u,ui
7280 01 01 343 (O_sV/bc 1 that final check with.
2 MR. BECKNER: Right. And I think our test, that 3 is, the final synergistic length that will give that element 4 together, will be distorted by 40 percent, or whatever, in 5 this area. That's right.
6 And, hopefully, the codes can handle that. I 7 don't know. We did a good job developing the model on our 8 separate effects test. The final test will be distorted 9 some in this area.
10 MR. CONDIE: Harold, you have to remember when 11 you look at all of these scaling relationships, they all 12 come to the same situation where the void fraction 13 distribution is maintained. So the void fraction 14 distribution in the steam generator is from top to bottom 15 consistent with what you have in the large plant, albeit it 16 shrunk. The actual lengths are, in fact, different.
17 The thing you gain from it is the size in terms 18 of the area. And the chance to develop multi-dimensional 19 concerns. And the fact that you don't have the tremendous 20 heat loss that we know has played havoc in the long-term 21 transients that we've done in Semiscale.
l 22 And any of the full height volume scale 1
23 facillites. Because that area to volume ratio is just 24 unreal. We've done everything we knew how then to control
( 25 that.
i
! Acti Fitolinai Riii oniiins. INc I 20:M x00 Nahonwide Cosaage W1%fM.
4
. _ _ _ _ . . _ . _ _ _ . _ _ _ _ , . . . _ . , . _ . .._.~.____ _ __. _ _ _ , _ . ~ _ _ _ _ . . . , _ . . _ _ _ . . . . . _ _ _ . . , _ , . _ . _ . _ . _ . _ . -
7280 01 01 344
()7/bc 1 In fact, we've almost eliminated it in a facility 2 this size. I reiterate it really complements the facilities 3 that we now have, complements the phenomena that we can 4 properly simulate.
5 MR. SULLIVAN: And I would agree with that.
6 (Slide.)
7 MR. WOLF: Let me finish up here by just 8 reviewing some of the comments over the time that we've been 9 putting our scaling report together that have been made by 10 our various review groups.
11 As Keith mentioned yesterday, one of the 12 questions was:
( 13 If you went back and viewed the phenomena 14 differently, would that change the overall ranking of your 15 various facility concepts?
16 As Keith mentioned yesterday, we did it about 17 four or five different ways, ranking the importance of 18 different phenomena.
19 And the relative rankings between top and 20 bottom stayed the same. There was a little bit of shuffling 21 in the middle, but nothing that really had much effect on 22 the overall ranking of the various facilities.
23 The linear concept that was proposed, I think the 24 concept was that this will best satisfy the type of data
() 25 needs that we're looking at because of its faithful Acn-FuniinAL Rui>oninns, INC.
202.m-Um Nationsside roserage xu o w u>m
. .__- .. _ . ~ _ -. - - . --
> 7280 01 01 345 j ( 7/bc 1 geometrical similitude. But you run into problems trying to 2 implement that in reality. And the fact that about the
( 3 largest type of exact linear facility you can afford is
, 4 somewhere on the order of a tenth scale.
- 5 And as I mentioned, there's a lot of problems l
l 6 that make something that small just impractical to build
- 7 from a physical point of view.
8 So, as a consequence, we back off slightly from l 9 an L over D ratio of 1 to something on the order of 2/1. In 10 the area of full-scale power versus size, requiring full-t- ,
i 11 scale power puts severe design constraints on the type of ,
I l.
12 facility you can build.
13 You just can't afford to power at full-scale I 14 power a big facility. You're limited to something on the 15 order probably of either the 10th linear, if you want to go 16 to that type of scaling, or something on the order of MIST 17 or MOBI if you want to go to the power volume scaling on the 18 other end.
I 19 MR. EBERSOLE: Can I ask a question?
20 Fission power is abruptly cut off except for 21 delayed fission. Without having full power, but with 22 storing energy in the system, at a smaller rate, of course, 1
1 23 and synthesizing the stored energy content, but not the l
! 24 power, can't you approach a large LOCA potential?
( 25 MR. WOLF: You could approach it, yes.
i l
ace-FEDERAL REPORTERS, INC.
202 347 3M) Nanonwide Coverage Mn 33MM6
. - - - . . - . _ , _ - , . . - - . . , . , ~ , , _ _ - _ _ . , . - . , , , - . . . - _ . . , . - - . - _ , . - - , - . . - . , , , , . _ . - - , . - . . , .
., . _ _ . - . _ . . , - =. . .. .,
7280 01 01 346
( V/bc 1 MR. EBERSOLE: Do you follow what I'm trying to 2 do? to do? I'm going to synthesize the stored energy to 3 take the place of the absence of full power by whatever 4 means I want.
5 I'm going to say my initial conditions, it 6 doesn't look like it was full power from the stored engine 7 concept. Then I'm going to argue that my fission power I'm 8 going to cut off anyway.
9 MR. WOLF: You could approach the energy
]
10 deposition conditions for a large break LOCA by doing that.
11 MR. EBERSOLE: Why isn't that worth something?
12 MR. WOLF: Because one of the major transients 13 that we've said that we can't do really well is the large 14 break LOCA because you don't have the right initial energy 15 deposition.
16 Now, you could probably work around that problem 17 by doing just as you suggest. There are some other l 18 transients that you can't do a good joo on the front end if 19 you just have decay heat.
20 For instance, a transient that results in an 21 initial primary over-pressurization, like on a feedwater, 22 where you have some number of full power minutes before you 23 trip on high primary pressure, for instance.
24 But, once you get past those initial few minutes,
( 25 you can set up the right type of initial conditions to 2
ACE FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coverage 800-336-(M6
7280 01 01 347
()V/bc 1 simulate the backend of the transient. In fact, decay heat 2 is how just about all of the Kerr facilities are operating.
3 The final conclusion was in keeping with the idea 4 that bigger is better, that you would be better off from a 5 phenomena point of view by having a larger physical-sized 6 facility and going with, you know, 5 percent decay heat for 7 initial conditions; as opposed to a smaller type of volume 8 facility and being able to simulate the full power.
- 9 So that concludes what I wanted to say. Our 10 working recommendation then, as I mentioned, this is a i
11 proposed recommendation. It's something for people that we 12 wanted critiqued. And before we actually are committed to
( 13 design work, we want to have the best possible compromises
! 14 that we can make in our scaling rationale.
- 15 So we're open for any type of comments and i
16 discussion that we can get.
1 17 MR. MICHELSON: Any comments? Or, any other l 18 comments?
l i 19 (Laughter.)
20 MR. WARD: Yes. It seems to me that what they're 21 trading off here, you know, is kind of a radical shift from
', 22 the Semiscale, the traditional concept, you know. What l
23 they're trading off is giving up timing, for similitude in
[
24 timing to get all the dimensional effects. That's sort of
() 25 obvious.
1 ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6M6 i
, . . . . . _ . _ - , ~ . . _ . _ . , - . , _ . . _ _ , . _ _ _ _ . . , _ . .
.i 7280 01 01 348
()V/bc 1 And I guess the question there is: Which is more 2 important to have for the purpose of an integral experiment?
3 If we remember the purpose of an integral experiment is to
! 4 give an opportunity for, I guess we could' call them 5 synergisms or interactions of parts of the system that show 6 up experimentally, first of all, and then to demonstrate 7 that the codes can in fact reproduce those,synergisms and 8 interactions within the system.
9 It seems to me it's not obvious giving up the 10 timing similitude is a good thing, except, you know, we've i r 11 got 20 years of history with that sort of similitude 12 available, and we haven't had the benefit of 3D similitude.
() 13 So it seems to me I guess I'm sort of persuaded 14 that it's a healthy thing. In essence, they're going to be 15 trading one set of blind spots for another set of blind 16 spots. That'c prooably a good thing to do as long as there 17 is some institutional memory that keeps going on.
18 The second part, I'm given this with high or low 19 pressure. And, again, when you go to low pressure, of 20 course you're gaining a lot in cost, practicality, 21 flexibility of the system.
22 What you seem to be giving up is this big 23 disconnect with phase change. But I guess there you have to 24 look again at the purpose of the integral system, the
( 25 integral experiment.
- AcEJFEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6M6
- . - . . _ , . - . _ . _ _ _ _ . , _ - , ~ . _ _ _ . _ . -
7280 01 01 349
( V/bc 1 Is the purpose of the integral experiment to 2 directly demonstrate phenomena, interactive phenomena, let's 3 say?
4 Or is it to provide a test for the codes, so that 5 the codes, so you can have some confidence that the codes 6 demonstrate these interactive phenomena.
7 MR. WOLF: Really, it's both, although the 8 transient is running at faster than real time. If you want 9 to look at any type of recovery procedures or accelerated 10 operator actions, you need to be able to simulate the 11 transient from front to back, without having to go through 12 the various steps that you would require with a low pressure
( 13 facility of either adjusting break sizes as the transient 14 progressed, or adjusting core power.
, 15 MR. WARD: It just seems to me then, I suspected i
j 16 it might be a reasonable thing to go to a low pressure 17 system, although it would be tougher to interpret the data, 18 to analyze the data.
19 I think you pointed out, Jim, that it would be 20 more difficult to make sense out of the data; that that's 21 only if you're trying to directly use the experimental 22 results in understanding the plant.
- 23 And if you're really serious about using the 24 codes to make that transition, then I'm not sure you're
( 25 giving up as much.
ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 8(0-336-6M6
. _ , . . , , _ - _ _ _ _ _ _ _ _ ~ _ _ _ , . -
7280 01 01 350 (h
(_s1/bc 1 MR. WOLF: I think there's a misconception on how 2 much you gain costwise and ease of operationwise by going to 3 the, quote, " low pressure".
4 Any time you 're talking several hundred psi, 5 steam water type systems, you don't gain that much cost 6 savings by having the same size facility operating at 2,000 7 psi. You're looking at Schedule 160 pipe or maybe Schedule 8 40, but it's still in all a welded system.
9 You have the same type of high pressure 10 instrumentation and you have to observe the same type of 11 safety procedures and safety precautions.
12 So I think, as soon as you step over into the
( 13, above atmospheric pressure, steam water type of facilities, 14 I think that the cost difference between 2,000 psi full 15 pressure and three or four hundred psi is really drastically 16 overrated. There's really not that much of a cost
(
1 17 difference involved.
l l
18 DR. CATTON: At what pressure can you go to a 19 bolted system?
i f
j 20 MR. WOLF: I don't know.
l 21 Steve, do you have any idea on what the code 1
22 requires for a flange and bolted system versus an all i
23 welded?
24 We have flange and bolts on Semiscale, flexitalic
() 25 gaskets. So you trade off system leakage there if you have l
ACE FEDERAL REPORTERS, INC.
j 202 347-3700 Nationwide Coserage 8%336-6M5
,, - - , , - - _ _ _ _ . . _ . _ _ _ . . . _ . _ - ~ _ - - _ _ _ - - - - _ . _ , . - - _ . , _____.~ . . _ _ _ . . -
0 7280 01 01 351
()7/bc 1 flanges, and things like that.
2 DR. CATTON: What pressure do you have to go down 3 to then where you could feel comfortable with a bolted 4 system and not worry about the leakage? I mean...
5 MR. WOLF: If you start looking at the number of 6 thermal cycles, it's the thermal cycles that really kill you 7 on a flange system; the continual heat-up and cool-down.
8 DR. CATTON: You see, I think what's going to j 9 happen in the future is you're going to wind up having to 10 look at something other than this B&W system. Oconee.
l 11 That's just as clear, and you know it.
12 So, flexibility is important. In my view, a lot
() 13 of emphasis ought to be given to what contains the device 14 you're going to study. Power supply is important.
15 Instrumentation. Above all, people.
16 Those are the important parts. I think you back 17 off a little bit on what you're doing with the actual device 18 you stick in there. And it should be flexible.
1 19 That's when you say all welded system, I imagine 20 something that's out there and it's going to be there for 20 21 years. And no matter what the problems we have, it's going 22 to be very difficult to rationalize a change because it's 23 going to be so expensive.
24 If you give up a little and go to a bolted
- () 25 system, well, you don't like the steam generator? Take it ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-(M6
. _ .- _= _ _ ._ _. ._ _
7 7280 01 01 352
( //bc 1 out. You've got flexibility.
i 2 We have a long history of trouble from exact i
3 interpretation of results from these systems. These guys 4 here know what's happened with Semiscale board notices and 5 all that shouldn't have happened because of this kind of 6 interpretation.
- 7 I would rather see a system where you just say 8 flatout
- We cannot interpret this as something that's going 9 to be defined in full scale. The code is in the middle, as 10 a necessary piece of the interpretation.
11 I'd rather see that. Then we'd avoid all this.
i
- 12 MR. WOLF
- We try to always make that disclaimer.
(_) 13 DR. CATTON: Well, you were making it a little f
14 while ago. Right? I'm not sure you do it conscious 1'1, but 15 it was being done here.
16 One of the reasons for full pressure is, one of 17 the reasons we want timing is:
- 18 Well, that is, is where we get into trouble.
l 19 MR. WOLF: You like to have those as close as
}
! 20 possible, I would think. You don't want to have an apple 21 and an orange.
22 DR. CATTON: Of course not. But your primary
( 23 concern is that when you run one of these transients, 24 whatever you're doing to this device, it's going through the 25 same kinds of processes. The timing is going to have to be i
! /\CEJFEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 804336-646
7280 01 01 353
( 7/bc 1 right. A lot of things don't have to be right, because the 2 code will adequately interpret it.
3 If the code can't adequately interpret it, we've 4 got other problems. That doesn't quite fit with the 5 philosophy that you're presenting.
6 MR. WOLF: For the flexibility question, if you 7 look obviously at the difference in design between AB&W and 8 a Westinghouse plant, for instance, several years ago, when 9 this concept first came up, the idea of a modular facility, 10 where you could set your tinker toy parts together to go 11 from a B&W to a Westinghouse was brought up.
12 The conclusion there was that the design
( 13 differences were drastic enough that we really couldn't do 14 that. We had decided that to handle -- well, as you say,
- 15 the differences that are going to come up, so that we're not 16 stuck with one facility to handle all problems is to have 17 the capability as far as support facilities -- the ECCS 18 systems, the data acquisition systems, the available power 19 supplies, core power instrumentation -- to be able to, if we 20 needed to 15 years down the road, replace the B&W plant with 21 the Westinghouse plant.
22 DR. CATTON: I don't think it's going to be 15 23 years. It's going to be a lot sooner. I can't imagine l
24 needing 15 years for the B&W configuration. We've got all
( 25 those Westinghouse and CE plants out there.
1 ACE FEDERAL REPORTERS, INC.
l 202-347-3700 Nationwide Cmerage 800-336-6M6
_ , . - - _ _ , _- . . _ . _ _ _ _ _ , _ _ _ _ - . _ _ _ _ , _ _ _ _ _ ~._ _ _ _ _ _ . . _ _ _ _ . . _ _.
7280 01 01 354 ,
A)//bc i
s_ 1 What about a boiling water reactor system?
2 Fifteen years. Well...
3 MR. WOLF: We're trying to plan as much 4 flexibility into the initial B&W facility as we can as far 5 as the control systems. Current plans call for leaving 6 Semiscale currently where it is and building a new facility 7 wherever the B&W design facility in the pit next to it.
8 Later on, if, for instance, we need a 9 Westinghouse plant, all those subsystems are then available.
10 And you don't have to replace. You just have to build the 11 loop itself.
12 MR. SULLIVAN: Could you go over the raised 13 loop / lowered loop decision?
15 MR. SULLIVAN: What you proposed is that you're 16 going to model Oconee. Right? How did you decide on the 17 raised loop / lowered loop issue?
18 MR. WOLF: I'm not sure that we consciously took 19 that into consideration. There are more of the lower loop 20 plants available.
21 MR. BECKNER: That hasn't been addressed, Harold.
22 We wanted a model to study. ,
23 MR. CONDIE: It was addressed in the front of the 24 report.
25 MR. BECKNER: We're not just deciding to build ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 80h33M646
d
't 7280 01 01 355 i ()//bc 1 that option, necessarily. We're deciding to study this 2 option. I don't know that if we came back to NRC and we 3 made the decision, then we'd have a lowered or raised loop.
4 MR. CHARLTON: I haven't heard that decision; 5 however, there are seven like Oconee and one the other way, 6 like Davis-Besse. I mean, that.
7 MR. WARD: Which is Belefonte?
8 MR. BOEHNERT: That's raised.
9 MR. CHARLTON: Belefonte is not like Davis-Besse 10 either. It's unique also.
11 MR. WARD: Isn't there also an argument, I guess, 12 that the lowered loop plants are inferior? I'd say they
( 13 are.
I 14 MR. MICHELSON: They have a different set of 15 problems and they're potentially more serious.
16 MR. BECKNER: I want to make a comment on the 17 pressure scaling. The facility Jim has described here has, 18 in effect, been built and is operating at lower pressure 19 right now at Maryland. I think it will be very interesting 20 over the next year or so to see how it gets around some of 21 the problems that were raised in the scaling issues.
22 When the scaling study was done, which was done 23 after the_ facility was built, we were very concerned about 24 this discontinuity. Since that time, Dr. Hsu has been
() 25 working on this.
ACE FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coserage 800-334646 t
,v.- - . . , . - ,n.. , , - , , - , . . - - , . , - . - - , - , - -
7280 01 01 356 m
(_jV/bc 1 And I think it will be very interesting. He can 2 tell us a lot about the scaling procedures, so see how 3 successful he is and how useful his data is.
4 We don't have to do design studies or code 5 calculations when we have the data coming out to see how 6 useful his data turns out to be.
7 And we'll also have some comparisons, or at least 8 an attempt to compare that scaling with the full power, full 9 heigth power to volume, full pressure MIST.
10 So I think that pressure issue is being addressed 11 well experimentally.
12 MR. SULLIVAN: Bill, how did you come up with the 13 B&W design as the one you wanted to do?
4 14 MR. BECKNER: At this point in time, that's the 15 plant that we're most concerned about. I don't think we 16 made the decision that this facility would be B&W 17 necessarily, but it's the facility that we're going to study 18 as far as designs and concepts.
f 19 Earlier on, when Dr. Shotkin put up his future 4 j 20 issues, he had the full thing, advanced plants and so forth.
21 I don't think we made a decision to cut metal on the B&W 22 facility yet. That's the one we're concerned about today.
23 MR. WOLF: Bill, wasn't that one of the reasons 24 that you did not request owners group participation in ISIF
() 25 was that it may turn out to be a Westinghouse plant when we 4
ACE FEDERAL REPORTERS, INC.
l 202-347-3700 Nationwide Coserage 800-336-6646
r 7280 01 01 357
//bc 1 built it?.
2 MR. BECKNER: Right. We specifically did not 3 request owners group participation at this point. We're in 4 a concept and, really, not even a facility conceptual 5 design. We're in a scaling type conceptual design.
6 DR. CATTON: I think your indecision is a very 7 strong argument for flexibility.
8 MR. BECKNER: I don't disagree, Ivan.
9 DR. CATTON: I think there were just three 10 parameters -- bigger, pressure and power. It seems to me 11 there's a fourth. That's the downstream cost of having to 12 go in and make big changes versus the cost today in ensuring
- ( 13 that you have the flexibility.
14 I think you ought to shoot for flexibility.
15 MR. SULLIVAN: Have you considered building a
! 16 Combustion design?
17 MR. BECKNER: Not specifically yet, no.
18 MR. SULLIVAN: You know, you do have data on 19 Westinghouse and MIST designs. However incomplete it is.
l 20 And you're going out and doing a whole bunch of separate l
- 21 effects or proposing to do separate effects on the steam e
i 22 generators in the B&W design.
23 Yet, we have never run a Combustion design.
1 l 24 MR. BECKNER: We've always piggybacked CE plants
() 25 on Westinghouse data.
I
! /\CEJFEDERAL REPORTERS, lNC.
'02 347-3700 Nationwide Coserage 800-336-(M6 I
7280 01 01 358 (s_hsV/bc 1 DR. CATTON: And your study that I referred to 4
2 yesterday showed that there is a distinct difference in the 3 margin that you have.
4 MR. CHARLTON: But is that really based upon the 5 way that the plant is thermally hydraulically different? Or 4
6 the boundary conditions and initial conditions that that 7 plant runs at?
8 DR. CATTON: I would hope the person who did the
- 9 study tried to do it on a consistent basis.
10 MR. CHARLTON: The plants operate differently, 11 whether they're on a consistent basis.
12 MR. BECKNER: These were different calculations.
13 The big difference is the accumulators. They've got 200-14 pound accumulators on the CE plant.
! 15 DR. CATTON: So they didn't do the study. Right?
16 MR. BECKNER: What was the purpose of the study?
j 17 DR. CATTON: All I know is what they presented.
18 You guys presented it at the information meeting. And there
- 19 was a comparison of the three plants.
20 MR. BECKNER: The purpose of the study was to i
21 find out what a typical CE plant did and what a typical 22 Westinghouse plant did.
23 DR. CATTON: And a typical B&W.
24 MR. BECKNER: Since CE plants typically have 200-25 pound accumulators and Westinghouse has higher pressure i
Ace FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cmerage 800-336-6M6
7280 01 01 359 (D
s_s//bc 1 accumulators, that's what we used. So they were different.
2 MR. MICHELSON: And no PORVs on some CE plants.
3 And that makes a difference. And it's a two-loop versus 4 four-loop.
5 MR. MICHELSON: Do you think, even with all those 6 differences, you understand CE plant responses well enough?
7 MR. BECKNER: The transients can be different 8 because they have different subcomponents. But things like 9 the steam generators tend to be the same.
10 So, if the codes can make those differences, then 11 we're successful.
12 MR. MICHELSON: On the scaling studies, it wasn't O
(_/ 13 said explicitly, but I assume that your scaling work is 14 assumed to be a two-loop B&W arrangement?
15 You're proceeding as if there were two loops, and 16 you will build two loops?
17 If you go like to another plant, like 18 Westinghouse, are you thinking in terms of having all four 19 loops?
20 MR. WOLF: Yes.
21 MR. MICHELSON: Thank you.
22 MR. WARD: If you went through an argument like 23 you did today for a Westinghouse simulation or a CE 24 simulation, would you come to kind of the same conclusions
() 25 versus something that would drive you to a very different ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage Suk33MM6
7280 01 01 360
()//bc 1 conclusion?
2 MR. WOLF: I think probably pretty much the same 3 type of conclusion. We did go through the scaling report.
- 4 MR. WARD
- Right. You did, but you didn't go for i
5 the selection argument.
1 6 MR. WOLF: I think we would probably still opt i
7 for the geometrical similitude and the L/D ratio somewhere 8 around 1, 1 to 2.
9 MR. SULLIVAN: Bill, it looks like there is high 10 level input. Maybe it's even political input into selection
- 11 of things like the plant design that you were trying to
, 12 model, f.() 13 Maybe it has some technical information in it, j 14 too, on how well we think the codes do in particular areas.
1 15 But it looks like you haven't given that to them yet.
16 MR. BECKNER: I didn't follow you, Harold.
17 MR. SULLIVAN: There's some political and high i
l 18 level technical input that's needed to this study. And that i
19 is, what kind of plant am I going to model because of
]
20 whatever reasons?
4 21 And, for some reason, you have selected the B&W i
22 design and it's at least not clear to me why you made that l 23 decision. Because what's going to happen, the way I look at 24 it is, that you're going to have eight plants out of 110
() 25 modeled in this facility.
,I
/\CEJFEDERAL REPORTERS, INC.
l _
__ __ m-
i t
l i
7280 01 01 361
( V/bc 1 And do we think the risk in B&W designs is that 2 high that we're willing to exclude, or have less data?
, 3 MR. BECKNER: We don't disagree with you, Harold.
4 Like I said, this is a concept to study. We have not made a 5 decision to proceed with this plant. What you're bringing 6 up is a larger issue of the whole B&W plant.
, 7 We've got a tremendous amount of research laid 8 out in a tremendous number of plants. That's a larger 9 issue. And we are attempting to deal with that.
10 MR. MICHELSON: I think, in order to get through
].
j 11 the rest of our schedule, we will have to cut off discussion l 12 on this item.
13 Do you have any further statements under item I 14 four?
- 15 MR. SOLBERG: Yes. Let me get up and say a few
- 16 words.
l 17 MR. MICHELSON: Because we're going to lose
! 18 essentially all of our consultants by 4 o' clock. So it I
19 would be impractical to go beyond that.
I
! 20 (Slide.)
21 MR. SOLBERG: On the scaling study, I think it's i
22 important to note you've questioned why we've done f 23 phenomenological studies.
! 24 There's a very good reason for that. We feel 25 that, in a way, our codes may be somewhat biased because 1
i i
4 /\CE. FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coserage MO-336-6M6
. . _ . . _ _ _ _ _ , . . _ _ , . _ _ _ . . _ _ _ _ _ _ . . - _ _ . _ . ~ , . _ -. _ _ _ , _ . _ . _ . _ _ _ _ _ _ _ _ _ _ _
i 7280 01 01 362
(-)s N, //bc 1 they have been, if you will, tuned against all the past 2 experimental facilities which were done to a particular 3 design.
4 Therefore, we think that the scaling study is an i
5 excellent report. It did what we wanted. We continue to 6 recognize the need for separate effects facilities and, i
l 7 hopefully, then we'll build up the site and it will have the 8 capability to do both separate effects and integral testing.
9 So you have to consider the context of the i
10 scaling study. It is only one part of our decision-making
. 11 and we're now proceeding into the second part, which is to 12 do studies with TRAC, to try to show us how the system
() 13 behaves as an integral facility.
14 Clearly, what we've chosen, models of multi-j 15 dimensional response, we recognize that it creates problems
.i 16 in heat transfer and it may create problems when we try to 17 use it in our integrated disciplined approach when we look 18 at human reliability.
19 But it also gives us some things in long-range i 20 capability as well. The idea of pressures has come up.
21 We're thinking seriously of the pressure. I think the 22 problem I would see is that we've asked this to be an 23 integral facility. And if reducing pressure interferes with j 24 that, then it has not met our objective.
() 25 So, to the extent that we can reduce pressure and ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6M6
7280 01 01 363
( V/bc 1 still give us a long-range complete transient -- not little 2 pieces of transients, as we've heard at the University of 3 Maryland -- then that meets our objectives.
a 4 The next perhaps is I think the reason I'm up 5 here at this present time. We're working on what we 6 consider the NRC needs.
7 What are the needs for this facility?
8 You've asked the question. We haven't been 9 completely prepared to answer that, so we need to put those 10 together.
11 We would probably have a draft of that completed 12 by June of this year. We will start into the TRAC analysis
( 13 of the transient response of the baseline facility, which 14 you've seen today. That will be to evaluate small break 15 LOCA, main steamline break and loss of feedwater.
16 These will be taking place this summer. There f 17 will be, I'm sure, further tradeoff studies to optimize the i
! 18 results from the facility versus what we can find as the 19 means.
20 We would propose then to prepare a SECY paper 21 sometime later this year, to get Commission approval for 22 whatever final design, we would go along with.
23 Now, there are still some remaining near-term 24 problems in addition to these. Certainly, the power level
) 25 system tradeoff volume is one of them. The type of reactor i /\CE. FEDERAL REPORTERS, INC.
l 202-347-3700 Nationwide Coserage 8(x)-3 %6M6
- - _- _ _ ~ -. _ .- -.
4 7280 01 01 364
( V/bc 1 model we've just been discussing, we're not convinced that 2 B&W is the answer. It could be Westinghouse.
3 So that's a staff decision that we're going to 4 have to make. We're still trying to obtain NRR support. We i
5 continue to be interested in how this facility can support 6 th multi-discipline approach that we've started in TIC. We 7 see this as being a support facility for the TIC.
8 Therefore, we'd like to have it do as much in the 9 area of human reliability and procedural investigations as
] 10 we possibly can.
i~
11 MR. WARD: Don, you mentioned you are trying to 12 get NRR support.
13 MR. SOLBERG: Yes.
14 MR. WARD: What does that mean?
i 15 MR. SOLBERG: As of the present time, the NRR has 1 16 not indicated that they fully support this facility. So 17 they recognize the need for a long-term experimental 18 capability. But, in terms of this being what they require i
19 and the costs involved , we've not convinced them yet.
1 20 MR. WARD: Let's see. A lot of people have 21 moved. Whose left at NRR?
i i- 22 (Laughtr.)
23 MR. SOLBERG: We'll try to figure that out when 24 we get back home. But, starting first at the top.
( 25 MR. SHOTKIN: That was our procedure to convince i
ACE FEDERAL REPORTERS, INC.
202 347 3700 Nationwide Coserage Mn-336#46
- , _ . _ - . . , , , . _ . _ , , , , , ,,,____._y, , . __. - ,- _,__-..,7,_
- - - - ---e-.w-w,,,,,,,r-, -a--. .%.,---.--,-,we-. ,.m.-,- , ~ , . . . , , , , 3_-_-_,_.
7280 01 01 365 7/bc 1 them.
2 MR. MICHELSON: I thought the need was already 3 expressed. It's just the details. Which is it?
- . 4 Westinghouse, B&W or that sort of thing. But there is no 5 question about them believing that need at this monetary 6 level exists.
l 7 Is that right?
l
! 8 MR. SHOTKIN: I believe I could speak better for
- 9 the EDO than I can for NRR. EDO supports this. And the top i I 10 management in the Office of Research. When we're talking l 11 about the new management in NRR, I think Don is accurate in j 12 saying I don't think they know about this yet.
! 13 Tom Murley may not have heard of it.
14 MR. MICHELSON: They have to bless the 15 expenditures of this what you call support?
{
16 MR. SHOTKIN: I think so, yes. I think we want i 17 to get them on board.
l 18 MR. SULLIVAN: Don, in the past, I think that NRR i
19 asked for full power in MIST.
5 20 MR. SOLBERG: Yes.
21 MR. SULLIVAN: So you guys feel comfortable with l 22 going to decay power in this test?
23 MR. SOLBERG: My first bullet up there of things i 24 to be decided yet is the boiling / power tradeoff. No, we're l 25 not totally convinced yet. I think part of the answer is i
ACE. FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-33MM6
i 1 7280101 01 366
( V/bc 1 going to come out of the TRAC analysis as to what the l 2 transient looks like. We have to figure out what are the 3 limitations on our future test capability.
4 MR. SULLIVAN: You see, if you want to run --
5 MR. WARD: Don, is that the right slide?
- 6 MR. SOLBERG
- Sorry about that.
7 MR. MICHELSON: Well, you can't read it anyway.
~
8 (Laughter.)
2 9 MR. SOLBERG: I may not have the right one here.
1 10 MR. MICHELSON: Maybe we won't need it.
l 11 MR. WARD: We've gotten this far without it.
. 12 ( Laughter. )
13 (Slide.)
l 14 MR. SOLBERG: Power level versus system volume 15 tradeoff, yes.
l 16 MR. SULLIVAN: The reason that people compromised 17 on MIST is certainly the cost, but also the types of i
f 18 transients that we're going to look at.
19 What you've done is you've ruled out a whole j 20 bunch of transients because you can't get started if you
! 21 have decay power.
I-i 22 MR. SOLBERG: You've ruled out a part of those
- 23 transients. You've ruled out a front part of those 24 transients.
25 MR. SULLIVAN: In some cases, you've ruled out l
l l
l ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6646 l
. -. - . -- - . - - - - . . - _ - ~ . . - -- ~. _. . ~ . _ . - .
l f
7280 01 01 367
( 1/bc 1 the whole transient because it makes a difference how you 2 get started about where it's going to go in the end.
i
- 3 MR. BECKNER
- Harold, we agree with you 100 4 percent and we are wrestling with that issue. We're in 5 total agreement with you.
{
6 MR. SULLIVAN: You've given up the modular i
- 7 concept. Have you decided on that?
4 8 MR. SOLBERG: No.
- 9 MR. SHOTKIN: I think probably we ought to stop 10 here. I think we're just going over the same questions, i
11 MR. MICHELSON: At 10:30 exactly, we're going to 12 stop. So, whatever you guys have got to say, do it in the 13 next two or three minutes.
14 MR. SOLBERG: One other comment I guess I'd like 15 to make. It seems to me we have with this multi-dimensional i
i 16 type of facility something that's going to be unique in the i
i 17 world. I think it goes a long way toward reestablishing the i
. 18 United States as a place where totally unique data will be 19 available, i
i 20 And I think it will reestablish leadership for 21 us.
22 MR. SULLIVAN: One of the things that I noticed
]
23 is it seemed to be three-dimensional.
24 MR. MICHELSON: We can't get into the details any l
) 25 more, Harold. I think he's got to get his summary, and i /\CE FEDERAL REPORTERS, INC.
j 202-347-37m) Nationwide Coserage SR)-336-(M6 i<
1 1
7280 01 01 368
//bc 1 that's it, because of time. We've got some other things 2 which I'm sure are of interest to everybody that are coming 3 up. And we really need to get on with it because this is 4 still in a fluid state, anyway.
5 6
7 8
9 10 11 12
, D
'd 13 14 15 16 17 18 19 20 21 22 23 i
l 24 25 I
i a
ACE FEDERAI. REPORTERS, INC.
! 202-347-3NN) Nationwide Coserage SI Eb 336-6646
A i
7280 07 07 368.1 MR. SOLBERG: Yes, it is.
()V/bw 1 2 MR. MICHELSON: So we will adjourn for a ten-3 minute break. Let's be back at 10:40 and proceed with the i 4 rest of the schedule.
I
- 5 (Recess.)
J 6 MR. MICHELSON: One of the things I need to poll ;
i
- 7 is when people need to leave here. What times do the 2
l 8 consultants have to leave?
j 9 (A pause.)
! I i 10 (Discussion off the record.)
11 MR. MICHELSON: Let's proceed.
12 I think the next statement is by Ivan Catton.
l
() 13 DR. CATTON: I will go through this rather 14 quickly. I just wanted to make some comments about water i r 15 hammer. l I 16 (Slide.)
17 And sort of try to encurage some itnerest in j
18 water hammer. So I just put together a few notes that are l 19 really based on an EPRI meeting and some of my own views.
20 These numbers a the result of an EPRI study.
l I
j 21 My slides really are not very good.
22 (Laughter.)
l 23 MR. MICHELSON: They are impossible to read.
24 DR. CATTON: That is true.
I
() 25 First, 30 fluid systems and four reactor designs 1
ACE FEoERAL REPORTERS, INC.
202 347 3700 Nationwide coserage 800-3346646
7280 07 07 369
( hV/bw 1 have the potential for water hammer events. Ten of the 30 2 fluid systems perform safety functions. One can make some 3 interesting observations. Many of the subsystems that are 4 designed to protect from water hammer have become water 5 hammer initiators and severe accident management 6 instructions have been postulated by some to lead to 7 circumstances f avoring water hammer.
8 Third, of f-normal plant configurations f requently 9 aggravate their potential water hammer initiators.
10 And finally, I think, which is probably most 11 important, is the removal of restraints, pipe whip 12 protection and seismic snubbers, will place new emphasis on
() 13 normal piping system load limits.
14 I think presently, the snubbers and other things 15 that are put in are designed for a load that is probably as i
16 great or greater than a water hammer load. So they have 17 been able to get away with a lot, and they haven't been too 18 important.
19 (Slide.)
20 During the period '81 to '85 there were 58 events l
21 and just to sort of put them in perspective, I tried to put I
- 22 them under various categories. 60 percent resulted in pipe 23 support damage, hangars, anchors and snubbers. 17 percen t l resulted in component damage, piping, pumps and valves. 10 24
() 25 percent results in reactor trip. That is .02 scrams per l
ACE FEDERAL REPORTERS, INC.
, l 202 347-37(O Nationwide Coserage Mn33MM6
7280 07 07 370
/~s
'(,)W/bw 1 year per plant. And 7 percent resulted in plant shutdowns.
2 MR. WARD: Ivan, these are water events that 3 initiated as water hammer?
4 DR. CATTON: Water hammer, yes.
5 MR. MICHELSON: I assume these are all LERs?
6 DR. CATTON: No, they are LERs and actual plant 7 visits that brought all these out.
8 MR. WARD: Those these are events initiated by 9 something or other that involved water hammer?
10 DR. CATTON: I have some information on what 11 initiated and led to the water hammer.
12 MR. MICHELSON: And you are going to tell us how
() 13 many events probably escaped reporting from water hammer 14 DR. CATTON: Yes.
15 (Slide.)
16 This, to me is an interesting view that came 17 about as a simple hand raising at the EPRI meeting. There 18 were 25 utilities present. It turns out that roughly 30 19 percent actually address water hammer, 4 percent have an 20 engineer assigned, 30 percent incorporate water hammer in to 21 operator training. 20 percent have check valve programs in 22 place. This is after the San Onofre incident and 20 percent 23 look on check valves as active components. And this is just 24 my own little thing on the bottom. The NRC pronouncement
() 25 that water hammer is not a safety issue, I think, has led to ACE FEDERAL REPORTERS, INC.
- o:-m-roo Nationwide coserage sm3m 6M6
4 i 7280 07 07 371
( )V/bw 1 utility lag of interest.
I 2 (Slide.)
j 3 I think this is what you were referring to, Carl.
^
4 The information that I have here came from the LER and INPO i
. 5 data, the back room wisdom, and this was among the utility 6 people who were at this meeting, is that only one in ten I 7 water hammer events is reported. So if there are 58 during I
8 that period I guess that means there are 580. The thing 9 that is important is that if water hammer yields its maximum 10 pressure pulse, failure will certainly occur. And the i 11 statistics of water hammer are not well defined.
12 MR. WARD: Failure of?
() 13 DR. CATTON: The pipe system.
14 MR. WARD: Loss of fluid or support?
15 DR. CATTON: Well, the maximum pressure is 16 extremely high, and one in 400 to 500 water hammer events
, 17 reaches its maximum or near its maximum. Most of the time 18 it is significantly less, and if you look at the 58 events,
, 19 you have a number of them with damage. So the statistics 20 sort of hang together. When we have 500 to look at, we l
21 certainly will have one with some significant damage. San 22 Onofre is one of those.
l i 23 MR. MICHELSON: It is important to dif ferentiate i
(
24 between significant pipe support damage versus significant l () 25 pressure boundary damage.
I ACE FEDERAL. REP'ORTERS, INC.
202-347-3700 Nationwide Coserage 831336-6M6
]
- -. --. _ - . - , - -- - - .. -- -- . , . - - - . , . ~ .-
7280 07 07 372 3
\V/bw 1 DR. CATTON: I think it certainly is, but I am 2 not sure we can do that.
3 MR. MICHELSON: I wondered in your little survey 4 whether you differentiated it.
5 DR. CATTON: No, I didn't try. But things are 6 going to change in that regard too. They are going to be 7 taking out numbers. They are essentially going to be 8 reducing the support for the piping systems. So the next 9 . water hammer, when the piping system has been changed, is 10 going to be dif ferent. It may well lose its pressure 11 boundary. I am not sure we know. We certainly will have 12 more pipeline.
() 13 (Slide.)
14 One of the difficulties, I think, is NUREG 582.
15 These were typed up at the last minute, and they 16 really reflect it.
17 MR. MICHELSON: Your word processor doesn't 18 expand the size of the letters or anything?
- l 4 19 ( Laug hter. )
20 DR. CATTON: I don't know how to work it.
21 MR. WARD: You got a big chip though; right?
l 22 (Laughter.)
23 DR. CATTON: At present, the requirement is that
! 24 the water hammer be catastrophic for it to be considered. I
() 25 think that is foolishness. The data base, as a result, is ACE-FEDERAL REPORTERS, INC.
202J47-3700 Nationwide Cmerage 800-336 4 646
~
7280 07 07 373
( ?N/bw 1 biased towards events that shouldn't happen. A lot of the 2 precursors, and so forth, just don't get reported. There is 3 no agreement among system designers as to what should be 4 considered normal or abnormal, so water hammer is usually 5 not considered in the piping syster design, other than that 6 steam should flow uphill and water downhill, and a few 7 simple things. l 8 Plant engineers wind up fixing many problems 9 during the pre-op phase and only those missed show up in the 10 data base. So the plant, as designed at the front end has a 11 lot of potential for water hammer, and they gradually get 12 eliminated. Some slip by. The ones that slip by turn up
() 13 later.
I 14 MR. MICHELSON: There is an important point here 15 that maybe you should be aware of, if you are not. And that 16 is, in all of this approach of leak before break, so you can
- 17 take out supports, and so forth, it is said that if you have 18 a potential for water hammer than leak before break doesn' t 19 apply. So it becomes very important now to decide which 20 systems have a potential for water hammer. So particularly 21 outside of containment, you can decide whether you have to 22 take away the extra supports and whatever. I mean, you 23 should be allowed to take them away.
24 This is more than just academic, to wonder 4
() 25 whether you have a potential for water hammer and how do you ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 80)-336W46
.. - _ _ = - -. . . _ . .
7280 07 07 374
( hv/bw 1 know you have a potential and which systems do you think 2 exist and which parts of systems? We don't have the 3 understanding to narrow it down that nicely, yet we are 4 allowing other people to do it for us. And the question of 5 water hammer is one of the significant ones, along with 6 stress corrosion cracking.
7 DR. CATTON: It is in teresting too. There was a 8 fellow from Bechtel, who was doing some calculations for 9 water hammer. When he was asked, what did you use, he said 10 RELAP-5. We all know condensation is a dif ficulty in these 11 codes. So he asked him, well, why did you use RELAP-5? He j 12 says, well, NRC developed it, didn't they?
() 13 I just throw that out for your information.
14 (Laughter.)
15 Somehow whatever you do with these codes, when it 16 slips over to industry, it is viewed as having a red ribbon 17 on it.
18 MR. EBERSOLE: When you say " water hammer," do 19 you mean just the particular kind of water hammer that you 20 get?
f 21 DR. CATTON: Yes.
i 22 MR. EBERSOLE: You are not talking about impact
! 23 loads due to interesting and infrequent sequences like pump 24 drain down?
() 25 DR. CATTON: Those kind, I think you can analyze.
/\CE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage wh336-6M6 i_ .__- . . _ _ _ _ . . _ __ _
7280 07 07 375
()N/bw 1 MR. EBERSOLE: Even so, even at this late data, 2 af ter standing around for five or ten years, plants have 3 gotten into interesting sequences they have never gotten in 4 before with prodigious results in water hammer. You know, 5 like this venting business because pumps stop because of 6 outside power failures, and it drains down, and then the 7 diesels pick it up and it rams the thing into the pipes.
8 MR. MICHELSON: Not unpredicted. Simply took a 9 while to get the experience.
10 DR. CATTON: If it is single-phased water hammer, 11 I think that is predictable. Every aerospace company in the 3 12 country has codes to do it.
() 13 MR. EBERSOLE: Nevertheless, that area of water j 14 hammer has been obscured by the infrequent nature of these i 15 transients. So they are still occurring.
16 DR. CATTON: They should be looked at then. They 17 are probably more important than some of the LOCA stuff that 18 we do.
19 Let me just try to finish this up. I think a 20 question for licensing is, what is preferable? Do you fix 21 it af ter it occurs or do you prevent it? Maybe it is 22 possible that a design basis water hammer is what is needed 23 for piping system designers. Something, in my view, 24 certainly needs to be done. And the head in the sand view
() 25 with respect to water hammer is, I think, going to lead to i
~
l' ACEJFEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cmerage 800-336-6M6
7280 07 07 376
.( %V/bw 1 difficulties .
2 MR. MICHELSON: Let me give you something to 3 think about. Then you can tell me if you have looked at it.
4 In the case of ECCS systems like core spray, for 5 instance, the spargers normally sit at 550 degrees or so and 6 during a large break, the system depressurizes, the spargers 7 flash hot water to steam, and so forth, and this occurs back 8 down as far as the pipe is hot. No then, all of a sudden, 9 we put in 70 degree core spray water or very low temperature 10 core spray water. Are we setting ourselves up to water 11 hammer the sparges, and so forth?
12 DR. CATTON: I think you have already answered
() 13 your own question. We have a history of these kind of 14 problems.
15 MR. MICHELSON: Very closely related, in case of 16 the J-2.
17 DR. CATTON: It is the same problem.
18 MR. MICHELSON: We've kind of put our head in the 19 sand and don't even consider the possibility that we won't 20 have core spray spargers when we throw water to them, i
21 DR. CATTON: I believe there is an I&B notice 22 that just went out on RHR restart in water hammer.
23 MR. MICHELSON: That is a cold water situation.
- 24 That is the one I referred to that was understood in 1973.
() 25 (Slide.)
ace FEDERAL REPORTERS, INC.
i 202-347-1700 Nationwide Coserage 80k336-6M6
- f 7280 07.07 377 i O u/hw 1 DR. CARTON: This ene ceme rioht erf mv orinter2 1
- 2 (Laughter.)
4
., 3 MR. MICHELSON: What kind of printer do you have?
4 DR. CATTON: It prints very fast, and I didn't 5 think to press the other button to get the nice print, and 6 the laser printer's at school. This one is at home.
7 I tried to put together some of the results of 1
8 the EPRI studies. You can get an idea of where these things 9 are happening. The causes are administrative, components, 10 design, operator, procedures and unknown. And the various i
- 11 systems they occur in, RHR, feedwater, turbine, steam 12 supply, ECCS. A significant fraction, almost 10 percent, is l
() 13 in the ECCS system. Main steam supply. I don't remember 14 what SGBD is.
15 Anyway, a lot of these fluid systems are 16 important to safety, and they have water hammers.
i l 17 MR. BOEHNERT: SGBD is probably steam generator
} 18 blowdown.
19 DR. CATTON: Okay. Anyway, Paul, I will give you 20 a set of these to do with as you wish.
1 l 21 DR. SULLIVAN: This doesn't occur very often.
)
i 22 ( Laughter. )
i i 23 So I am going to take the opportunity --
)
24 DR. CATTON: Take whatever opportunity you can.
() 25 DR. SULLIVAN: You said you went to a meeting i
i
/\CE FEDERAL REPORTERS, INC.
1 202 347 3700 Nationwide Coverage 800-336-(M6
7280 07 07 378
( XV/bw 1 with the indusrty, right? Do they have a plan to rectify 2 this?
3 DR. CATTON: Yes. As a matter of fact, EPRI has 4 issued an RFP and the program is quite extensive. They are 5 talking in excess of $1 million, at least the manpower 6 loading looks like that. ' And their intention is to 7 eventually come up with a handbook. I don' t know what the 8 time scale is on it. And I don't know how much response 9 they are going to get from the industry. It is my view, 10 that as long as NRC doesn't take an interest, then the 11 utilities aren't going to spend any money on it, because 12 they can't pass it through to the rate base.
() 13 And I think an excellent example of this is what 14 happened at San Onofre The people at San onofre decided 15 that what they really needed to have was a liquid sensor in l 16 the line to the steam generator. They went to a person who l 17 developed that for them. It caused 100K for the in s trumen t.
18 They even bought spares. They came down to the point where
(
- 19 they were going to do something, and NRC says, well, gee, 20 well, we don't consider water hammer as an issue.
21 Well, all of a sudden the utility can't spend any 22 money. They just write the $100,000 off. And maybe it is 23 not the proper view for the utilities, but they do not do 24 things that they can't pass through to the rate base. So
() 25 these things sit on a shelf. They were never installed.
ACE-FEDERAL REPORTERS, INC.
.02-347-3700 Nationwi& Coserage 8t o-3346M6 l
7280 07 07 379 The engineers want to install them.
(XV/bw 1 2 I think that is a sad state of affairs.
3 MR. MICHELSON: There is an uncertainty in ECCS 4 response, because it is during that transient situation that 5 you are setting yourself up potentially for some of this 6 water hammer. If you look at the fine structure in some 7 parts, you spend millions of bucks on it and ignore other 8 areas, because somebody said it is a resolved issue, and 9 therefore, there is no more problem.
10 DR. CATTON: Three Mile Island, I don't think you 11 want to forget that what caused the difficult there, they 12 were pumping rather nasty water through the steam
() 13 generators, and they had to blow the steam overboard, 14 because they have a water hammer that disabled pumps. That 15 water hammer put things a lot close to trouble.
16 DR. SULLIVAN: Is somebody systematically going 17 through a plant and looking for places that it could occur?
18 DR. CATTON: I don't think so. And I would hope 19 that maybe the EPRI work would lead to that, but it is not 20 clear to me that will occur.
21 DR. SULLIVAN: How about codes, the ability to 22 calculate? Is anybody looking at that?
23 DR. CATTON: The codes, I think, can calculate 24 the maximum, because you just assume that the steam is a
() 25 vacuum and that the water is going to come crashing in and ACE FEDERAL REPORTERS, INC.
202347-3700 Nationwide Cmerage N o.33MM6
7280 07 07 380
( hV/bw 1 calculate the spike, but that is too drastic, I think. You 2 get two large a pressure. And I think you will wind up 3 painting a picture that is much too bleak. It is not like 4 that. It is really statistical. If you are going to get 5 one out of 400 that are of some magnitude, then you can 6 decide whether or not you want to do something, based on 7 risk or something else. Right now, I don' t think you can do 8 anything.
9 DR. SULLIVAN: But nobody is developing a code.
10 They are not doing any significant code development.
11 DR. CATTON: Not that I know of. Now Peter 12 Griffith took a look at what you need to do, if you have an
() 13 empty line and you want to fill it without a water hammer, 14 and he has some limiting curves, flow rate, and so forth. I 15 am not sure where he is carrying this, but other than that, 16 I don't know of a systematic load.
17 DR. SCHROCK: Ivan, I just wanted to make one 18 comment about your characterization of the utility 19 attitudes. Eventually, they are not all thinking in that 20 way, because Peter's work on water hammer that led to his 21 paper that won the award was funded by a utility.
22 DR. CATTON: That is true, but it was to resolve 23 an existing problem in the plant. That is dif ferent. They 24 had the problem. They had to do something. They wanted to i
() 25 know how rapidly they could fill the system.
ACE-FEDERAL REronTuns. INC.
202 347-37tW) Nationwide Cmerage S(n336W46
7280 07 07 381
( )V/bw 1 DR. SCHROCK: That is dif ferent, in your view, 2 from San Onofre?
3 DR. CATTON: I think so. If you look at the 4 numbers, these percentages of the people who had an 5 interest, only 30 percent of the 45, it was 7 or 8, actually 6 had a program. Well, people looked at the water hammer 7 question. Only 4 percent, just a small number had an 8 engineer who was told, you have a responsibility for water 9 hammer.
10 Now maybe the right 25 weren't there, but I would 11 suspect that the utilities have sent their people there, the 12 ones who are more in te re s ted .
() 13 MR. BOEHNERT: Harold, I think it is worth noting 14 that we are looking at a discussion of water hammer on the 15 July 16th subcommittee. I have looked to EPRI and John Kim 16 who is the project manager. They have indicated they are 17 willing to conme in and talk to us.
18 MR. MICHELSON: Thank you, Ivan.
19 If there is no other discussion, let's proceed to 20 the next agenda item.
21 (Slide.)
22 MR. LILES: I am going to talk a little bit about 23 some versions of TRAC. The current released version is PF-24 1, Mod 1. We currently have under development a version,
() 25 Mod 2, which is starting up developmental assessment. And I ace FEDERAL. REPORTERS, INC.
g 202 347-3700 Nationwide Coserage 80).33MM6
._7280 07 07 382 will chat just a little bit about that.
hV/bw 1 2 (Slide.)
3 When we looked at producing a new version of the 4 manual, we wen t down , essentially, the list of people that 5 we had planned to send it to, and we got to about 350.
6 There are a number of people out ther using the code, and as 7 a consequence, maintenance is a large effort for the code.
8 The Mod 1 code is, of course, a number of programs, both in-9 house and outside at ICAP. 2d/3D, MIST, ROSA-IV.
10 Presumably, it will be supporting Appendix A and a number of j 11 other things, i
12 We have got a frozen version which, as you will 13 see in a minute, really isn't completely frozen, but we do 14 try to severely restrict the sorts of changes we put in the 15 code. Error corrections, of course, are allowable. It is 16 find to add user convenience features, and it is okay to add l 17 new models, as long as they' re optior.s. And I will describe i
!, 18 briefly one or two.
19 There is another thing that we are trying to do 20 to the frozen version, and that is, continue to improve the 21 portability of the code, to be able to run it on a number of j 22 different computersw. The original versions of TRAC were
! 23 designed for a 7600. We have retained some of that large 24 core, small core stuff, simply because there are still a few O 2s veoo oeere eet there.
l i
l Acu FEDERAL REvonTEns, INC.
202-3 C-3 W Nationwide Cmerage 8m-33MM6 l
1
7280 07 07 383
( >V/bw 1 Now we, of course, at Los Alamos, run strictly on 2 one of the eight Los Alamos Crays that are available, but I
3 there are users that are running on 205 Amnols, and what i
4 have you. We don't maintain specifically IBM versions, for 5 example, but we have tried, just as part of any sort of 6 error corrections or updates, to gradually try to make the j 7 whole code more transportable, make it more easily 8 transported to computers.
f 9 (Slide.)
, 10 We estimate that we had in the past 17 months 11 about 1100 separate user contacts. And what we define as 12 user contact is through our User Liaison office. I am not I
() 13 talking about other phone calls that may come in, but it is 14 specific user contacts asking for in formation , complaining i 15 about something, asking for help.
i 16 The 1100 is based on sampling. We' don't log i
17 overy single phone call we get in our User Liaison Section.
t l 18 We did, for a couple week period, and extrapolated to get '
19 the 1100.
20 DR. CATTON: Who supports the User Liaison 21 Office?
22 MR. LI LES : Essentially, it is ICAP. It is part l 23 of ICAP. It is combined not only with ICAP, but part of the 24 TRAC code developmen t. It is sort of a mixed thing, because
() 25 a lot of the users are ICAP.
ace-FEDERAL REPORTERS, INC.
202 347 3700 Nationwide Coserage NV)-336-M>46 l
i 7280 07 07 384 (hv/bw 1 DR. CATTON: What if someone off the street calls 2 you, do you give them this free service?
3 MR. LILES: We have had that happen, but they
, 4 usually seem to have been involved with too much Ripple at 1
7 5 the time, so we normally don' t.
6 (Laughter.)
1 1 7 DR. CATTON: Is that a requirement for losing j 8 TRAC?
9 (Laughter.)
} 10 MR. LILES: It is a definite help. It depends on 11 the nature of the question. If it'is somebody that is not l 12 in some way contributing -- let me point out something else.
() 13 We have got a Users Group now, a domestic Users Group, that l 14 we charge a modest fee. So if it is a member of that 15 domestic Users Group, then, again, we essentially assume l
i 16 that we will answer their questions. If it looks like it is i 17 going to involve a lot of work, they are going to have to
- 18 pay something extra.
19 But it depends on the nature of the question. If j 20 it is something we can resolve very simply, ten to fifteen
} 21 minutes, we will go ahead and do it. If it looks like it is 1
22 anything more involved than that, no.
i 23 So anyway, based on our estimates, we think there 24 is about 3 percent of user problems that are unresolved and 1
() 25 some of those are significant work to fix. So I don' t want
! /\CE-FEDERAL REPORTERS, lNC.
202-347-37th) Nationwide Coser. ige Mn336-6M6
7280 07 07 385 1 to underestimate or assume that somehow the 3 percent, you
(()V/bw 2 cna linearally weight the magnitude of the problem by 3 percentages.
4 Nonetheless, I think most of them were able to be 5 resolved. We catalogued what we think are significant 6 things in to trouble forms, and we've got 270 of those 7 currently. That doesn't mean they are unresolved trouble 8 forms. It just means that we think are significant.
9 (Slide.)
10 Over about the last 17 months, I think it has 11 been a bit less than that, we have put in 170 --
12 approximately 170 separate idents or chucks of update coding
() 13 into the code. The large number of lines of coding. It 14 looks like a very large number. A lot of those are 15 comments, but there are a couple of new models we put in --
i 16 a CCFL model, a separator model, which I will talk about in 17 a minute. A lot of the lines of coding are asosciated with l 18 those new models.
I
- 19 So the number is larger, much larger than it i
20 would be, if we did not include a couple of new models in 21 the code.
! 22 These percentages are appropriate percentages.
I
< 23 And it is approximate, not by lines of FORTRAN, but it is 24 approximate by the number of idents.
() 25 DR. CATTON: What is the total number of lines?
l ACE. FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Cmerage Muh336 6M6
1 2
7280 07 07 386 i
( )N/bw 1 MR. LILES: In the code? The basic TRAC code is i 2 probably on the order of 70,000, 75,000 lines. There are 3 also, then, about an equal number of lines of graphics codes i
4 and some utilities, and some of these error corrections went
[ 5 into these things. We lump it as one thing.
6 DR. CATTON: So you have replaced a third of the 7 code.
8 MR. LILES: Not quite, but a lot of its comments 1
i 9 too.
4 10 But, yes, it is more than 10 percent.
) 11 (Slide.)
l 12 So some of the recent improvements that we put
() 13 into the code are a self-initialization procedure. We've 14 got now, multiple source terms in the 3-D vessel, and what I
]
j 15 mean by that is that you can hook up more than one pipe into 16 a single 3-D mesh cell, which is something users have f
17 requested for sometime, and it is conceptually easy to do.
1
! 18 It turns out that actually implementing the logic takes a l 19 bit of work.
20 The CCFL model and separator model. In paralle,,
21 we have been working on a Mod 2 version. The primary i
! 22 distinguishing feature of the Mod 2 is a 3-D-2 step, which 23 will cause reduced running time on a lot of transients, l
j 24 particularly slower transients.
() 25 DR. CATTON: Is Mod 2 under development now?
i ACE FEDERAL REPORTERS. INC.
. 202-347 37(o Nationwide Coserage 8m-33MM6
7280 07 07 387
( )N/bw 1 MR. LILES: Yes.
2 DR. CATTON: When are you going to finish Mod I?
3 MR. LILES: It depends on what you mean by 4 " finish?"
5 DR. CATTON: By " finish," it is the normal finish 6 of a code development program, which means documentation, 7 characteristics, it's limitations. That ends the program.
8 MR. LILES: We think that the code itself, if yu 9 consider that it is a frozen code, is essentially finished, 10 except for some error corrections. The documentation is 11 finished. The manual is finished. The assessment documents 12 are finished. The OA documents are finished.
() 13 DR. CATTON: When will those be out?
14 MR. LILES: We had a set of meetings with the 15 NRC on this very matter the previous two days. I think they 16 are going to decide whether to temporarily suspend work on 17 Mod 2 and finish the documentation or summary show. In 18 other words what I suspect will happen, and this is just an 19 unsolicitated speculation, I suspect that Mod 2 assessment 20 offectively will be slowed up somewhat to allow some of 21 these other activities to finish.
22 DR. CATTON: Because Mod 2 is not going to be 23 done anyway, in time to enter in to this. Certain things are 24 needed for Mod 1. If they are not supplied, then, in a way,
() 25 you are not doing the job.
Acti FitorinAi. Rtiponittas, INC.
L 202 347-3700 Nationwide Coserage NW)-336 M46
. . =. - - - - - . . . . . . -. .. _ . -
i 7280 07 07 388
( )V/bw 1 MR. LILES : You understand that we have 189.
i 2 DR. CATTON: I don't know what those are. Well,
! 3 I sort of do.
4 MR. LI LES : Well, the 189, it is our contract for
. S the NRC to do this stuff. That is about a year in advance.
6 Effectively, it requires a modification, and I think that is 4
7 going to happen. A major model addition, major model 8 improvemen t in Mod 2 is an improved core void model, and I i
9 will talk a little bit about that.
10 DR. TIEN: Will there be a Mod 2 numerics? Are 11 you considering some parallel processing?
1 j 12 MR. LILES: Not directly. We have some
() 13 vectorizatio, because, of course, the Cray is a vector 14 machine, but that is in 3-D. We have a proposal that is
! 15 temporarily shelved to permit vectorization in the 1-D, and 1
16 also at the same time to make it amenable to parallel i
17 processors, but that is sort of one the shelf.right at the l .
18 moment. We are not doing any work on that. We know what we
! 19 would do, i f the time, money et cetera, were there.
] 20 We have done some benchmarking on parallel r
i 21 machines, just a very low level effort in the past.
i
! 22 (Slide.)
j 23 I will tell you what I mean by self-l 24 initialization. It is not a separate algorithm. It is not I
() 25 a separate steady stage algorithm. What it is, though is a i
- ace. FEDERAL REPORTERS, INC. '
202 347 3?(o Nationwnie coverage soo-33 M M 6 o
- - - - .w,.mm-.-w-wm+ ,,,
w,~4-,,v.yr. ,m.,-3 w. -mnmm...r,.-- , . , , - - -.
7280 07 07 389
( lN/bw 1 standard set of controllers. It is a bunch of controllers 2 that make it much easier, essentially, to run the code most 3 types of experimental facilities, and certainly on PWR type 4 plants, to be able to get a steady state, it makes the code 5 achieve a steady state somewhat faster. It makes it easier 6 to hit certain set point parameters. So it sort of macros, 7 if you want to look at it. It is sort of macro controllers.
8 The multiple source terms I described earlier 9 simply allow several pipes to be connected to a given 3-D 10 mesh cell, not difficult in principle, but it did take a 11 little bit of work to actually get it in.
12 (Slide.)
() 13 We have gone ahead and implemented the CCFL 14 model. This has been under discussion for a while. This is 15 in Mod 2 also. Mod 2 has this, but of course, this is also 16 in Mod 1 now. The version that we actually put it in was 17 what we call our internal version 12.8.
- 18 I will show some assessment against some 19 saturated and some subcooled flooding data. It was reviewed 20 in Germany. They have run it on a couple of bundles 21 effectively, and it did very well, although apparently, it 22 is not working as well in UPTP, but that may not be due to 23 the CCFL model per se. That may be due to some other 24 things.
() 25 (Slide.)
ACE. FEDERAL. REPORTERS, INC.
20L 3474700 Nationwide Coserage N o-3 3MM6
7280 07 07 390 1 You have a choice. You can use J-star or K-star
( )W/bw 2 scaling ef fectively. You can use a mixture. It is not a 3 particularly complex model. The idea, of course, is to, 4 under certain conditions, replace, basically, the liquid 5 momentum equation, in order to be able to specify the amount 6 of liquid downflow.
7 DR. TIEN: What is the factor B? That is not a 8 bounding number 9 MR. LILES: No, it is an interpolating parameter, 10 which allows you to mix J-star and K-star scaling. If it 11 is zero, it is J-star, if it is one, it is K-star. And for 12 small scale facilities, if you think you could get a better O
(,/ 13 fit by using something in termedia te , you could do that.
14 DR. CATTON: You meant J-star plus K-star right 15 underneath there, not "or"?
16 MR. LILES: Right.
17 DR. TIEN: Actually, it is related to the 18 bounding. The M and the C both are functions of the 19 bounding.
20 MR. LILES: Yes, but that is not what I am 21 talking about.
22 (Slide.)
f 23 I am going to show you one or two of Bankof f's 24 experiments.
() 25 (Slide.)
Act -Fl!Dl!RAl. RitronTrias, INc.
! 202 m 37m Nationwide cmerage sm-3M 6646
d i
7280 07 07 391
()N/bw 1 I guess I don't have the saturated one in my set 2 of slides. I thought I brought it. I will show you the l- 3 subcooled one. The saturated one looked excellent. This
- 4 one looks very good also. What we used were essentially the 1
1 5 correlation constants were the same for the saturated and.
6 the subcooled. So the condensation model is accounting for I;
1 7 the difference. The basic condensation model in the code.
. 8 (Slide.).
! 9 DR. TIENs For the condensation case, I remember 1 10 Yuen and Bankoff's things. They observed the pressure I
11 bounding hysteresis effect.
12 MR. LILES: As I recall, that is right.
() 13 DR. TIEN: There is a recent paper by Taitel, et 14 al., that shows that you get hysteresis.
15 MR. LILES: There is no formal provision for 16 hysteresis directly in the model. It is due to coupling i
i 17 with the other stuff.
18 DR. TIEN: But that would have some effect, i 19 because whether you go to flooding or deflooding, you have a i 20 dif ferent formula.
l 21 MR. LILES: That is right. Well, obviously, the l
- 22 notion of hysterosis is, one curve won't do the whole thing.
. 23 Anyway, we didn' t get the weep point. The weep 24 poin t is, of course, the point at which you can barely l
() 25 observe a little bit of liquid before the I plate. We were i
1 ACE. FEDERAL REPORTERS, INC.
i 202 4 7 37m Nanonwide coserage sto-13u646 s~r
7280 07 07 392
( )V/bw 1 able to do what looks like a good job of that, and then we 2 finally got the actual- dumping point, at which there was 3 significant dumping down. It also seems to work very well.
4 DR. CATTON: What value of B?
5 MR. LILES: To tell you the truth, I don't know.
6 It was a fixed value. We weren't varying from one data 7 point to the next.
8 DR. CATTON: But do you vary it from one 9 experiment to the next?
10 MR. LILES: Yes, you might. It depends on the 11 geometry. The whole idea is -- the whole idea of putting a 12 CCFL model in, or at least a large portion of it is, that
() 13 there may be geometry dependence for CCF that isn't in the F
14 code otherwise.
15 MR. WARD: So this fact or B would be in the code 16 as a function of something?
17 MR. LILES: It is a user iraput. You have to know 18 something about that. Now you can switch it off and the 19 code will do something for CCFL. So it is not a case where 20 you get nothing, but'what you tend to got, and it isn't 21 shown on this particular slide -- let me see if my next one
, 22 has it.
i l 23 DR. CATTON: Could I use one value of B for a tie i
24 plate and another value of B for something else?
i
() 25 I I
! Acn-Funtinat Riti>onTnns, INC.
202 347 37(u) Nationwide Cmcrage N Xb)36 6M6
)
. . . _ . - ~ . - _ . _ _ __ _ . _ _ . _ _ _ _ _ . _ - - -
r 7280 07 07 393
()N/bw 1 MR. LILES: Yes.
2 DR. CATTON: So I could put in a whole array of 3 them at the outset.
4 MR. LILES: You could. I will show you what the 5 code does by itself for CCFL.
6 (Slide.)
7 MR. TIEN: That becomes just like a fudging 8 constant from zero to one. That's a big factor. I think it 9 should have some rationale behind it, perhaps with certain 10 allowed spread, in terms of the value. Otherwise it becomes 11 totally empirical.
12 MR. LILES: What we anticipato doing is putting
() 13 some suggestions in the Users Guide, okay? That would be 14 what we would suggest. Understand, what we have assessed 15 the CCFL model with so far, two sets of data, Bankoff's 16 data, a couple of dif ferent of Bankof f's runs and the 17 German's ran this one for some Karlstein data.
18 So as we get more experience or see what other 19 people with similar models are using, we develop user 20 quidelines.
21 DR. SCilROCK : I would like to echo this kind of 22 protest here. B, in fact, is a function of the bond number.
23 Why won't you tell us how you do that? That is documented 24 in the literature. B has to be a fuention of the bond
() 25 number.
ace-Fitorinat. RiivonTrins, INC.
202-347 3700 Nationwide Cmerage N n 13MM6
7280 07 07 394
( )N/bw 1 MR. LILES: What we do, in terms of the model, is 2 simply predict a vapor velocity, based on this model, as 3 described, and what I mean by "as described" is, you are 4 really calculating JG and JL, based on this particular kind 5 of fit, okay, for a particular geometry. If it turns out 6 that the JG is large enough to preclude or predict less 7 liquid downflow than the equation is calculating, then you 8 simply limit it to this value. It is a simple model.
9 DR. SCHROCK: I don't think you have answered the 10 question, though. It is known phenomenologically that, in 11 fact, the change of the flooding relationship from the 12 Wallis correlation, in terms of Wallis paramotors, to the
() 13 correlation, in terms of the Kutta-Joukowski parameters, is 14 dependent upon the bounding number.
15 Will that guido you in how you calculate it?
16 MR. LILES : We don't calculate it in the code.
17 That is what I am trying to tell you. It is an input 18 parameter. The user simply inputs it.
19 DR. TIEN: That is precisely the point. There 20 are so many data floating around for CCF, you have a pretty 21 good idea how you bound that value of B.
22 MR. LILES: That is right. We have a very good 23 idea, if we are looking at a Creari downcomer.
24 DR. TIEN: So you have to put some rationale in to
() 25 it and not just say that varies from zero to one. That Acti 17ti n ti n a i. R i:PonTtins. INC.
202 14 4 00 NanonmJe Cmerage m .1% M46 n
7280 07 07 395 1 could be some order of magnitude, just from zero to one. So
()1V/bw 2 you have got to have a lot of data, you know, to put a 3 number there, but with some guidance about under what kind 4 of geometric configurations will you got what kind of 5 spread.
6 MR. LILES: That is what we plan to try to do at 7 some poin t, but in the user guido, what we are not going to 8 try to do is to, from the input, the geometric input, that 9 the code currently requires, we are not going to try to, 10 interior to the code, calculate that number.
11 DR. SCilROCK: flow can you have quality assurance 12 of the result of the code calculations, if you leave it up t
() 13 the user to haphazardly pick that number and put it in 14 there.
15 MR. LILES: The point is, when you look at a 16 calculation, you have to try to look at tho way the user 17 modeled the system. This has always been true, and in any 18 of those codos, you can't preclude the user doing something 19 that turns out to be a very poor choice. You can' t do that. i 20 DR. SCilROCK : Somohow you have to, if you are 21 going to have confidence in the result.
22 MR. LILES: The confidence in the result has to 23 como f ro.n an examination of the input dock for the 24 particular system the user is attempting to model.
1
() 25 DR. TIEN: I would fool much more comfortable, if Acti Fimi Rai. Riii>oluitas, INC.
. i 202 .147 4 700 Naimnwnle cmcrage M u L .i to (M6
f l
7280 07 07 396
( )V/bw 1 you have some kind of formula, which, again, I mentioned, so 2 much work has boon done in CCPL to say how you calculate 3 that factor B and with an allowed kind of deviation and 4 given, say, some users guido or whatever, say, under certain 5 kinds of geometrical configuration, you would expect certain 6 variations. So not to completely leave it to the user to 7 vary it from zero to one, then you don't have any assurance 8 really.
9 MR. LILES: Well, when we try to model, for 10 examplo, an orifico, the user can put generally any K factor 11 he wants to in the code. If he selects a K factor that is 12 way out of lino, I think it is pretty obvious that tho
() 13 results may not be very meaningful, and if you look at any 14 of the codos, this is not an isolated examplo.
15 DR. TIEN: That is exactly -- I disagree with 16 that point. I think we have so many floating constants with 17 out any real rationale about how to pick thoso constants, 18 and that gives a lot of people -- somo people love it, 19 because they can vary whatever they want to chock with the 20 experimontal observations, but that is just not the way to 21 build confidence or credibility of a very powerful codo like 22 T RAC .
23 MR. LILES: Well, it has boon our experience that 24 when wo try, the usor doonn't really have much option, in
() 25 terms of models, okay. In other words, the modol soloction Acit.17 ii>iti<Ai Iliti>oitriti<s, INC, 202. m m oo Nanonwide nnerage mow vaa
7280 07 07 397
( >V/bw 1 in the code. This is one example where they do.
2 DR. TIEN: Models very closely relate to those 3 , constants you are talking about like B, and so on.
4 MR. LILES: Yes, I understand that.
5 MR. MICHELSON: Can we proceed? 12:15 is the 6 drop dead time.
7 (Slide.)
8 MR. LILES: Basically, the separator model was 9 separated out of the TRAC B code, the BWR code. It did turn 10 out that we had to, in order to got it in to TRAC P, thoro 11 were enough dif ferences that a lot of FORTRAN had to be 12 rewritten. Essentially, the thing is, the model that is in
() 13 T RAC B .
14 (Slido.)
15 I don't intend to spend very much time on this, 16 except to point out that just as in TRAC B, we have several 17 options. An ideal separator is still in thore, if you caro 18 to use that. Thoro is a partially mechanistic separator 19 modol. What I really mean by that is that it is based on 20 the GE data for their stoam-water separators. It turns out 21 that the UPTP separator was built by GB. Wo think this 22 should work well for UPTP, and then you can supply again 23 your own performanco data.
24 (Slido.)
() 25 The coro void modol. Wo are starting assessment Acit Fiti)iinat. Riti>onit:as, INC.
202mmm Nanonmae rmaage smav, u,u, c
I 7280 07 07 398
( hV/bw I with a core void modol. It is entirely possible, as we go f
2 from assessment that we may make some changes for it. . Right 3 now on the very, very limited assessment we have had, it 4 certainly seems to represent a noticeable improvement over 5 the old codo version with at least a few caveats, which I 6 will insert in a minute.
7 The thing we wanted to do was to try to eliminate 8 the interf ace sharponer, which was a device to try to 9 provant essentially excessive liquid carryover, which we had 10 obsorved years ago in some of the earlier versions of TRAC, 11 and it probably works too well. The current interf ace 12 sharponer does not probably provide enough carryover,
() 13 although it was originally tune up with the low flooding 14 rato floc tats which woro very high power hot and did not 15 show nearly as much water up in the upper reachos as CCTF or 16 SCTP.
17 Anyway, the basic idea was to try to improve the 18 prediction of void fractions. And wo decided that despite 19 questions about cold wall offects, wo decided to heavily 20 depend o'n the largest scalo data we had, CCTF and SCTP, to 21 try to use this to improve the void predictions.
22 (slido.)
23 Wo tried soveral dif ferent models. Wo had a low ,
24 levol ef fort going on for a number of months on this. Ken
() 25 Williams and I had a modol which used a modified form of Aci I;itolinai. Ill:PonTiins, INC.
- 02a47 37m Nationwide rmerage moJwua6 ,
7280 07 07 399
( )V/bw 1 Blasius below the quench front, because it appeared to us, 2 if you looked at single bubble drag, that you get too much 3 drag. And we needed something that was more like sort of a 4 bubble sheet, in order to be able to do this.
5 We didn' t actually code up the Anderson /Ishii 6 model, but what we did do is write a little separate PC 7 driver code and drive it, because once we had done some hand 8 calculations, we had at least some idea of the drag numbers 9 that we needed at different stages of reflood, as we went up 10 the core, in order to match the data.
11 We looked at the Murao/Iguchi model. We had some 12 difficulties implementing that, because there were some
() 13 discrepancies tha we had in the published reports. We 14 finally got what wo think is a model of that in the code.
15 Then we used somathing that EPRI, Lalouche, basically, had 16 come up with. Full-range reflux correlation, which, as_the 17 name suggests, is a drift flux thing that he matched to 18 quite a variety of raw data. I will show you what that 19 looks like.
20 As of right now, we are using the EPRI model. it 21 is not perfect, by any means, but it seems to have provided 22 the best void profile results to date. We do, of course, 23 use it to extract an interf acial drag, based on the usual 24 drift flux implementation.
() 25 And what I will do is show a few results. What I ace-FEDERAL REPORTERS, INC.
202147 37tx) Nationwide cowrage soo-3w6M6
7280 07 07 400
( )V/bw 1 am going to show is the old code version.
2 DR. CATTON: You know, EPRI has a model also in 3 the MMS.
4 MR. LILES: That simulation code they've got?
5 DR. CATTON: It is actually four rod bundles.
6 MR. LILES: He had used rod bundle datas before 7 this. He had used some of this data. ,
i 8 DR. CATTON: I am talking about the other one.
9 It's been checked against SCTF flex, also the high pressure 10 stuff at Oak Ridge.
11 MR. LILES: He's got some sort of fitting 12 coef ficients on the end that are essentially just that,
() 13 fitting coefficients to get it to collapse all the data. We 14 had some misgivings, but in our particular code, with the 15 particular gavel we are using, it is all coupled. It 16 happens to give the best results.
17 (Slide.)
18 What I showing is a CCTF run 14 data that we got.
19 This is the original Mod 1 version at the top, and this is 20 starting from the bottom of the core, working our way up.
21 What you can see is, that we have too much liquid in the
, 22 bottom. The dashed lines. You can see, I think that the i
23 second model provides a significantly improved delta P 24 response. That initial little spike you get is due to the j
i
() 25 fact that for this particular simplified model for that i
ACE-funeral REPORTERS, INC.
202-347-370) Nationwide Coserage Nn33MM6
7280 07 07 401 1 particular run that we used, we force fed the boundary
( )N/bw 2 conditions in the bottom. It is not soft. It's got gravity 3 refill. That is fairly typical. When you force a certai 4 amount of liquid, a hard boundary condition onto the heated 5 rods, you get a little bit of a belch, in itially . -
6 MR. LILES: Anyway, after it settles out, the 7 results look pretty good. As you move up, you will see the 8 same sort of thing. You will see that Mod 1 has got too 9 much liquid. We are still down near the bottom of the 10 thing, and you will see that the Mod 2 version of this 11 particular model is slightly over predicting the liquid 12 amount, but I think it is a substantial improvemen t.
() 13 DR. CATTON: Have you changed the heat tr sfer 14 coefficient?
15 MR. LILES: Yes, we did. Basically, what we have 16 been doing, we started off, essentially, with something that 17 was similar but not absolutely identical to what Murau had 18 done in their code, which is essentially a modified Bromley, 19 basically, if you get enough liquid. We are not completely 20 happen with that. That's effectively what we are still 21 looking at.
, 22 DR. CATTON: MMS has a set of correlations. ;
23 MR. LILES: I guess I am surprised he can do 24 reflood as a lump parameter thing.
i
() 25 (Slide.)
i ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-(M6
7280 07 07 402 If we go still higher in the core, we can see
()N/bw 1 2 that if you neglect that initial, small initial belch that 3 goes through, that we are doing a much better job of being 4 able to predict the thing.
5 (Slide.)
6 And finally, this is the upper plenum, basically.
7 You can see -- it looks like we are getting about the right 8 amount of liquid in the upper plenum, which we were not 9 doing before. We get too little liquid before.
10 (Slide.)
11 The rod temperatures are not bad. I am just 12 showing one set. The rod temperatures aren't bad, but they
() 13 are not quite as good as the Mod 1 code. What we are trying 14 to do is to look at heat transfer options.
15 DR. CATTON: I think you have got to do it as a 16 package.
17 MR. LILES: You do. There is no question about 18 that, absolutely. And we did that f rom the very beginning.
19 DR. CATTON: Heat transfers dropping, that will I
20 drive your void fractions of f.
I 21 MR. LILES: Absolutely. The point is that when 22 we look at it, some of them quench a little too soon. The i
23 net energy removed from the core is about right.
, 24 DR. CATTON: Do you have a delta T for the
[ () 25 minimum point? If you have, that is your problem. You l
1 ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6M6
. _ _ . _ . ~ , . _ _ , _ . _ _ _
.- ~ _ -_ . ._. .- -
7280 07 07 403
( ,1V/bw 1 can't hang it together that way. You have got to in tersect 2 heat fluxes.
3 MR. LILES: We are running some assessment with 4 this particular version of the code. We are also working on 5 heat transfer.
6 DR. CATTON: You really ought to check the EPRI
- 7 study. They did much better than this.
8 MR. LILES: For what?
9 DR. CATTON: For the temperatures and the void 10 fractions.
11 MR. LILES: It is hard for me to believe it did 12 much better on the void fractions.
() 13 (Slide.)
14 We are planning to run CCTF runs 54, 77. We are 15 running some LOFT large breaks, and we are going to start 16 running some LOFT small braks. We intend to present this at 17 the 2D/3D coordination meeting. We intend to present some 18 results at the next coordination meeting.
19 (Slide.)
20 I will just show you one example of run time 21 difference. The dashed line is Mod 2, and the solid line is 22 Mod 1. And this was a steam generator tube rupture, H. B.
23 Robinson. A steam generator tube rupture. The thing ran 24 about three times faster. What we have observed is that for
() 25 large breaks, where you have rapid transients, rapid i
ace FEDERAL REPORTERS, INC.
, 202-347-3700 Nationwide Coverage 8(n336-6M6
7280 07 07 404
(,JV/bw 1 oscillations, particularly when you start getting 2 condensation and wiggles going on af ter the accumulators 3 start emptying, we get about a 25 to 30 percent speed up.
4 That is most due to vectorization, but when we get to small 5 breaks or we are looking at a steady state transient, we 6 will get anywhere from factors of 3 to sometimes as high as6 7 speed up. In other words, reduction in running time.
8 So for small breaks, with the 3-D vessel, this 9 appears to significantly reduce the running time.
10 DR. SCHROCK: What happens in about 260 seconds 11 that causes Mod 2 to actually take more time than Mod 1 for 12 the remaining period?
(~T
(_j 13 MR. LILES: I didn't actually look at the 14 tran sien t. It has something to do with the two-phase flow 15 occurring at that point in the primary. I don' t know 16 exactly what it is. It has to do with voiding that is 17 occurring.
18 DR. SCHROCK: You mean there is no voiding up to 19 that time?
20 MR. LILES: There may be voding , but it isn't 21 much. We have run other cases that are small breaks, where 22 it looked like the thing was running a good bit faster.
! 23 DR. SCHROCK: Here it looks like it is actually 24 taking longer with Mod 2 than Mod 1.
i
() 25 MR. LILES: Over that portion of the transient, ACE FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coserage Sm336-6646
7280 07 07 405
( hV/bw 1 it may slow up. That is certainly true. But I don't think 2 we have seen a case yet on any problem we've run where it is 3 slower, net slower, including refill, where you are two-4 phased.
5 Thank you.
6 I actually got through that a bit faster than I 7 . thought I would. So.....
8 MR. MICHELSON: Are there any questions?
9 (No response.)
10 MR. MICHELSON: Does that essentially complete 11 our Agenda Item 8?
12 MR. SHOTKIN: Yes.
13 MR. MICHELSON: I would suggest that we start a 14 little bit on Agenda Item 10, because it is my understanding 15 that the noise level increases, because of an event next 16 door, along about 12:00 o' clock or so. I suspect it doesn ' t 17 break up until 1:00 o' clock or so.
18 So we ought to try to schedule our lunch while 19 the noise level is highest. That means it would be nice to 20 start the overview, and wherever it seems like a convenient 21 breakpoint on Item 10, and we will break when you say so.
22
- 23 24 l () 25 1
! /\CE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6M6
.. . - _ , ._. . .. -. . - . . . - - . - _ = - - - . - _ _ , - , . . - .,_-
[ 30 10 10 406 1
- DAV/bw 1 MR. BECKNER: I think I can probably give my 2 presentation before lunch, and then af ter lunch, the 3 committee can come back and ask questions.
4 MR. MICHELSON: Okay. I think a lot of the 5 Compendium is going to be discussion.
6 (Slide.)
7 MR. BECKNER: I am Bill Beckner. I am going to 8 provide a brief instruction on just what the Compendium is, 9 what we have done, and so forth and kick off any discussion 10 the committee might have.
11 I had to bring it up here just to show it to you.
1 () 12 I carried this over from the Stardust this morning, because 13 Zuber had my car.
- 14 (Laughter.)
15 Luckily, I was provided a copy here. If anyone 16 wants my copy, so I don't have to carry it home, you're 17 welcome to it.
18 (Laughter.)
, 19 DR. CATTON: I'd like to make the same deal.
20 (Laughter.)
21 MR. BECKNER: I am sure everybody has had it for J
i i 22 probably about two weeks no. So you may or may not have 23 gotten through it. I hope somebody will have at least
() 24 started to thumb through it.
25 Before I get in to the Compendium itself, I want i
ACE FEDERAL REPORTERS, INC.
202-347-3XX) Nationwide Coserage MG336-6M6
7280 10 10 407 to give you a brief update, as far as what the rule says.
()V/bw 1 2 (Slide.)
3 The rule was formally published on March 3rd as a 4 notice of proposed rulemaking in the Federal Register and 5 listed a public comment period through July 1, 1987. Also ,
6 at that time the notice indicated that we would have two 4 7 more documen ts coming out. First of all, the Reg Guide and 8 the Compendium, we indicated would be available in about 30 9 days. We were successful. The Reg Guide came cut and was 10 actually published for comment on April 2. That was given 11 fairly wide distribution. It was put in the Public Document 12 Room and was announced in the Federal Register. And I
() 13 believe it was sent out to all the licensees on the standard 14 distribution list. If you don't have the Reg Guide, why it 15 is available from NRC at this point.
16 Likewise, the Compendium was placed in the Public 17 Document Room in early April. This was a draft version 18 which essentially included everything except for the part of 19 Chapter 4, which we call our code uncertainty and scaling I
20 methodology. That section was still being prepared, and we i 21 will be discussing that both later on in my presentation, 22 and also Dr. Zuber, I think, has most of the afternoon 23 scheduled to discuss that.
24 As far as the current status of the report, it is
() 25 publicly available, although because of its volume here, we i
ACE FEDERAL REPORTERS, INC.
I 202-347-3700 Nationwide Cmerage 800-336-6M6
7280 10 10 408
( ?N/bw 1 have not made a lot of Xerox copies. We gave it to the 2 subcommittee so that they could read it. We are planning to
) 3 pubish this thing as a draf t for comment, so that we can 4 give it fairly wide distribution.
5 Currently, the document is completed, as far as
) 6 the draf t that we want to go out for comment. It is in our 7 publications shop right now. We have a rather wide 8 distribution list prepared.
9 Basically, we started out with one of our 10 standard R4, I think, distribution lists and added to it 11 just about everyone we could think of who would be 12 interested in this Compendium.
() 13 So we do plan to get this thing out to a large 14 number of people, so they will be able to look at it. We 15 have had some minor hold up in the past couple of days. I i 16 think we can probably get the thing printed in about half a 17 day, but we are working on the references right now. We l 18 have had some problems with references that we wanted to get 19 up-to-date, and hopefully, we will get that done over the 20 next couple of days.
21 MR. WARD: What are you going to do about the 22 section on the uncertainty methodology?
, 23 MR. BECKNER: It is in there.
24 MR. WARD: It is in there?
() 25 MR. BECKNER: I will be discussing it a little 1 ace FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coverage 800-336-u>16
7280 10 10 409 bit more in a momen t.
()V/bw 1 2 MR. WARD: Is it in this?
3 MR. BECKNER: It was probably sent out to you as 4 a separate package later on.
5 MR. MICHELSON: I never got my mine. Maybe you 6 never got yours either.
7 DR. CATTON: I got two.
8 (Laughter.)
9 DR. CATTON: Where is it going to be mailed?
10 MR. BECKNER: What do you mean?
11 DR. CATTON: I am deciding whether or not I can 12 leave my copy here.
() 13 MR. BECKNER: Let me give you an indication of 14 our problem. We have a requirement in NRC that all 15 references and NUREGs be publicly available in the 16 literature, in the Public Document Room, and so forth. Ther 17 are nearly 1000 references in there. I looked through each 18 one individually. I found 250 lab reports that I think we 19 need to reference and historically document them, but they 20 ao not publicly available. I think we will probably get by 21 tha t po in t , because it is desirable to document those 22 references, even though they are generally available only if 23 you go to the laboratory. When I went through it again last 24 n igh t , I found 41 references that were indicated as draf ts
() 25 or to be published. Just looking at those, I would say ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6646
1 6
7280 10 10 410 l %V/bw 1 probably all those 41 references probably have been 2 published since the time that was written.
i.'
3 So it is just a matter of how quickly we can get 4 those references updated. I called back home this morning 5 and started people working on them, hoping within a day or 6 two we could get those things corrected, and then we could 7 get this thing printed. It is just a matter of how much 8 trouble we have getting some of these references.
9 DR. CATTON: That is a week or so.
10 DR. SCHROCK: I have some fairly serious comments 11 on the decay heat thing, unfortunately. I would hope that 12 particular section, which is very short, could be fixed
(/ 13 before this thing goes to press.
14 MR. BECKNER: You will have to talk to me today I 15 then.
16 DR. SCHROCK: F in e . I would like to put it in 17 the record, what my problems are, at the appropriate time in 18 the discussion.
19 MR. BECKNER: So that is the status of where we 20 are.
21 The next couple of slides I won' t go through.
] 22 They are simply a listing of the table of contents, and you 23 can use them for reference when I go through and tell you 24 exactly what we've done and how we've changed the reports,
() 25 since you last saw it, which is about a year ago. That is
! ACE FEDERAL REPORTERS, INC.
l 202-347 3700 Nationwide Coserage Mx)-336-6M6
. - . . - - , . . - - - - . .. , - -- _ - --__ - . - - _ _ _ . . . - - , ~ . , . - - . . - . . , . -
7280 10 10 411
( XV/bw 1 what I will get into next.
2 (Slide.)
3 As you recall, we did show you what I would call 4 a very rough draft about a year ago, and we have had a 5 number of commen ts. I think the biggest comment, in terms 6 of, not rewriting, but reorganizing, was probably almost a 7 comment. Well, it caused a lot of work for us to to, and it 8 ws also a very useful one. That was a combination of what 9 was previously Chapters 4 and 7 to one chapter, which is now 10 6.
11 Previously, we had the LOCA issues and phenomena 4 12 described in one chapter, and then it was several chapters
() 13 later that we started describing the response to those 14 issues and work that we did. It looked good at the time to 15 organize it that way, but based on the committee's i
16 recommendation, we will bind those two sections.
17 In addition, there was a comment that the second 18 part, the response was almost solely experimental work and l 19 experimental response. For instance, critical flow, not 20 only the experimental work that was done, but also how the 21 codes respond to it. With regard to that latter comment, we 22 have tried to include a little bit more analysis, a little i 23 bit more code capability in that second part, although I
- 24 think that probably we would like to have done a lot more, I
l () 25 but we have at least put a little bit of additional 1
l ACE FEDERAL REPORTERS, INC.
l 202-347-3700 Nationwide Coserage 800-336-6646
,_ ,_.____ _ _._ __ _ .~ . - - --. . _ _ . _ . _ _ . - , . - . _ . . _ - . _ . . _ _ . _ _ .___. _
1 7280 10 10 412 (XV/bw 1 information in that section as far as analyses.
2 We also did a little bit of reorganizational in 3 Chapters 1 and 2, not too much, just basically some 4 combining of things. Also, as we were looking at the 5 organization, particularly, the change in the organization ,
6 based on the committee's recommendtions, we had a section
, 7 there on instrumentation, which seems to get in the way. We 8 felt we needed to discuss the instrumentation development 9 tha was needed in order to do this research, but it just 10 didn't go with the flow. So we have taken that and made it 11 into an appendix now.
4 12 The next major change that was done is in Chapter
() _13 4, which includes basically an introduction into the
- 14 philosophy of CL and code development. That has been 15 extensively worked on by Novak Zuber. In addition, he has 16 added an addition to our plans on code uncertainty and 17 scaling work. This section is unique to the Compendium. It 18 is unique, in that the Compendium describes completed 19 research, because the research has been completed, with the 20 exception of this one section, in which, of course, the work 21 is ongoing. And it is really a plan at this point in time.
22 And Zuber will be talking about that after this i
~
23 presentation.
j 24 In addition, we have reserved a section in
- () 25 Chapter 5 to place the results of that study, once that work f
l ACE. FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6646
,_. , _ _ _ _ _ . . . _ _ . . ~ . . _ . . _ _ _ . - - .. . - . _ , . . _ , _ _ - . _ . _
7280 10 10 413
( >N/bw 1 is completed, for the later version of this report. We have 2 also, in the version you've gotten, included some appendices 4
3 that you haven't seen before. The f acilities in appendices, 4 we have gone ahead and put in as an appendix, the EPRI 5 report that decribes their work. And of course, the 6 instrumentation appendix is in there. We have done a lot of 7 . editing. We have gotten a lot of comments from ACRS 8 consultants. Harold Sullivan came back and spent about a 9 week with us, talking to dif ferent people. So we have done 10 a lot of rewriting of what is now Chapter 6. So I will 11 think that you will see the report has changed since you've 12 seen it a year ago.
() 13 (Slide.)
14 I would also like to address a few questions that 15 Paul asked me to look at. These are some things that the 16 ACRS has brought up before.
17 The first item is on the addressing of test data 18 on certainty, and I have really put two responses for that.
19 First of all, in the method we are currently formulating to 20 determine code uncertainty, we have indicated where we would i
21 be considering data uncertainty, when we are looking at our
(
22 comparison between data and codes.
l 23 Again, Zuber can talk about that. We haven ' t
! 24 made a detailed formulation. It is a plan at this point.
() 25 It is really an overview of how we hope to proceed.
t l
! /\CE-FEDERAL REPORTERS, INC.
202 447-3700 Nationwide Coserage m)-336-6646
7280 10 10 414 (hV/bw 1 The second thing on data uncertainty is that in 2 the Regulatory Guide, we have listed a number of really not 3 models but mostly data that we are asking for public comment 4 on their acceptability on developing and comparing the best 5 estimate computer codes.
6 As it turned out, EPRI performed a study to 7 qualify experimental data, and a large portion of the data 8 that we listed in the Regulatory Guide was included in this l 9 EPRI study. I think it was done by Sol Levy, Inc. We are 10 considering trying to extend that study to the extent i
11 practical, to go ahead and at least look at the data that
- 12 may end up eventually in the final Reg Guide to make sure
() 13 that we do, indeed, have that data in the Reg Guide. That 14 is currently in the planning stages, as far as that 15 extension of what EPRI has already done.
16 That is what I am going to address on the data i
i 17 uncer tain ty .
18 The next item refers to really indexing of the 19 Compendium with the ECCS rule and the Regulatory Guide.
20 Let me point out, that there is no real index. I 21 don't think this is what the ACRS really meant. The 22 documents are dif ferent. They each serve a dif ferent but 23 related purpose. So there is really no one-to-one 24 correspondence; however, we are very sensitive to the need
() 25 that we don't have any inconsistencies between these three i
^
/\CE. FEDERAL REPORTERS, INC.
I 202-347 3700 Nationwide Cmerage 86-336-646
7280 10 10 415 (xlN/bw 1 documents. For in s tance , the Regulatory Guide, obviously, 2 should not contradict the rule. If the Regulatory Guide 3 indicates this about a certain piece of data, why that 4 should be supported by the ECCS report. We have tried to be 5 very alert to this fact, as we have gone along. And I think 6 that during the last round, while we were alerted to a 7 number of inconsistencies, we tried to correct what we 8 found.
9 So the only real response I could give to this is 10 that we are sensitive to the need of these things to be 11 consistent. If we've got any inconsistencies in there that 12 remain, we would like to know about them. We don't know it
() 13 right now, but we have gone through, particularly when there 14 was a key poin t. For instance, the Dudel-Rosenau, which is 15 being highlighted in the proposed rule revision. We were 16 very careful that the treatment of Dudel-Rosenau in the 17 Compendium was consistent with what we were saying.
18 The last item is peer review and what our plans 19 are. As I indicated, we have already put this thing out for 20 public comment. Now we are trying to get it fairly widely 21 distributed. We expect, I think, significan t commen t from 22 organizations such as the ACRS, the vendors and EPRI. At 23 least we hope we will get some significant comments.
24 We are also going to transient this to the ICAP
() 25 members and formally ask their view. Whether or not we want ACE FEDERAL REPORTERS, INC.
l 202-347-3700 Nationwide Coserage Sm-336#46
7280 10 10 416
['\p/bw v
1 a further or more formal peer review of this document, is 2 still under consideration. There have been some 3 recommendations that we have this formally peer-reviewed, 4 and we are considering it, although the type of organization 5 we might set up, we really have not settled down on at this 6 poin t in time.
7 MR. MICHELSON: Would it be useful to have some 8 kind of workshop with the industry, sort of in formally 9 discuss the document, the problems the industry may have?
10 MR. BECKNER: I hadn't thought of that before.
11 That is potentially useful.
12 MR. MICHELSON: Normal commen ting is going to (o) 13 give you a different flavor than you might get from an 14 informal discussion.
15 MR. BECKNER: We had toyed with the idea of 16 meeting with industry on the overall Rule, getting their 17 feedback and so forth.
18 MR. MICHELSON: But I think this ought to be 19 separate from the rule. There is going to be a dif feren t 20 set of people that get involved with the rules versus 21 ge tt ing involved with the details of the Compendium.
22 MR. BECKNER: That is a new idea that I think we 23 will certainly give consideration to.
24 (slide.)
() 25 The last thing is just really conclusions and ACE FEDERAL REPORTERS, INC.
202 347-370) Nationwide Coserage 8(X)-33MM6
7280 10 10 417 I really my views as far as the Compendium. The report is
()1V/bw 2 three inches thick, single-spaced, two-sided, 1300 pages.
3 As I indicated, it's got about -- almost 1000 references.
4 We feel that we have probably put in almost a staff year of 5 NRC support and probably spent about a half million dollars 6 worth of contractor support, developing this document.
7 Certainly, the quantity is there. In fact, 8 that's been part of the problem. It has been a tremendous 9 volume of report.
10 As far as the report, as we indicated to you last 11 time we gave you this report, it was originally written by a 12 large number of authors, and it was edited by a large number
+
() 13 of people. It's really suffered at times from having two 14 many authors. In fact, during the rewrites I am not always 15 sure that we have improved it. Sometimes we went the wrong 16 way.
l 17 The real problem is that we have been unable to
! 18 place one key person on it to go through the whole thing for 19 any length of time. I have looked probably at the whole l 20 report at least once, probably two or three times, but I 21 can't say that I have looked at the final version in its 22 entirety. It is simply too big.
23 We have remedied that problem, I think, to some i 24 degree. In fact, we did have some key people over the past
() 25 year that we have had them looking at least at large ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 8fKk336-6M6
7280 10 10 418 (ltV/bw 1 portions of it for long periods of time. So I hope that we 2 have remedied some of that problem, as far a consistency, a 3 far as at least people like myself being aware of the slant 4 and viewpoint that's been put forth in this document. But I 5 think only review will tell if people have done a good job 6 in that regard.
7 I think at least as far as my conclusion, as far 8 as what I have seen, I think that the product is far from 9 perfect, but I think we are ready to put it out for comment.
10 I think we can probably spend another year, if we wanted to, 11 trying to improve it and getting all the typos out, making 12 sure there's no embarrassing statements. But I think the
() 13 substance of the report is there, and we are ready to go out 14 for comment on the substance of the report. We do expect 15 criticism, and I think we desire that criticism, in order to 16 make this a good report.
- 17 I think that is really my personal view of where 18 we stand at this point in time.
i 19 MR. MICHELSON: Could I refresh my memory on the 20 use of the Compendium? I had the impression at past 21 meetings that if a person is developing a best estimate code 22 and comes in for your blessing, and to the extent that the 23 material has been used out of the Compendium, it is going to 24 be accepted more readily than if you have developed some new O 25 compilation.
(_j I
I ACE FEDERAL REPORTERS, INC.
I 202-347 370t) Nationwide Cmcrage MN)-336-6M6
. _. . . .=. _. =_ ._ , .. - . _ - .
l 7230 10 10 419
(,\V/bw 1 MR. BECKNER: Definitely not.
l 2 MR. MICHELSON: What does the Compendium really 3 mean then?
4 MR. BECKNER: The Regulatory Guide references the 5 Compendium and says that the Compendium is a source of 6 information as far as what research has been done, but it 7 also indicates that there is no guarantee that because the 8 research has been in there, in the Compendium, that it will 9 necessarily be acceptable.
J 10 MR. MICHELSON: Is there any inference that 11 because it is in the Compendium, that you will more readily 12 accept it?
() 13 MR. BECKNER: I hope not. Basically, the purpose
, 14 of the Compendium really is a historical document to i
15 document what has been done.
I 16 MR. MICHELSON: The early problem, I thought, r
! 17 with some this rewrite of Appendix K, was that you would i
18 like to be prescriptive in Appendix K and gave which l
19 correlations are acceptable and that sort of thing, and then l
l l 20 you played around finally with the idea that no, we won't be l
21 prescriptive, but that we would put out a Compendium l
22 describing the kinds of things that have been examined and 23 reviewed and whatever. But I guess it has drifted away from 24 that idea.
() 25 MR. BECKNER: As a matter of fact, we even ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage MX)-33M616
- , , _ _ . . _ = _ . _ - _ . _ . . _ _ . . _ _ _ _ . _ . . _ . , _ . . _ _ - ~ _ _ _ _ _ _ _ . _ _ _ ___.~. ~ __---_ _ _
7280 10 10 420 (y,AV/bw 1 changed the name of the document for that very purpose. We 2 did not want to go through and, in ef fect, sort date a 3 review data and give a stamp of approval, because that was 4 not the purpose of the document. The purpose of the 5 document was to document what had been done, both the good 6 and the bad, to show what had been done.
7 MR. MICHELSON: It is just a Compendium for 8 information only now.
9 MR. SHOTKIN: There is a third document, the Reg 10 Guide, I think you are confusing with the compendium. In 11 the Reg Guide, we do reference specific models for 12 acceptability.
() 13 MR. MICHELSON: I haven' t seen the current 14 version. Did you change it from the previous? In the 15 previous, you were debating it. And I didn't get a new 16 copy. Unfortunately, the mails don' t move too fast in my 17 direction.
If MR. SHOTKIN: It is in the Reg Guide, where we do 19 recommend specific models.
20 MR. MICHELSON: And in the Compendium, you 21 describe some of the details on those models.
22 MR. SHOTKIN: The Compendium is, just as Bill 23 said, a road map to all the research that we know of that 24 has been done over the past decade or so on ECCS.
() 25 MR. MICHELSON: Even using the correlations in ACE FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coserage 800-334 6646
. ~ . _ .
7280 10 10 421
( )N/bw 1 the Compendium does not give you any more edge than dreaming 2 up a new one.
3 MR. SHOTKIN: The ones in the Reg Guide do.
4 MR. MICHELSON: I hope if you prescribe them 5 there --
6 MR. BECKNER: It is certainly useful that if you 7 were wanting to develop a model in a certain area, wanted to 8 know what was available, that is the purpose of it, but you 9 would have to be the one who would have to say to the 10 licensing people that this is acceptable, and here is why.
11 MR. MICHELSON: What is now thought to be the 12 purpose of the Compendium? Just to make it handy for
() 13 people? Could you tell me what the purpose is
- 14 MR. BECKNER
- I think, actually, even before the 15 ECCS rule was linked with the Compendium, we had a user need 16 from NRR to write such a document, because we had performed 17 almost $1 billion worth of research, and it was the desire 18 to try to document that research before everyone died or
! 19 went away, and they all forgot where all these things were 20 stored in all these reports. That effort to write such a
- 21 historical document was really hastened along by the fact 22 that we were getting into the ECCS rule.
23 MR. MICHELSON: Then, in terms of reviewing the 24 Appendix K rulemaking papers, there is no linkage to the
() 25 Compendium; is that right?
. ACEJFEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coserage 800 336-6646
- . - . _ . .._, _. _ __ . . . _ _ _ _ . - . _ . . _ _ _. _. _ . _ . __ ___ . _ _ _ . _ _ _ _ . _ _ . . ~ . _ _ _
7280 10 10 422
( hV/bw 1 MR. SHOTKIN: There is. It is a diminishing 2 linkage, the rule. Then the Reg Guide is a stronger linkage 3 MR. MICHELSON: You are getting back in to a 4 dif ferent answer than I thought I heard a few minutes ago.
5 MR. SHOTKIN: The Compendium used to be called 6 the Technical Basis. So if somebody, the public, say, how 7 do you know that you have done enough research that you can 8 come and change the rule?
9 They will say, well, here is the Compendium.
10 This is all the research that has been done. That is the 11 technical basis for it.
12 DR. SULLIVAN: Is that what the document is?
) 13 MR. SHOTKIN: No, it is called the Compendium, 14 because the lawyers didn' t want us to call it a Technical 15 Basis, due to legal problems.
16 MR. MICHELSON: In citing your technical basis, 17 do you cite the Compendium, specifically? I mean, you will l- 18 pick the parts of the Compendium that you are using and cite 19 those. Then the NRC will review those parts for your 20 particular application; is that the idea?
21 But we don't need to review the Compendium at 22 all, as far as the rulemaking process. We are not going to 23 writo you a letter on the Compendium, because it is not 24 cited as part of the process.
() 25 ACE FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coserage KO-336-6M6
I 7280 10 10 423 p$V/bw
(, 1 MR. WARD: What is sort of puzzling is, you have 2 said it is sort of decoupled from the process, but you want 3 to go through this kind of elaborate peer review, and it 4 sounds like, on one hand you are describing just something 5 that goes on the shelf. It is a history. You feel 6 responsible for getting it published. Why all the peer 7 review, then?
8 MR. BECKNER: I think this document is linked to 9 the rulemaking, in the sense that when we wrote the original 10 rule, we did not have the research to support more 11 realistic. We had to use conservative calculations, because 12 we didn't have the research basis. Now we are revising the
() 13 rule, proposing to revise the rule, using more realistic and 14 more accurate calculations. The reason we can do that is 15 that we have done, in effect, this research represented by 16 this report. Now where we want to break the linkage here is 17 really in two areas. Number one, we did not want to review 18 this ten years worth of research to put a stamp of approval 19 on each and every piece of data in this report. Likewise we 20 had some legal concerns that if we said this is the basis 21 for our rulemaking, if someone found a flawed piece of data 22 in here, this would jeopardize our rulemaking. This wasn' t 23 the purpose.
24 Taken as a whole, this is our basis, but
() 25 individually, it is really a summary of what's been done.
ACE. FEDERAL REPORTERS, INC.
202 347 37m Nationwide cmerage m3356646
I 7280 10 10 424
()N/bw 1 MR. LAUBEN, Norm Lauben. I just wanted to make 2 one comment, in terms of historical perspective about this, 3 and that is, when the ECCS hearing began in 1971, the first 4 one, one of the serious criticisms we got from technical 5 members of the public was that they didn't know where to go 6 to find out about ECCS. And I think when we are talking 7 about peer review, there's different kinds of peerage, if 8 you will.
9 There are peers who are very familiar with ECCS, 10 and there's technically capable people who may want to 11 become in te rested , and it would have been extremely valuable 12 at that time, for instance, to have something like this
() 13 available for them to look at. So we are, in way, I think, 14 satisfying some of the criticisms we had at that time for 15 some document that describes the work that's been done on 16 ECCS. So some technical member of the public, not 17 necessarily that closely familiar with ECCS, would have a 18 single document and source to go to to begin to look at 19 these issues.
20 MR. MICHELSON: It is a fine document. It is .
21 very helpful to me, because I am sort of a member of the 22 public that doesn't understand all the details, and it is 23 nice to have it package 3 together.
24 I was trying to determine, though, what legal or
() 25 regulatory stature, if any, it has. I gather it is not real ace.Funnart. RnvanTnas, INC.
202. W-MW Nanonwide Cmerage N3 MW6
_ . _ _ _ . _ _ . _ . . . ..- - _ . . _ _. - - ~ _ _ _ . _ _ _ _
4 1
7280 10 10 425 j (hV/bw I clear, from what you said, it still isn't real clear to me I 2 what regulatory stature it has, in terms of a person coming 3 in with an application and a new analysis for Appendix K.
- 4 DR. SCHROCK
- It seems to me, it doesn't have 5 such stature, but I would agree wholeheartedly with what 6 Norm just expressed. It will serve an invaluable need, if l 7 this rule revision goes to hearing, such as we had in the s, 8 '70s.
1 .
i 9 MR. LAUBEN: I don't believe -- Bill may know as i
10 well as I do, but I don't believe it really does have any 11 " clear regulatory stature" at all nor was there any l
12 intention that it did. But I think it will also provide a l
f () 13 very helpful document, not only for the general public, but
- 14 also for people in NRR or anyone else who maybe gets 1
15 transferred into this or out of it, that they have some i
j 16 place to start with.
i
- 17 We will talk about Novak's uncertainty, but that
]
18 is a very, very good example of how one might go through 19 what one needs to go through to comply, as a methodology, f 20 Eor example, to help you comply with what the rule J
l 21 requiremen ts are.
22 MR. MICHELSON: In terms of the subcommittee, it
. 23 appears to me, and I will solicit other comments, it appears
- 24 to me that the subcommittee, if they write a letter on the
() 25 Compendium, it should be kept separate f rom a letter on the l
j ACE FEDERAL REPORTERS, INC.
J 202 347-3700 Nationwide Coserage 800-336-6646 1
a 7280 10 10 426 (hV/bw 1 rulemaking process, because those are two separate subjects.
2 Regulatorywise, they are two separate considerations, and 3 the uncertainty business will come into our rulemaking 4 comments, to the extent of how we believe we can play a role 5 in terms of its technical detail. We will comment as a part 6 of the Compendium.
7 I think we need two separate letters, if, indeed, j 8 we even write a separate letter, on the Compendium, but I 9 think we should not link the two together. Clearly, you are 10 naot linking them together in your regulatory thinking, at i
11 least from what you said just a moment ago.
12 So I think we want to be doing the same.
() 13 Do any members have any comments?
14 MR. EBERSOLE: Let me just read the sentences in l 15 the front end in the abstract.
16 "The United States emergency core cooling system 17 will require, for lightwater reactors, to provide cooling of
! 18 the reactor core in the event of a break or a leak in the 19 reactor pipeline. These accidents are called loss-of-l 4
1
- 20 coolant accidents, LOCAs, and they range from small to a
! 21 postulated full break of the largest pipe."
22 Well, we stop at that point.
23 Most of the loss-of-coolant accidents can be
]
,i 24 generated, of course, by machinery turning off. It is not l
i
() 25 here.
- ACE FEDERAL REPORTERS, INC.
2t)2-347 37(X) Ltionwide Coserage 8m33M446
_ -- _..____._.__.. _ _--_ _ -._. _.____.___.._. _ _ .-,.__ - . _ _ , _ _ . , . . . , . - _ _ . _ , . ~ . _ _ . _ - _ _ _
7280 10 10 427 1 MR. BECKNER: This is LOCA, as defined by 50.36.
(JV/bw 2 MR. EBERSOLE: So it is delineated very sharply.
3 It doesn' t even get deeply into the secondary.
4 MR. BECKNER: That is right. And this document 5 has been limited to ECCS.
6 MR. EBERSOLE: I don't know whether there's 7 enough emphasis on the degree of limitation therefore, the 8 fractional aspect of reactor safety, which this covers. It 9 only covers a tiny fraction.
10 MR. BECKNER: That is correct.
11 MR. EBERSOLE: But one reading this, not aware of 12 the percentage of reactor safety, will tend to get a
() 13 distorted opinion of what he is looking at.
14 MR. BECKNER: They may also think this is of the i
15 most concern, since we spent so much money on it.
i 16 MR. EBERSOLE: That is right. So there is a 17 distorted generalized viewpoint that you get when you pick
! 18 this up by itself.
I 19 MR. BECKNER: It focuses on a very specific role.
20 DR. SULLIVAN: Bill, when I read it, I thought l
21 that your purpose for having it was to provide the technical 22 base for giving up the conservative approach that. was 23 Appendix K, now, and going to a best estimate. But if it is
- 24 a historical document of what has been done, it doesn't even
()
l j 25 i
I l
l
/\CE. FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-33M64
i 7280 10 10 428 N/bw 1 do that.
2 MR. BECKNER: I don't know, Harold. We are 3 skirting legal semantics, maybe. I view it as a three-stage 4 thing. At the top, you have the rule, which is a fixed 5 requiremen t, an absolute. The next step is the regulatory 6 guide, which has a lesser degree of regulatory significance, 7 but it provides an acceptable method of meeting that fixed 8 rule. The next layer down is a Compendium, which really has 9 no requirements for acceptability, but is a broad-base of j 10 information that supports this whole pyramid. That is the 11 way I view it.
12 We have had to be very careful not to give this
() 13 Compendium any regulatory significance, because we don't 14 want to go through and put the NRC stamp of approval on 15 every page of this document. It is an information base.
16 MR. MICHELSON: I think we are reaching the point 17 where we have to go to lunch.
18 MR. BECKNER: Someone must agree with me.
19 (Laughter.)
l 20 MR. MICHELSON: Are there any comments that can' t 21 wait until after lunch?
22 (No response.)
23 MR. MICHELSON: We will adjourn now and reconvene 24 at 1:15.
() 25 (Whereupon, at 12:15 p.m., the meeting was ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6646
. . _--_ _ __ _ -- ~ _._ _ . _ - .. - - _ . - . . _ _ _ _ _ . . , - .
7280 10 10 429 recessed, to reconvene at 1:15 p.m., this same day.)
hV/bw 1 2
3 4
5 6
7 8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 O 25 .
ACE FEDERAL REPORTERS, INC.
202-347-37(x) Nationwide Coserage 800-336-6646
7280 10 10 430 AFTERNOON SESSION
()N/bw 1 2 (1:15 p.m.)
3 MR. MICHELSON: We will reconvene now and proceed 4 with the discussion of the Compendium.
5 I think the Staff ought to be here,-more or less.
6 I don't know what is the best way to proceed. I 7 am sure that all of our members and consultants have 8 comments. So I think the simplest thing to do is to go from 9 left to right. And if you wish to reserve until later, then 10 just pass on to the next, and we will pick you up later.
11 DR. TIEN: This is about the Compendium?
12 MR. MICHELSON: That is right.
() 13 DR. TIEN: I just want to make a general comment.
- 14 I think this is supposed to be a documentation of the 15 technical bases and history. Coming from academia, I took a j 16 -quick look, especially in terms of the references. This is i 17 a general comment.
- 18 Most of them are really reports. I think if you 19 want to give some stability and so on, I think we should do 20 the best we can to try to get some published journal 21 literature instead of all the report forms, which is 22 generally regarded as not really rigorously peer reviewed.
23 I can see many, many places, correlations of all different i
24 kinds, but they were always referred through some kind of l
l
() 25 NUREG report, which can be replaced by some standard i
ace FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coverage Mn336 6646
7280 10 10 431 I-s1V/bw v 1 literature and that would help a lot.
2 I think this is the feeling. If you are really 3 talking about the technical basis, to give some credence, 4 and so on, I think every ef fort should be made to make 5 referee-published literature instead of reports. That is 6 one general comment. Later I can come back to some other 7 comments.
8 DR. SCHROCK: My general comment, first of all, 9 is that I have done only a cursory review. I have read some 10 sections and skipped over others. It is a big improvement 11 over the document that we reviewed last summer.
12 With respect to the decay heat section, 6.12,
() 13 there are some serious problems that I would like to 14 discuss, but before doing that, I would like to say that I 15 feel somewhat responsible, because I think that the person 16 who write this became preoccupied with one specific aspect 17 of the decay heat problem, which I erroneously commented on 18 at our next meeting. And so I apologize to all concerned 19 for my error and the agony that inflicted on a number of 20 people in getting it resolved.
21 So the things I want to comment on in the decay 22 heat section involve, essentially two aspects. One is the 23 general nature of that section and its importance to the 24 whole procesw of rule change and the specifics of a couple
() 25 of errors that are contained in it, as it is now written.
ACE FEDERAL REPORTERS, INC.
l 202-347-3700 Nationwide Coseraec mb3364M6
4 i
i 7280 10 10 432 1 First of all, the general aspect of it, it seems 4
(])V/bw
] 2 to me, that the topic deserves a little more in-depth 3 discussion than it received, inasmuch as decay heat 4 improvements in the last 15 years are one of the major, if 5 not the major justifications for going to a rule change.
6 And so recognizing its importance in that sense, it seems 7 out of proportion to have two pages of discussion about the t
8 decay heat problem, you know, in a document that is four
- 9 inches thick that is providing the technical basis for a 10 rule change.
11 So that is the general comment. I suspect that
! 12 that would take more time to fix, but I certainly would hope i
() 13 that these two pages of the document, the errors that they 14 con tain before you send this out for any peer review that 4
l 15 you are talking about this morning -- so let me run quickly i 16 through what the problems are.
- 17 In the first paragraph, section 6.12.1, entitled 18 " Phenomena," there is an attempt to identify what are the 19 key phenomena that need to be addressed. This falls far 20 short of doing that task. There are five items that ought i
21 to be addressed. First, there is the fission products decay a
22 heating, as a result of fissioning of several dif ferent l 23 nuclides that have dif ferent yields, and a consequence of a
- 24 the different yields is that their contributions to the
() 25 total decay heating are dif ferent. What that means is that ACE FEDERAL REPORTERS, INC.
- 202-347 3700 Nationwide Cmcrase Mn33MM6
, 7280 10 10 433 as burn up occurs in the reactor, there is a change in the
()V/bw 1 2 decay heat, which is in consequence of the composition 3 change.
] 4 So U 235 and U 238 and plutonium 239 are the main 5 contributors plus other less significzant contributors have 6 to be treated individually, in order to get the kind of 7 accuracy that people sought in decay heat evaluation.
8 The second item is the one that the author of J
9 this section became preoccupied with. That is the decay i
10 heat from actinides. Actinides are a part of the overall 11 category of activation products. N-gamma reactions in fuel
- 12 and structural material result in radionuclides, which
() 13 produce a contribution to the decay power and that has to be
! 14 evaluated in some detail.
15 The dominant contributors to that in the time 16 frame for the ECCS considerations in an uranium fuel reactor 17 wil be uranium 239 and 239 neptunium. Those are the I
18 conversion intermediate nuclides in the conversion from U 19 239 to plutonium 239.
20 The third item is the effect of neutron capture i
i 21 in the fission products themselves, which is not a major 22 factor during the short term in reactor safety
! 23 considerations but for small break loss-of-coolan t acciden ts 24 we tend to get in a time frame where that is a significant
() 25 in fluence on the fission product contribution.
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-33MM6
7280 10 10- 434
(?N/bw 1 The fourth item is the fact that there is a 2 space-time dependence of the fuel fissionable isotope 1
3 concentrations. This results in different normalized decay 4 power curves for every pellet in the reactor. That means i
5 that the computation of the decay power is seriously
! 6 complicated by the evolving concentrations that are present 7 in the reactor.
8 The fifth item is the spatial distribution of the i 9 energy deposition, which is different from the spatial l 10 deposition -- spatial distribution rather -- of the 11 fissioning itself, going to the range of the gamma rays.
1
- 12 In the operating mode, you have prompt gammas as 1
i
() 13 well as delayed gammas, and in the decay heating problem, 14 you have delayed gammas only, of course, but the fraction of 15 the total energy release in the shutdown mode, which is N 16 gamma radiation, becomes much higher that during normal j 17 operation.
- 18 So that is another spatial effect that has to be 1
19 evaluated in order to get accurate values of the spatial 20 deposition of energy throughout the reactor column.
21 So those are the fifth key phenomena that really 22 need to be addressed in looking at the decay heat problem.
23 Now as to the errors that are in the current 24 documen t in the Article 612.2, the statement is made that 2
() 25 neither the ANS 1971-73 draft standard nor the 1979 ANSI, ACE-FEDERAL REPORTERS, INC.02-347 3700 Nationwide CoseraFe BOO-336-M46 J
. . . - . _ . - _ . . - - , _ - . . . - . . . . _ , , . , - _ - , _ _ _ _ , _ , _ _ , , _ _ _ . - . . . . . ~ . , . . _ . _ , , - , _ , _ _ . , . . , - _ , , . _
7280 10 10 435
()N/bw 1 ANS approved standard addressed the actinides. This
- 2 statement is blatantly false. Both of those standards 3 addressed the actinides. Both of them contain explicit 4 equations for the contribution to the decay heating from 239 5 uranium and 239 neptunium.
6 So that statement is simply incorrect and it i
7 ought to be changed.
l 8 In addition to that, the 1979 version of the ANS 9 standard requires that the user shall calculate the other 10 actin ide s' contribution, as well as all activation products.
l 11 That has to be added to the decay power calculated for the 12 fission products.
() 13 The second error is a quotation later in the same 14 article. I will give the quotation. It says " Actinide 15 contributions to decay heat began to be considered in the t
16 mid-1970s."
17 That statement is similarly false.
18 MR. BECKNER: Excuse me, Virgil, where are you on l 19 tha t?
20 DR. SCHROCK: I am sorry. It is in 12.3. 612.3,
! 21 the second paragraph, first sentence.
22 In the 1971 standard, as I have already l
1 23 indicated, there was a prescription f rom the 239 actinide i 24 contribution to the decay power.
l I
() 25 In the early literature, it was a well-recognized l
ACEJPEoERAL REPORTERS, INC.
202-347 37(t) Nationwide Coserage 8(X)-336446
?.,__.--. . . , _ _ _ _ _ _ _ _ _ _ , . . , _ , _ . _ . _ _ _ . _ _ _ _ . ,_-, __
7280 10 10 436 fact, even in the '50s, that there is a contribution to
()N/bw 1 2 decay power from actinides. So it is not as though we woke 3 up and discovered the existence of actinides as a decay heat 4 source'in the mid '70s, which is af ter we were already into 5 the business of licensing the reactors.
6 In addition to that, it is explicitly stated in 7 10 CFR 50, Appendix K, I.A.3, that the actinides and 8 activation products must be calculated to the fission 9 product con tribution.
10 The difficulty here is that this error leaves the 11 impression that there is a significant contributor to the 12 decay power that may have been omitted from early, if not
() 13 recent license evaluations.
14 I don't think you want to make that in ference 15 I would note, however, in passing, that other 16 actinides and activation products are not treated in TRAC 17 and RELAP-5, if the current documentation is accurate. So 18 it may well be that there are calculations that have been 19 used for comparisons in which other actinides and activation 20 products have not been included as they should be.
21 Lastly, the document omits to comment on the 22 error component which is associated with simplifying the 23 space-time decay heat problem to a generic core average, 24 which is what the NRR Staf f has indicated they think is a
() 25 suitable thing to do and which, indeed, has already been ace FEDERAL REPORTERS, INC.
202 347 37u) Nationwide coserage 89)-3346u6
. _ - - - , - - - - . - - - - - _ _ - . _ _ . _ . . , ~ . - - . , , _ , , . - . . . - _ , - _ ,. , .__._, .. . . - _ -
7280 10 10 437
( )N/bw 1 done in the approval of the licensing basis for GESSAR-2.
2 This is serious, since the rule change proposes 3 to use the best estimate on a 95 percent confidence level.
4 And the additional uncertainty which is associated with a 5 simplification of the spatially dependent decay heat curves 6 into a single generic curve for a given irradiation history, 7 is something that has to be addressed in error assessment.
8 I thought about your problem of references on 9 this. I noted one which maybe you can't use. Maybe you can 10 find a way of using it. There is a review of GESSAR-2, l 11 Appendix 6-A, which is cited in a letter f rom Erlinger to 12 Sharon, dated June 16, 1982, which is something that this
() 13 committee reviewed at the time the GESSAR thing was under 14 discussion.
15 A second possibility is a paper which I will 16 present at the ANS meeting in Dallas. One of my students 17 did a thesis this past year on the effect of the spatial 18 variations in core composition on addition error in a 19 generic average curve.
20 Those are the comments that I wanted to make.
21 MR. MICHELSON: Jesse, do you have any comments?
22 MR. EBERSOLE: I have a sof t comment to make. It 23 comes from some 20 years back, when I picked up papers that 24 said nuclear safety research and defined contained therein
() 25 virtually nothing except that due to large LOCA ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cmcrage 800-33M6M
7280 10 10 438 1 experimentation and analysis. It took me a while to learn
()aV/bw 2 that I didn' t need to confine much of my attention to that, 3 but elsewhere. I thought perhaps you might add, as an 4 interf acing statement here, after the first sentence in the 5 extract, some sentence that says, there are many other and 6 more probably ways of causing loss of core cooling, such as 7 by loss of flow or malfunction of secondary and tertiary 8 cooling systems for which emergency systems and procedures 9 are also provided.
10 And merely acknowledging this as being part of a 11 larger problem, it has a tendency to be loss within itself 12 as a topic.
( )) 13 MR. MICHELSON: David?
14 MR. WARD: I will pass.
15 MR. MICHELSON: Ivan?
16 DR. CATTON: I think the report is much better --
17 at least it's table of contents is much better. I didn't 18 read it in much depth. I am a little disappointed that it 19 is only going to be a history of reference work. It seems 20 to me that the taxpayers' money deserves a little more than 21 that.
22 I am not sure how useful further review by myself 23 would, if it is not to be part of the Reg Guide or reference 24 to the Reg Guide.
() 25 MR. MICHELSON: Harold?
l i ACE FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coserage 8m-33MM6
. - . - - - . - , - --__-----_,~- - -.
7280 10 10 439 1 DR. SULLIVAN: I'd just like to say that I think
-()N/bw 2 you guys did an excellent job trying to put this thing 3 together. The weight has gone up from the last one that I 4 saw. I think the improvements that you put in it have 5 really helped it out.
6 So I commend you on that.
i 7 The comments that I am going to give you are
[ 8 meant to be constructive. I don't know even exactly how yu 1
9 would use them, because it is going to go out without these j 10 commen ts in them. So if you can somehow incorporate them 11 right in the f ront of the document someplace. There ought i
j 12 to be a statement about what the document is. I thought I l
J
() 13 knew. I dare say each of us thought they knew when they 14 were reading it.
]
15 So if it is a history document or how it is going l
i l
16 to be used to support a rulemaking hearing in however way 17 that you decide, I think it ought to be stated.
- 18 I guess the thing that I read in it is the July 19 date, still the date at which it is issued?
20 MR. BECKNER: The date that it is issued?
21 DR. SULLIVAN: I think it says in here, your l
7 22 comments about July 1.
1 23 MR. BECKNER: Right. We have tried to tie the 24 whole package together and have comments when the rule i
() 25 comments were out.
i[
ACE FEDERAL REPORTERS, INC.
! L 202-347-37m Nationwide Coverage 8m33w46
7280 10 10 440 Ou/a- 1 1 aes e xee - eeooie mev met aeve taie aecomeet 2 for an adequate length of time, and they may well request an 3 extension.
4 5
6 1
7 8
9 10 11 12 O 13 14 15 16 17 18 19 20 21 l
22 23 l
24 l
l O 2s l ace FEDERAI. REPORTERS, INC.
l 202 347-37m Nationwide Coserage NWL3 tMM6
7280 13 16 441 DR. SULLIVAN: There was a program run in
()Nbur 1 2 Research called Heavy Section Steel. That showed the 3 instantaneous break and all the problems that that caused.
4 I thought you ought to put a write-up on the heavy section 5 steel program in there because it also discusses the 6 probabilities of having a circumferential break and actually 7 have the guillotine rupture or split and the probability of 8 the break itself.
9 I believe that strengthens the document, and the 10 fact that what you are doing is analyzing a very improbable 11 event, it allows you to look at it probabilistically, which 12 you do a little bit in the back.
() 13 Another addition I think you ought to put in 14 there is in the code section, the RELAP-4 code. It is not 15 mentioned, but it is later on when you start looking at some 16 data comparisons.
17 So, particularly the MOD-6 and MOD-7 line, which 18 is the end of that line, is, I think, important to reference 19 in there, and it gives you some base for what really 20 advanced codes are, and if you could compare those.
21 I guess maybe this comment is not really that 22 constructive any more, but if you look on page 1-1, it gives 23 you an indication, I think, of the extent that you need to 24 do some modifications to the document. You are going
() 25 through there and listing things that are in the Appendix K, ACE FEDERAL. REroRTens, INC.
I 202 347-3700 Nationwide Coserage 80h33MM6
7280 13 16 442 1 but one of the things you left out is coolable geometry. It
()Vbur 2 is left out almost all the way through the document.
3 That was important. There are other 4 characteristics that were required by the regulations to 5 demonstrate. Coolable geometry was one of those, also.
6 If you look at a list, say 2-5, you really never 7 get into coolable geometry. You are still going to require 8 that, and I hope that the reg guide does say that.
9 I think the implication is that if you meet the 10 oxidation criteria -- the two oxidation criteria in the 11 inlet peak clad temperature, that you meet the coolable j 12 geometry. I don't that was the intent.
] () 13 on page 430, you wind up coming up with a list.
j 14 What you have down is required thermal hydraulic capability 15 of best estimate codes. There is a list of items down i- 16 there, and one of them is radiation by steam, and you are '
) 17 talking about the core heat transfer and drop by radiation.
l 18 Those are things that are currently not in some I
k 19 of the best estimate codes, and by saying they are required, 20 are the current codes unacceptable?
21 What I am trying to get you to do is to relook at i
22 this list and the things that are on there that are not l
i I 23 required. In fact, you probably ought to change the word
! 24 " required.
() 25 If you go on and look at those paragraphs that l
ACE-FEnERAL REPORTERS, INC.
202-347-37tio Nationwide Coserage Sm33trM>M I
7280 13 16 443
( ,N/~
a Vbur 1 discuss the others, such as the BWR, there is no list at all 2 of the required features.
3 Then, if you go to the end of that chapter and 4 look at the requirements, there is a section called 5 " Requirements," and it says go to this list. Those things 6 ought to be in the other ones, and with these additions.
7 There are no additions in the other list.
8 So there is a problem with the consistency of the 9 writing between sections.
10 In Chapter 5 you talk about the codes again, and 11 some of the codes that you describe are never really ever 12 used. You never made a data comparison for them. Some of
() 13 the data comparisons that you did make, the codes aren't in 14 there.
15 There was RELAP-4, MOD-5. You never discussed 16 the RELAP family at all. Then later on you used it as a 17 reference.
18 I notice that you tried to get a consistent 19 format for when you are discussing a code, that it had the 20 objectives, background, et cetera. But it was like the 21 objectives of the code, which I thought were reasonably 22 impo r tan t , like FRAPCON and COBRA and the TRAC BWR codes 23 didn't have an objective. They are just not in there.
24 And such as a background. RELAP-5 was the only
() 25 one that had a background reference. So although even ACE.FunnRAL IturonTuns. INC.
202 347 37m Nationwide cmerage s o-336-r M 6
. . - - . . . . - -- .- . _ - - . --- -. . . ~ - . ~ . _
)
- 7280 13 16 444 1 though you tried to make them consistent, there is still a 1
()Vbur 2 f air amount of inconsistency.
i 2
3 The developmental assessment is another area like i 4 RELAP-5. Didn't have a developmental assessment, and not i
j 5 but one of them had a quality assurance, and that was the 4
6 TRAC BWR code.
! 7 If you want to go to Chapter 6 -- and I am just 1 8 giving you the overall comments. I will give you a 9 document, and you can carry it home with comments.
,' 10 The biggest objection I have to the chapter is t
j 11 still th'e one that I had before. That is that there is very I
j 12 little code data comparisons, and if you take critical flow, I
() 13 the first section of it, and you look through it, the only
! 14 code that was used in there is MODSENS. To me, you need to 15 get a good base of looking at these codes versus data.
16 If you go to the heat transfer, there are places 17 in there that there is no write-up of how any code did, and j 18 particularly the heat transfer section talks about grids, i
{ 19 radiation, steam, the natural convection if you have an i
20 early stalled flow, yet you never say if these things are i
l 21 included in the codes or not. So not only are they not in i 22 code comparisons, they are not in descriptions of. Does the
, 23 code have all this stuff in it, and how important is it?
24 probably the worst section like that is the 1
() 25 reflood heat transfer. I think that whole section didn't i
j ACE-FEDERAL. REPORTERS, INC. l f 202 347 37(A) Nationwide Coserage MG 33MM6 i
1 ,_.r ,. -- -. --c.. ---.-,cm.-.m . . _ _ .-
7280 13 16 445
()Wbur 1 have but one code comparison, and it was with COBRA TF. So 2 you just don' t have any feeling for how much the codes 3 compared with the data at all.
4 That to me is the section that really ought to 5 give you the demonstration that the codes actually will do a 6 job on all of this.
7 MR. BECKNER: We recognize that, Harold, 8 throughout there. We did that a lot less than we desired to 9 do it.
10 DR. SULLIVAN: Bill, there are still sections 11 that don't have any.
12 MR. BECKNER: Where are those?
() 13 DR. SULLIVAN: I think in every case you can put 14 something in there, particularly the reflood, because of all 15 the CCTF, SCTF work that has been done.
16 One thing that I noticed is that you talk about 17 --
in fact, it is the beginning of 4, I guess -- you talk 18 about the things that you were looking at in terms of large 19 breaks, small breaks. BWR large breaks, and it doesn't 20 mention small breaks. Neither does it say anything -- that 21 they are not a problem or anything. Later on in the small 22 break section you bring up small breaks again.
23 Is there a reason why it was left out that I 24 don't know?
() 25 MR. BECKNER: Yes. Originally, it said that BWR ace. FEDERAL REPORTERS, INC.
202 m-Um Nationwide Cmerage N00_116 6M6
7280 13 16 446
( )Nbur 1 small breaks turned into large breaks. That is why it 2 wasn' t discussed.
3 Whether that wording is still in there I don' t 4 know. It was located in two places, the part that Novak 5 wrote and also the section in Chapter 6, and I don' t know 6 whether that wording is still in there or not, but that was 7 the intent, that the small break turn into a large steam 8 line break.
9 MR. 2UBER: That section was written by several 10 people, several cooks, and we just didn't have time to go 11 through it consistently to put it in a uniform way. If we 12 had maybe another month or two months, it would have been
() 13 much better, but as it was written, with several people, 14 some of them are self-consistent and there is a 15 discontinuity. They were written by different people with 16 dif f eren t experience, and they were really not very 17 integrated.
18 An example is this BWR section.
19 DR. SULLIVAN: When you look at the test 20 facilities for BWR, there is no TLTA. Later on, that is 21 about one of the biggest uses of the code data comparison.
22 MR. BECKNER: I guess there's two sections that 23 have that. There's some introductory material in 4. Then 24 there's details in Chapter 6. You probably don' t know what
() 25 you are referring to either. We can go back and check.
ace-FEDERAL REPORTERS, INC.
202 m.37m Nationwide cmerage sm 34ua6
.. . .-. . ~ . _ _ - _ _ ,_ _ . _ . . . - .
i P
j 7280 13 16 447 Wherever it was, it discussed the l ()Vbur 1 DR. SULLIVAN:
j 2 integral experiments and the separate ef fects.
3 MR. BECKNER: TRT is discussed in detail in j 4 Chapter 6.5, also the small break LOCA.
3 5 DR. SULLIVAN: The facility is not described up 6 front. That is the problem.
t l 7 MR. BECKNER: It is in the appendix.
i 8 I think what you are doing is you have read the 9 introductory material to Chapter 6, which is basically an q
10 extract of some of the stuff -- I am sorry, that is in I 11 Chapter 4. It is an extract of the stuff that is in Chapter
! 12 6.
() 13 DR. SULLIVAN: Even the introduction, the 14 abstract -- well, not the abstract. It looks like to me l 15 there was a conscious effort to take TLTA out of it. I 16 don't know why.
i 17 MR. BECKNER: Harold, Chapters 6.4 and 6.5 have
', 18 extensive tests, large break tests and small break tests.
19 DR. SULLIVAN: In 6, look on page 2. You put 20 FIST in it, but you didn't put TLTA.
I 21 MR. BECKNER: Where are you, Harold?
l 22 DR. SULLIVAN: Page 2.
I 23 MR. BECKNER: What chapter?
l I
24 DR. SULLIVAN: Page 2, Executive Summary.
() 25 So I don't know why, is there a reason?
l l
ACE-FEDERAL REPORTERS, INC.
202 347 3700 Nationwide Cmcrage MO-3 %(M6
7280 13 16 448 1 Then in the in troduc tion it is the same way.
( )Nbur 2 I just wondered if for some reason you maybe felt 3 the data wasn't good enough.
4 Then it is again on 1.3, kind of that middle 5 paragraph. Again, you have got a list of experiments.
6 Well, I have a fair number of other comments, but 7 I think -- I thought Chapter 6 was probably the most 8 important one, and the code data comparisons were generally 9 not in there.
10 MR. BECKNER: That is a valid comment. It 11 continues to bother us. We accept that.
12 DR. SULLIVAN: And another thing that I think you
() 13 really ought to work on is the conclusion. Every single one 14 of the conclusions in 6.
15 You know, the conclusion ought to be that the 16 codes that model this, they do well or it is not important 17 and it was left out for a purpose, or something.
18 And I think, let's see, I guess the thing came 19 home on the once-through steam generator section, and the 20 conclusions to that section say the data is there and an 21 advanced code can be verified by it, and that is the 22 conclusion, and it doesn't say anything about the 23 capabilities of the code. There is very little data in 24 there to start with. That can give you a whole bunch of
() 25 problems, and then you end up by saying the code -- you ace-FEnERAE REPORTERS, INC.
- 202-347-370t) Nationwide Cmerage 8(u)-336-6M6
t 7280 13'16i 449 1 can' t verify it.
- ()Vbur 2 And the U-tube section is the same way. The U-
- 3 tube steam generator. It talks about a lot of different 4 things in there, things like natural circulation, and then 5 there are no code data comparisons in there at all,'and it i
j 6 leaves you hanging because it doesn' t say anything about the 7 code capabilities.
8 Is that enough?
I' l 9 DR. CATTON: If it is only a historical document, 10 why don't you delay a month and fix some of these things?
4 11 MR. BECKNER: Ivan, I think we are going to take 12 serious exception with saying that this isn't important to l
() 13 rulemaking. That is not at all what we are saying.
l 14 What we are saying is that this thing is not j 15 going to carry a regulatory requirement with it, but we
- 16 think this is very important to rulemaking, and we want this 17 type of criticism to try to whip this thing in to shape. .
i 18 So we would appreciate some attention to this.
)
l 19 DR. CATTON: I didn't see anything about 20 condensation.
l i
l 21 MR. BECKNER: There may not be. I am trying to i
i 22 think.
l 23 DR. CATTON: It seems to me that there ought to f
l 24 be one of the sections in Chapter 6 called out specifically
() 25 to address that.
l ace FEDERAI. REPORTERS, INC.
202a47a700 Nationwide coverage axw3w/M6
]
7280'13 16 450 a Vbur 1 MR. BECKNER: A major shortcoming of this whole 2 document that is missing so far.
3 DR. TIEN: I think the document is very good.
4 However, I think particularly it is a good document you want ,
5 to make it improved, perhaps to a point in which it is 6 becoming a little asymptotic. Perhaps one month's or two I
j 7 months' time will buy you a lot of improvement at this i
- 8 stage.
J' 9 I read more on Chapter 6 because that is perhaps 10 local ECC phenomena. One problem I find is consistency.
- 11 You have dif ferent people writing dif ferent things, and i
12 particularly PWR and BWR, and some of them are common i
O 13 one o eo det deceeee taev ere eteteo av aterereet aro oe, 14 a totally dif ferent kind of flavor, and it looks like quite i
i 15 different phenomena.
i 16 I think it needs a lot of improvements, touching i
i 17 up, and so on.
18 To give you a specific example, Section 6.3, PWR
- 19 ECCS bypass. I am pretty sure I read that, and that is from 20 two organizations that have major input, Battelle and also 21 Creari. It is written in such a way it looks, you know, j 22 like just from their quarterly report and audit report, a l
- 23 lot of work in that area is now coded, a lot of it has been i
i 24 publishad in the literature.
)
i O 2s to rect- t oereooettv ateeoree "o eoa t eatex
.i
- ACE FEDERAL Runonreas, INC.
, 202 347 3700 Nationwide Coserage Nx)-336W46
]
7280 13 16 451 the technical community disagrees with perhaps what Creari's
()Nbur 1 2 interpretation is on ECC bypass, or some of them.
3 Then look at 6.5, BWR refill. That comes from 4 General Electric., and their interpretation is quite 5 different. I consulted with them a lot, and they reflect 6 more my view.
7 So I think there is a real inconsistency about 8 some of the common phenomena. I find this actually 9 scattered around Chapter 6 a lot.
10 MR. BECKNER: You say those sections are not 11 written by Crearl or General Electric but they are written 12 by the Staff?
() 13 DR. TIEN: Totally influenced by those two 14 organizations.
15 MR. BECKNER: No, they are written by the Staff 16 and myself.
17 DR. TIEN: Maybe I road them dif ferently. If I 18 tako a look at 6.3 I can tear that into pieces. Many 19 interpretations are just not right because thoso codes are 20 from Croari and Battollo's quarterly reports.
21 MR. BECKNER: I disagroo, but we can have the 22 technical argument lator.
l 23 DR. TIEN: Suro, we can discuss that.
24 MR. BECKNER: That was taken from a NUREG which I
() 25 road myself and which Ross Anderson wroto, basically.
Acit.Friotinai. Iliti>oniitas. INC.
I 202-W 37m Nanonwide Cmcrave W3MIMi
7280 13 16 452 1 DR. TIEN: I am familiar with this field, and I
( )Wbur 2 have worked on this a lot myself, and I find that that is 3 just not right.
4 Now, I found the comments of Harold's on 5 radiativo modos -- I soo very, very little, also, and other 6 condensation models are not touched at all. I think that 7 noods to be improved.
8 DR. SULLIVAN: Bill, if you turn to page 2.3, I 9 think this will demonstrate that you nood an editor. You 10 should hire somebody to edit this thing.
11 I hope you have a figure on page 2.3, your page 12 2.3. I didn't soo one.
() 13 First of all, it gives a nomenclature. Then it 14 says WEST, and if you go back into the text, which is on 15 2.2, in the first paragraph they reference this figuro.
16 Well, it is referenced on 2.2 and if you are not really into 17 it, you can't figure out what this thing means or even who 18 ran WEST until you go back to the reforence.
19 So one of the things is if you look at that 20 curvo, the thing that is NRC is approaching infinity on a 21 rather fast rato, okay? Thoro is no explanation of why, why 22 it was ovon stopped or, you know, did ovorybody just got 23 tirod?
24 And look at -- I quons it is the Exxon. It just
() 25 stops, too, and thoro is no explanation of it. Yet this Act 17itolinal Itt:Pou ri:Rs, INC.
g 202.nwm woonwide cmcrage sunn ua6
7280 13 16 453 (O,Nbur 1 figure is one of the ones that you rely on heavily in saying 2 that there is some data in the codes.
3 MR. BECKNER: I am missing your point, Harold.
4 DR. SULLIVAN: I am saying this is one of the 5 things you are relying on heavily, showing that there is 6 conservatism, and those things are just not explained. I 7 really believe that a figure that you put in there, you 8 ought to try to explain why there is that much conservatism 9 in it and all the problems that went into doing those 10 calculations and why this one stops and why Exxon didn' t run 11 theirs further.
12 MR. BECKNER: I am not sure. I can tell you
() 13 about NRC.
14 MR. LAUBEN: We don't know about the Exxon one.
15 There are certainly extremo conservatisms in 16 there which we could discuss.
17 MR. BECKNER: I am not sure, liarold, the point 18 you are showing.
19 DR. SULLIVAN: I am saying that without a really 20 good understanding of what that figure really means I would 21 be questioning whether you ought to put it in hero or not.
22 MR. BECKNER: The point of that figure was to 23 show with Appendix K you are up high. It might be useful to 24 try to explain why they are all over the map. I don't know
() 25 whether we can reconstruct that or not.
Acn.17EnnnAi. ItneoRTEns. INC.
202.m noo Nanonwide cmcrage n-3AtM6
i l 7280 13 16 454 DR. SULLIVAN:
f()Vbur 1 These were not advanced codes j 2 either.
i l 3 MR. BECKNER: I don't think it shows that. It i
4 says using conservative assumptions.
l 5 DR. SULLIVAN: Maybe a nice thing to do is to say I
j 6 here is all the conservatism, put an advanced code in there, j 7 and show, you know, that it does predict the data fairly i 8 well because that is what you are trying to get at, to me, 9 is to say that the old methods were very, very conservative l 10 and try to explain why to some degree and then show that the i
i 11 new codos fit the data really well.
i 12 MR. BECKNER: We will talk about it. To me, that i
() 13 illustratos the point I am trying to make. This is the only j 14 problem that we have, and that is a terrible figure. I i 15 think that illustratos the point.
i 4 16 DR. CATTON: It illustrates, also, that you i
17 probably can't calculate it.
)!
j 18 MR. BECKNER: This is not current work. This is 19 historical stuff.
i
]' 20 MR. LAUBENs Also, let me say something about I 21 that.
I i 22 Those woro calculations. Each person who l
l 23 calculated this did not use a uniform set of conservatisms i
24 for the calculation, partly because LOFT really is not a PWR
() 25 and there are some featuros about LOFT, including, as we all Ace.IT EnunAL REPonTEns, INC.
202 447 4 700 Nationwide coserage 8mnwM6 f
7280 13 16 455 (hVbur 1 know, the core design and the size of the core, which make 2 it not exactly amenable to the kinds of things that Appendix 3 K calculations were devised for.
4 So I think it makes even a greater variation than 5 the kinds of calculations you can get. Maybe that needs to 6 be said, but we need to talk about it, I am sure.
7 DR. SULLIVAN: If you look at this figure, the 8 NRC calculation looks the worst, but the Westinghouse, and 9 particularly the Exxon, they obviously did something 10 dif feren t.
11 MR. LAUBEN: Yes, they did. I can say for the 12 record that we know that there are prescriptions about
() 13 densification, in particular, that the NRC tried to use very 14 specifically in their calculation, which really wouldn't
{
15 apply to present day fuel and present day large PWRs. If i
l 16 you applied that literally to the fuel that Exxon gives for i
j 17 LOPT, it is going to give you high temperatures like that 9
- 18 overy time, and that is what happened.
19 MR. BECKNER: I guess, Harold, I don't understand 20 the problem because this figure is not attempting to r
i 1 21 illustrate where we are now, just the opposite. It is l 22 trying to illustrate the problems that we have back-using 71 23 features.
l 24 DR. SULLIVAN: But, Bill, the problem is that the l
l
() 25 figuro -- you are trying to demonstrate that, but there is a ace. FEDERAL RevonTEns, INC.
- ! 202-347-3700 Nationwide roserage 800-336-6646
7280 13 16 456
( )Nbur 1 big, big difference between all those calculations.
2 MR. BECKNER: Why is that bad?
3 MR. LAUBEN: Harold, part of that is the way you 4 apply Appendix K to a nonpower reactor such as LOFT. That 5 is part of the reason there is a big disparity.
6 DR. SULLIVAN: And that is what I am trying to 7 get you guys to put in there.
8 (Pause.)
9 MR. MICHELSON: Any other comments?
10 (No response.)
11 MR. MICHELSON: I think what I suggest we do is 12 go ahead to the uncertainty analysis portion because it
() 13 takes longer. At that time, if there is any time left, if 14 you want to come back for any more comments at 4:00 o' clock, 15 we can.
16 (Slide.)
17 MR. ZUBER: I am Novak Zuber, from NRC Research.
18 I would like to discuss today a section of Chapter 4 which 19 deals with the method of determining the capability of the 20 code to scale, the capability of the code in a particular 21 scenario and the uncertainty analysis.
22 (Slide.)
23 NRC held three meetings with the national 24 laboratories. These meetings were also attended by people
() 25 from the industry and consultants. One meeting was held in ace FEDERAL. REPORTERS, INC.
202-347 3700 Nationwide Coserage 800 336W46
7280 13 16 457 1 October '86 to discuss a method which would provide us with
[ )Vbur 2 the capability to address these three aspects -- scaling, 3 applicability, and uncertainty.
4 At the meeting held in January we applied this 5 method to large break LOCA and identified how we could :
6 determine the capability of the code to scale. We held a 7 third meeting on March 26 and 27, and we developed a method 8 to determine the uncertainty of the codes to calculate PCT.
9 We did apply it to a large break LOCA for two 10 reasons:
11 One, we had the largest amount of data and 12 calculations for that scenario. We wanted to make it
() 13 relevant to the Appendix K discussions, and this is the 14 problem which we are most familiar with.
15 The purpose of my presentation here today is to 16 discuss the results of this third meeting, but in order to 17 put it in perspective and to give you the rationale of how 18 we arrived at these criteria and this method, I would like i 19 to discuss a little bit of background material and what the 20 problems are when we want to address this phenomenon and 21 answer these questions.
22 So I will go a little bit further in-depth, and 23 then I will discuss the results of these previous meetings 24 to give you a better perspective.
+
() 25 (slide.)
i I
/\CE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800 336-6646
7280 13 16 458
('iVbur
-V 1 There are three questions we have to answer when 2 we want to apply a code to a nuclear reactor. ,
l 3 First, does the code have the capability to scale 4 up phenomena from a test facility to a full size?
5 Two, does the code have the capability to address 6 a particular scenario or a family of scenarios?
7 And, three, what is the uncertainty with which 8 the code calculates a particular parameter -- let's say the 9 PCT for a full nuclear power plant?
10 These are the relevant questions we have to ask 11 whenever we apply a code to a nuclear phenomenon, whether it 12 is our code or the vendor's code.
() 13 (Slide.)
14 Let me address the problem of scaling because it 15 was brought 15 years ago in the study of the Physical 16 Society as one of the important issues, and I would like to 17 discuss it because it bears on the discussion and the method 18 we developed.
l 19 Let me say the method was developed with the 20 national laboratories, with the consultants. We drew very 21 heavily on the experience which CSNI had in Europe during 22 the last two years.
l 23 So if you really want to look at this as a 24 consensus meeting of the technical community here and l
() 25 abroad, 13 years ago when we started this business in 1974, ACE FEDERAL REPORTERS, INC.
i 202-347-3700 Nationwide Cmcrage 800-334 6646
_ . - - --~ . . _ _. _ . _ - -
7280 13 16 459
)Nbur 1 there was a naive perception to believe that the code will 2 solve this problem of scaling. People said the code will do 3 the job.
4 The question which was not really asked, which is 5 a relevant question is: where is the capability of the code 6 to address this problem?
7 This was not really brought up until probably two 8 years ago, and I would like to show you where the problem 9 lies.
10 The scaling is really the central problem in our 11 method of analysis and experiments for two reasons. It 12 affects the design of the test facility. It specifies the
() 13 test conditions, under which conditions you are going to run 14 the tests, and also affects directly the code assessment 15 process.
16 What is the effect of scale distortion?
17 Usually, we learn in school that we can scale 18 something, but if we have any parameters, any similarity 19 between parameters, we cannot satisfy all of them at the 20 same time.
21 So when we design the f acility, some parameters 22 are satisfied and we have good scaling. Some parameters are 23 not satisfied and we have distortions.
24 In this particular program, we have so many
() 25 parameters that we design the facility according to some ACE FEDERAL REPORTERS, INC.
2J2-347-3700 Nationwide Coserage 800-336-6646
- 7280 13 16 460 1 criteria, and the question then goes: what are the effects
(}Nbur
- 2 of distortions?
3 And the distortions can have effects on test-4 results because they may affect the phenomena and the 4
i 5 magnitude of these phenomena and timing. They may affect 6 code assessments because the code is used to assess the code 7 against data from the. facility which may have distortions.
i 8 So you may get good agreement with the facility, 9 but it doesn't mean that you will have a good agreement if 10 you apply it to full scale. One cannot use a code and just 11 run it and say it applies to a full test facility.
12 This also particularly true -- and I will go back k () 13 to this -- this is a fact of life. Our codes have been 14 tuned. We have many coef ficients in the codes, and often 15 these coef ficients are adjusted to a set of parameters, to a l 16 set of experiments, and if these parameters come from ,
17 distorted facilities, it would be better than this item.
i 18 The question is what is the effect when you scale
- 19 this to full plant? This has to be addressed.
20 There is also the problem of nodalization. There l 21 is another problem. If the facility doesn't take part in the 22 code phenomenon, if it is scaled out, we cannot really 23 assess the code.
- l. 24 (Slide.)
() 25 This is really a diagram. It shows really our
/\CE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6616
. _ _ _ _ . _ _ _ - . . _ . - ~ _ . . . . _ _ . - _ . . - - - . . , . . . _ _ , _ , _ _ _ _ - _ . . . - . , . - . - _ _ . . _ - . . . . . . . - . - _ _ - , - - - . _ - .
i 4
7280 13 16 461
().Vbur 1 research and development approach, which we have been using 2 for the last 14, 15 years.
3 We have a nuclear power plant. We want to design 4 an experiment. So you had a discussion for a day and a half 5 of how we design the facilities. We developed scaling 6 criteria. We used the scaling criteria to design the 7 facility and run tests. We specified according to some 8 group of test conditions. Then we run separate effects 9 tests and integral ef fects tests.
10 The separate effects tests are used to provide 11 data and develop models, which are used then for assessing 12 the codes. There, indeed, we often set parameters in the
() 13 code through the code base. We do the same thing in this 14 part of the room.
15 So if the facility is distorted, it will affect 16 our codes, if somebody doesn't take care during the 17 development. Then we test the code and apply it to the full 18 plant.
I 19 So as I said, scaling is an integral part of the 20 experiments and of code development.
21 ,
(Slide.)
22 Most of our facilities have been designed 23 according to the power to volume scaling. I think it is a 24 very, very powerful method. It was designed essentially for
() 25 addressing the large break LOCA. It is a very powerful tool ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage WO-336-6M6
- - - - - _ _ _ _ ~ , .- ____ _. - . _ _ _ , , _., , . , _ _ - . - _ . . _ _ . _ , _ _ _ _ _ _ , _ . _ _ _ __ -
i 7280 13 16 462 because using this method, you preserve time, you preserve
()Vbur 1 2 fluid distribution and energy distribution, velocity 3 acceleration and length.
4 So you satisfy many things. You satisfy the 1
i 5 power to volume. You satisfy velocity and timing. So you 6 really have a good understanding of the facility and of the 7 phenomena that may occur in the full plant.
8 So it is a very, very powerful method. You 9 preserve volume and length, but you distort areas.
10 Therefore, there are components in the system, in the 11 reactor, which are completely distorted.
12 You heard the discussion yesterday and today.
A 13 You have a distorted downcomer. Your plena are distorted.
, (,j 14 Your structural heat transfer is distorted and heat losses 15 are distorted.
d 16 This has to be taken into account both in the 17 experiments and in the code analysis. As a consequence of l
[ 18 these pluses and these minuses, some part of the transients l 19 are well-scaled. Other parts of the transients are af fected 20 by distortions. This, then, you have to take into account 21 if you wan t to apply a code to a full-sized reactor.
22 (Slide.)
23 Let me talk a little bit about the code because l
, 24 this is the topic of my presentation. Let me go back in l
() 25 history.
l
/\CE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6M6
7280 13 16 463 Ig )Nbur 1 In 1975, we had a family of codes, RELAP-3 and 2 RELAP-4, three field equations, and I think these parameters 3 I have taken from a British report. I don' t guarantee that 4 there were 21, but there may be 21 or 19. I don't know.
5 RELAP-4 I think had 21 parameters.
6 The new generation of codes which we developed 7 since 1975 and which we call best estimate, the TRAC family 8 and the RELAP-5 family, they have unequal velocity, six to
, 9 seven field equations compared to three and, according to 10 Ian Britten, 175 parameters.
11 We then have the capability to model many 12 phenomena, many processes, but we never seem to use many
() 13 parameters which we had to evaluate.
14 Let me talk about the characteristics of our 1
15 codes. This influences the method to determine the scaling 16 capability and the uncertainty analysis. There are really 17 six characteristics.
18 First, all separate integral facilities have 19 distortion s . We have to take this into account.
- 20 Two, our best estimate codes have a large number 21 of closure equations, correlations, and very often these 22 correlations have been shown through experiments.
23 The question then arises can this then be scaled 24 to a full-sized reactor. The consequence of Item 2 is the
() 25 ef fect of compensating errors. Once you start turning ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6 4 6
7280 13 16 464 1 coefficients, one of two or several, you may introduce b'~hVbur 2 compensating errors. Some people don' t want to discuss it.
3 It is a fact of life, and we as a regulatory agency should 4 be cognizant of the shortcomings so we can either address 5 the problem or resolve it. We cannot hide our heads in the 6 sand and not look at the problem.
7 We have compensating errors in the code. In the 8 compendium there is a section which discusses an effect.
9 For example, if you have compensating errors between heat 10 transfer coefficients and void fractions, for some 11 transients the code will be very good but this is not a 12 grave error. For some transients it may be a very bad p) q 13 error.
14 Therefore, we have to understand which conditions 15 we can use the code and when we cannot use,it. That is an 16 important item.
17 (Slide.)
18 Codes have been compared to a large number of 19 data and experiments, and by and large we obtained 20 reasonable agreement. There are some instances where the 21 agreement is not that good, but it is important for us as a 22 regulatory agency to understand why the shortcoming, why the
! 23 code under some conditions fail, the reasons being that we 24 have to apply our codes to a large number of transients, l
() 25 different transients, different accident scenarios.
ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6646
7280 13 16 465
(,)Nbur 1 We have to understand why a code failed under 2 given conditions so that we may expect when it may fail 3 under some different conditions when we apply it to a 4 reactor. So this cannot be overlooked.
5 Also, our best estimate codes are sensitive to 6 nodalization. This has to be also addressed.
7 Item 6, which is very important -- they did go on 8 recently, maybe during the last two or three years -- the 9 correlations in our codes, the closure equations of our 10 codes either don' t cover the range of interest of 11 applicability, either in terms of fluid properties or 12 geometries or they don't have any experimental data base.
/~T
(_) 13 We again, as a regulatory agency, we must know 14 what are the conditions where we have these correlations.
15 We have to know this because we have to know what is the 16 effect of these shortcomings when we apply this code to a 17 full-sized reactor. As a regulatory agency, this is a 18 number one requirement.
19 DR. TIEN: Novak, I have a general comment. I am 20 very pleased to see your emphasis on the scale distortion.
21 However, I see there is also a fine difference between, you 22 know, the Kerman studies, where you always got to a point 23 where certain scaling laws and we say how successful that 24 scaling law will apply to certain separate facts or O
()
25 individual phenomena.
ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage MX)-336-6M6
l
-7280 13 16 466 1 If we ask ourselves where would be the biggest
(])Vbur 2 distortion for certain scaling laws, seek those phenomena.
3 MR. ZUBER: Absolutely.
4 DR.'TIEN: I think there is a very fine 5 difference there. We haven't really emphasized that much.
6 This is the first time I have seen the emphasis on the scale
- 7 distortion.
8 MR. ZUBER: Okay. Okay, I am glad you brought it 9 up. This is exactly what our method -- I said "our method,"
10 the best case method, the one we developed as a group --
11 addresses. It addresses what is well-scaled and what is not 12 well-scaled.
13 Then you can say what transient or what scenario
_ ()
14 cannot be addressed with confidence when you apply it to a 15 full-scale reactor. This is what we are addressing in maybe l 16 five or six more slides.
17 But there are two ways to look at this problem.
18 One is to say I shall use a statistical analysis, I shall 19 run 100 sensitivity analyses. When you ask how is the 20 scaling addressed, it is usually arm-waving.
21 The other way is to look at the physics of the l 22 problem and to see what the codes do. Where is the l 23 capability to scale and what can we calculate with ,
24 confidence?
I
() 25 This is where engineering thinking starts. The i
/\CE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cmerage 83)-336-6646
7280 13 16 467
(,)Nbur 1 other one is really a code that can be satisfied with number 2 crunching.
3 (Slide.)
4 Now, when we discuss codes to see what is the 5 capability of the code, the capability of the code rests in 6 four items -- in the field equations, which provide the 7 general capability for this process; the closure equations -
8 - and this is the item that was addressed. It is provided 9 with the closure equations which provide the capability to 10 address the particular processes or the particular 11 phenomena.
12 This is to a great extent where the scaling
() 13 capability of the code rests. If this capability is not 14 here, the code cannot scale, period. You cannot really 15 nodalize something and resolve the scaling problem if the
- 16 scaling is not embedded in these correlations.
17 A third element is the numerics, and the fourth i
18 element is the nodalization and structure of the code.
4 19 DR. CATTON: Two, three and four. The closure 20 equations sometimes are adjusted to compensate for the 21 nodalization.
22 MR. ZUBER: No, no. You are absolutely right.
23 There are two questions to distinguish. First, when you
. 24 take a correlation, what is the data base of these
() 25 correlations? Is it broad enough or not? Can it be applied ACE FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coserage 800-336-6646
7280 13 16 468 1 to a full-scale reactor or not?
(}Nbur 2 This you can address. Suppose it is an excellent 3 correlation. Then the question is how is it implemented 'n 4 the codos?
5 This also has to be addressed, and when you 6 address this, finally you have to look at the combination of 7 these two. This is the reason you have to see what is in 8 the code and how it is listed. It is not enough to say that 9 this is the correlation if the correlation is really 10 modified in a form that doesn't provide this capability.
11 (Slide.)
12 Let me address the impact of closure relations;
() 13 first, the capability of the closure equations with the 14 closure capability to address the phenomena.
15 We have to determine whether the code can 16 calculate a process in the scenario. We have to see whether 17 there are equations which account for this phenomena and 18 what is the scale-up capability of these correlations. We 19 have to also identify which of these correlations or closure
- 20 equations have been tuned to a certain amount of data. This 21 is what we want to apply to scaling.
22 We have to determine what is the ef fect of tuning l 23 on scale distortion and distortion of the facility when we 24 apply it to full-scale plants. We have to assess the effect
() 25 of tuning and determine in particular whether there are l
l ACEJFEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage m)-336-6M6 l -
7280 13 16 469 em
( ,lNbur 1 compensating errors in the code.
2 (Slide.)
3 Now, what do we need when we want to address 4 these questions?
5 We need really complete code documentation. If a 6 code is not fully documented, there is no basis to pass any 7 judgment on the code, either by peer review group or by 8 regular correlations.
9 What you need is a code manual. You need a model 10 and correlations assurance document. I shall discuss this 11 in a while. We need user guidance, and we need assessment 12 reports.
r~m
!s _) 13 Once this is available, we can then perform the 14 code scaling, the applicability and uncertainty analysis and 15 provide a fif th document which will tell us this code is 16 applicable to this scenario because this scenario gives the 17 biggest uncertainty, 18 With these five documents, these four and that 19 one, I think we have a complete document of code capability 20 to address a scenario or a group of scenarios.
21 MR. MICHELSON: Is that fifth document you 22 referred to essentially going to accompany an application to 23 using a particular evaluation model?
24 MR. ZUBER: This would document the results of f'
\ 25 the analysis we are undertaking now, and the hope would be ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6M6
7280 13 16 470 1 to finish by the end of this fiscal year. This will be for
()eVbur 2 a particular code to address a large break LOCA.
3 MR. MICHELSON: And that document will be outside 4 the compendium of yet another document?
5 MR. ZUBER: It will be a document which will be 6 part of a TRAC BF-1 MOD-1 code.
7 MR. MICHELSON: Is that document in any way part 8 of the regulatory process?
9 MR. LAUBEN: In the same way as the compendium 10 is, it would be something available to reviewers in'NRR to 11 use for their information so they could pick up some 12 insights on how they might review a particular submittal and
() 13 service submission by the utility or vendor.
14 MR. MICHELSON: Now, is this document going to be 15 reviewed at all prior to that point?
16 MR. LAUBEN: Reviewed by NRR, for instance?
17 MR. MICHELSON: For instance.
18 MR. LAUBEN: We would certainly hope so.
19 MR. MICHELSON: When it has been reviewed by NRR, 20 why isn't it then a recognized document and you would have 21 to review further when you apply for approval of a 22 particular evaluation model?
23 MR. LAUBEN: I think you would have to talk to 24 someone in NRR about that, how they would want to use this.
() 25 MR. WARD: Just think back.
ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6M6
7280 13 16 471 i
t'TVbur 1 (Laughter.)
V 2 MR. LAUBEN: Reviewers in NRR would like 3 sufficient freedom to in terpre t , especially at this early 4 stage in the game, sufficient freedom to interpret how they 5 would want to review a particular submittal.
6 MR. MICHELSON: I think what I am hearing is that 7 really there are no preapproved documents. You have 8 submitted an evaluation model that you are going to use, and 9 the Staff says, yes, no, maybe or whatever.
10 MR. LAUBEN: As I see it, the real value in this 11 for everybody and as I would see it for NRR is that we have 12 been going through this process not arriving at the end of
() 13 the line. We have been going through this process. We 14 would uncover things that one should pay attention to and be 15 concerned with.
16 I think without going through the process of 17 doing it, that becomes somewhat difficult. So I think the 18 process of going through this is what I consider to be of j 19 greatest value for our reviewer.
! 20 MR. MICHELSON: How many of these documents do l
! 21 you expect to see by the end of this year?
22 MR. ZUBER: We should have one for all codes.
i 23 MR. MICHELSON: You will see it for the presently 24 established codes?
(} 25 MR. ZUBER: Oh, you have to do this for each l /\CEJFEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coverage 800-336-6M6
7280 13 16 472
()Vbur 1 frozen released code. If you freeze the code, you perform 2 these four documents, you perform this evaluation, you 3 publish a fif th document. If you want to apply to the large 4 break LOCA for this uncertainty, then you have another code, 5 you have another statement.
6 (Slide.)
7 Let me talk a little bit about these quality 8 assurance documents, and this was the document which was 9 mentioned in the instructions, the quality assurance 10 documents that we have to address.
11 The quality assurance documents has really three 12 objectives. One is to provide detailed information on J
() 13 closure equations; that is, on the correlations which are in 14 the code. Then it is to describe how these closure 15 equations are coded, listed in the manual in a very nice 16 way. The point is how they are coded in the program and to 17 say what is in the code manual that is acceptable and to 18 provide a technical rationale. Why is this correlation, why 19 is this model selected to give a rational basis why it is 20 used?
21 These are the main objectives of these documents.
22 (Slide.)
23 Now, what it has to do for each correlation, each 24 closure equation, this document must provide the original
() 25 source, give the reference of the data, if it is published ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 8W 336-6616
7280 13 16 473
()Vbur 1 in the journal. What is the data base? What is the 2 accuracy of the correlation, and is it applicable to nuclear 3 power plant conditions?
4 That means whether the fluid conditions and 5 geometry really are applicable. Provide an assessment of 6 effects if it is used-beyond the range. Describe how it is 7 implemented in the code, any modification which has been 8 made in order to accommodate the numerics, and provide the 9 effects of some of-these distortions on the changes that we 10 made.
11 Such a document was requested by a letter from 12 the program manager in October of '85 to LASL -- in March, I l ) 13 believe, but first the request was made. Essentially, these
- 14 requirements were sent to LASL by the program manager in 15 October of '85.
16 A second request was sent by the Division 17 Director in March of '86. At that time NRC allocated f
18 $70,000 to have this document written and published by 19 October of '85.
20 The document was delivered for the code TRAC BF-1 21 MOD-1 in December of '86. The document was reviewed by a
.22 group peer review consisting of, I think, 11 members in this 23 country and abroad. The comments were very, very critical.
24 My commen t would be it is a shoddy, very poor
() 25 document. It doesn't do credit to the standards of a great i
i
/\CEJFEDERAL REPORTERS, INC.
j 202-347-3700 Nationwide Coserage 800-336-6M6
7280 13 16 474 1 national laboratory like Los Alamos. It doesn't do credit
(}Vbur 2 to the amount of money that we have spent on the code, and 3 it is not a document which a regulatory agency can use for 4 regulations.
5 As a consequence, and which really we appraised 6 about a month and a half ago, NRC management held a big 7 meeting the last two days here at INEL with the management 8 from INEL, Sandia, and Los Alamos. Before this meeting, we 9 requested Los Alamos to give us a work schedule and costs to 10 improve and bring this document to the standards which a 11 regulatory agency requires.
i 12 We were given yesterday the cost -- that the new
() 13 document will cost $450,000 and that it will be delivered to 14 NRC in February of '88. We shall make some decision next 15 week, but as to the best of my knowledge, what upper
- 16 management of NRC has decided is that this document will be 17 requested to be delivered to NRC by December of this year.
l l
i 18 We cannot wait till March or February of next year, and this l
19 document will be reviewed by the same group of people who 20 reviewed the previous document, and it shall meet the 21 standards needed by a regulatory agency.
22 In order to save money for the future, these were 23 the comments made by Dr. Ross and Dr. Sheron, who were at 24 the meeting, and then Dr. Sheron and Dr. Ross brought up l
() 25 something very important, that we won' t do this in the 1
ACEJFEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cmerage m)-336-6M6
--, -__ .___._ . _ .__ ._ ,_ _ ~ , _ _ _ _ _ . . . _ . . _ , _ . . . _ . .
7280 13 16 475
(%Vbur 1 future, that we shall make a request for more code 4
2 developers in the future to meet the standards which NRC 3 will require from our code developers. That will be the i
4 standard which NRC requires today from the vendors.
5 In addition to this, we are a regulatory agency.
6 We shall set higher standards, and a commitment was made 7 that by the end of the summer NRC will inform the 8 laboratories, give them additional specifications of the 9 requirements which our laboratories have to meet, and 10 Bosworth, the Chairman, gave out at this meeting as follows:
11 That the code has to be traceable and verifiable 12 by independent reviewers. There should be a history of
() 13 documented changes the same way as with everything in the 14 vendors.
, 15 We have to have the same requiremenit of our code 16 developers as we require from the vendors. This is the 17 minimum requirement from a regulatory agency. Otherwise, we 18 are not doing our business.
, 19 (Slide.)
1 20 I won't go into a discussion here because we have 21 also changed as a consequence of the change of this Appendix I
22 K. We are also increasing our demand on our code 23 assessments.
24 I will not go to these bullets. It is in a i
() 25 handout. It is in the compendium. So you can read it, and 4
ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 80tL3346M6
7280 13 16 476
[ )Nbur 1 these are the additional requirements we are asking code 2 assessors to put in the code assessment reports.
3 The emphasis is on detecting the compensating 4 errors. We must know under which conditions the code will 5 f ail and to find where are the compensating errors, so that 6 we may anticipate where we can and cannot apply the code.
7 Let me give you a good example which came to the 8 attention of NRC just last week.
9 Two years ago we asked our laboratories to 10 perform some calculations you are all familiar with. One 11 was done at Calvert Cliffs and the other at H. B. Robinson.
12 And we did perform these calculations, RELAP-5 MOD-1, I
(),
13 believe, and the other with TRAC.
14 It was during the decry heat we had flow 15 stagnation in both reactors. We did the calculations with 16 both codes.
- 17 The calculations of RELAP-5, which were done on i
i 18 H. B. Rob in son , resumed circulation after some time. So 19 there were interruptions, flow stagnation, then t
! 20 recirculation, and so on.
l 21 The calculations which were made by TRAC said i
{ 22 once the stagnation started flow was never reestablished. I l 23 don't think it was in this body, august body of saviors or l
l 24 gurus -- really, I am glad that you brought this question --
(
l
() 25 that this question was raised:
l l ACE FEDERAL REPORTERS, INC.
! 202-347-3700 Nationwide Coserage 804336-6M6 l
7280 13 16 477 What was the reason for this discrepancy, for (pVbur 1 2 this really dif ference -- same conditions but dif feren t 3 codes? What was the difference?
4 Prompted by this stimulus, NRC gave a contract to 5 an independent company to see what the problem was and 6 assess it. We were informed last week the error comes from 7 a very poor heat transfer correlation for con'densation.
8 Professor Tien, if you want to know what is the 9 importance of condensation, because you brought it up -- it 10 is a good question -- the reason is that TRAC really never 11 established flow circulation. This is what I was told by 12 phone. There are going tomorrow to have a program review of
() 13 these results. So I may be a little bit wrong, but this is 14 the information we had over the phone, that the heat 15 transferscoefficient for condensation was very poor.
16 Now, let me say this. It doesn't mean that TRAC I 17 BF-1 MOD-1 code is a bad code. It is a good code, but we
! 18 have to know when we want to apply it. We can say it is 19 good for large break LOCA. If we look at the phenomena, it 20 is modified, we can apply it with confidence.
i l 21 But we also have to know under which conditions l
22 this code may fail. A good example was just exactly this l
l 23 case here. We should have anticipated the heat transfer l
24 coefficient is poor in condensation. We made progress
() 25 monitoring the steam generators.
ACE FEDERAL REPORTERS, INC.
! 202-347-3700 Nationwide Coserage 800-336-6646 l
____ _ _-~ . _ _ _ _ _ _ - _ _ _ . - . - - - - . _ -
7280 13 16 478 1 The next question is to ask for what scenario
( JVbur 2 this will be important, and what I present now is a method.
3 This method -- we are trying to address this process to 4 identify the conditions when- the code is good and conditions 5 when the code is poor so that we can really use this code as 6 a powerful tool.
7 Our codes -- I don' t want to be critical -- they I
8 are very good, much better than we had several years ago.
9 They need improvement in some cases, but we have to document t
10 what is the capability so that we can use them with 11 confidence.
12 DR. CATTON: Who did the work?
() 13 MR. ZUBER: SAI and Professor Theofanous.
- 14 DR. CATTON
- Is that Ken Williams?
15 MR. ZUBER: Ken Williams.
16 DR. CATTON: Did he do some of the original
- 17 calculations, too?
18 MR. ZUBER: I really don' t know, honest. We are 19 going to have a program review the day after tomorrow to see 20 exactly what was the effect.
, 21 I was also told that once the heat transfer l
22 coef ficient was changed flow was restored. There was no way i 23 for a program manager or any technical man in NRC to detect 24 this error unless we had a good code review, and this is the
() 25 role of a QA document, to know what are the strong points of ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6M6
-. -, , _ _ _ . _ _ __ . _ - . ~ .
i ..
7280 13 16 479 1 a code and what are the weak.
(}Vbur 2 Now, as I said, I will not go and discuss the
! 3 requirements of the assessment reports.
4 (Slide.)
5 Let me now address the problem of scaling and the i t
6 - methods which have been developed with our consultants and i
7 the national laboratories.
8 If you want to assess the code and the capability 9 of the code, there are really two ways to do it. One is a l.'
10 bottom-up approach, where you examine each component 11 separately, then assess the effect of this particular 12 correlation or model on the calculated results.
l
() 13 Essentially, you perform a sensitivity analysis.
j 14 The other way to do it is the top-down approach,
- 15 where you first identify important macroscopic processes,
- 16 then identify correlations which model these processes and i 17 analyze that ef fect.
l 18 Obviously, the second approach has advantages 3
19 because it reduces the number of calculations.
20 (Slide.)
21 In the bottom-up approach we examine each 22 parameter -- and there is a report from Sandia -- and if you 23 have important parameters, you get to perform 2 times N plus 24 1-squared calculations to complete the sensitivity study.
t
(} 25 Our codes have approximately maybe a hundred l
/\CE. FEDERAL REPORTERS, INC.
- l. 202 347 3700 Nationwide Coserage M10-336-6M6
[
7280 13 16 480 coefficients. There is no way we can perform this (JVbur 1 2 sensitivity analysis on each parameter. We are forced to 4
3 use an approach which is a top-down approach.
i 4 The requirements here are then to identify what 5 are the important phenomena, then perturb them and analyze 6 their effect.
7 (Slide.)
1 8 What we also need is to be able to address the i 9 problem of scaling, and what we are really trying to do in 10 this method, which was developed in three meetings, is to 11 take the advantages of both approaches. We shall do a 12 detailed analysis of the phenomena which are important.
() 13 This is the bottom-up approach. And we shall
! 14 only do this on the bottom phenomena.
1 i
15 This is the top-down approach. We shall take the 16 process of both and try to put it in to a synthesis.
i 17 (Slide.)
18 The method, the CSAU -- this is an acronym --
l 19 contains four elements or four stages, four important l
20 features.
21 Ic first is based and identifies all important 22 phenomena. It addresses only important phenomena. It 23 requires that we have to analyze for a scenario or a group 24 of scenarios what are the important processes.
() 25 Then the second feature is that it uses separate i
ACE FEDERAL REPORTERS, INC.
- I 202-347-3700 Nationwide Coserage 800-336-6M6
,L 1,
L 7280 13 16 481
()7 bur 1 effects tests for two purposes. One is to address the 1
2 capability of the code to model a process or a scenario.
3 The second thing is to use separate effects to assess the 4 capability of the code to scale.
5 Since we have now data from UPTF, which is a 6 large, full-scale facility, we shall be able to assess code 4
7 scaling capability up to full scale, and these can be used
, 8 for separate effects tests.
2 9 Number three, there was always a question in the 10 community how people can extrapolate data from a test
! 11 facility to a full scale. That was always a controversy 12 because some of these facilities have been scaled, and the
() 13 argument was you cannot scale it to full scale.
i 14 The approach we have come up with is a little bit 15 different. We pose the question, what are the phenomena, i
l 16 what are the parts of the transient which are scaled by the i
17 facility.
18 Our facility, I would like to remind, was really 19 designed to preserve and was scaled for the first stage of a
] 20 large break LOCA -- let me discuss this in a moment -- the 21 reason being that the pressure, power, fluid conditions were i 22 scaled. The distortions which are in these facilities are i
i 23 in the plenum and in the downcomer. If the phenomena in the j 24 plenum and downcomer are not important during that part of t
() 25 the transient, therefore the transient is influenced by the ACE FEDERAL REPORTERS, INC.
202-347-M00 Nationwide Coverage 800-336-6646
7280 13 16 482
( Ybur 1 components which are not scaled.
2 This method then uses this capability and uses 3 this to scale to full scale the difference.
4 What this method does, Professor Tien, is 5 differentiates the parts of the transient which are scaled 6 and the parts of the transient which are affected by 7 distortion. So those parts which are well-scaled we can 8 extrapolate. Those which are not well-scaled we have to 9 address and look at in more detail.
10 (Slide.)
11 The other feature of this process -- it is a 12 process -- if you want to call it systematic -- to look at
() 13 in a systematic way different problems which are connected 14 with code applications. This can be used to perform a 15 diagnostic analysis of the code capability. In performing
- 16 the calculations, we said performing the calculations and 1 17 obtaining results the code doesn't get the capability to 4 18 model this process and we are wasting money.
'l 19 This is a way to save money. Then one can see 20 whether our code is applicable to the scenario.
I 21 Because of the test matrix, assessment matrix, in 22 addition, we can detect compensating errors, and we can also 1 evaluate the qualitative effect of tuning.
24 (Slide.)
() 25 The flow chart of this method is shown on this
- ACEJFEDERAL REPOR fERS, INC.
i 202-347-3700 Nationwide Coverage 800-3364M6
5 7280 13 16 483
( Vbur 1 slide here. It consists of four elements. For a specified 2 scenario, we first identify phenomena which are important 3 and rank them both for components and for the whole system.
4 The next important step is we apply this to 5 evaluate whether the code can contrast a particular 6 scenario, whether it has this particular capability. If the j 7 code doesn't have the capability, we should not use it. If i
8 it has the capability, we can compare then the code by a set 9 of separate effects tests to address whether the code can 10 scale.
I 11 We assess it at Stage-5. We evaluate whether the 12 code has compensating errors. If it does not, we look at 13 whether the test facility has scale distortions. If there
[ 14 are no scale distortions, we can use the results with 15 certainty. If there are distortions, we can determine the 16 significance.
17 We do this for the separate effects tests, and we
- 18 do this for the internal effects tests.
l l 19 If you look at this chart, it addresses the l 20 importance of phenomena in Stage 1, scaling in Stage 4, l
l 21 compensating errors, Stage 6, scaling, Stage 7, and total 22 uncertainties, Stage 9.
23 (Slide.)
i 24 We developed at two meetings in October and l
I
( 25 February -- we applied this method to a large break LOCA.
ACE-FEDERAL REPORTERS, INC.
, 202-347-3700 Nationwide Coserage 800-336-6646 l
7280 13 16 484 1 We identified two phases of LOCA, blowdown phase and a
( l1 bur 2 reflood phase.
3 Over 24 participants were at the meeting. There 4 was unanimous agreement among all participants that the 5 blowdown phases were scaled, so there was no distortion in 6 the data for the blowdown tests and the blowdown peak.
\ 7 The refill /reflood phases have distortion because 8 the plans are distorted. The downcomer is distorted, and 9 this affects deployment in the upper plenum. This has to be 10 looked at differently.
i 11 At this meeting we also developed an evaluation 1
12 and a method to determine the uncertainty, and this is what
() 13 I will discuss on the next viewgraph.
14 (Slide.)
f 15 The uncertainty method is shown on this 16 viewgraph. It is applicable to the large break LOCA because 17 it is scenario dependent.
i 18 Items 1, 2, and 3 really stress the requirements 19 to perform code uncertainty. First, you identify the range i
20 of the conditions of the scenario, you select the nuclear lr 21 power plant you want to apply it to. It can be generic or a l
> 22 particular one. And you have to have a completely frozen 23 code. If the code is not frozen, there is no way you can 24 i determine uncertainty.
() 25 That means you have to have complete l
l ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage Mxt3364 M6 l l
\
7280 13 16 485 llhVbur 1 documentation. It has to be frozen. You have to have 2 complete documentation. But it means you must have these 3 documents that I mentioned before.
4 There are three problems to be addressed when you 5 do a code uncertainty. One is the uncertainty of the code 6 models. The second, what is the effect of nodalization? A 7 third one, what is the effect of scaling? Fourth, what is 8 the effect of plant parameters for a particular plant?
9 There are four uncertainties which have to be 10 addressed. This method really addresses the problem of 11 nodalization at Stage 4. You perform code calculations for i 12 a nuclear power plant with a given authorization. This is
- () 13 fixed. You document it. This is put aside.
14 Then at Stage 4 you select separate effects 15 tests. You calculate and compare code capabilities and what 16 are the important phenomena to this scenario.
17 And separate' effects tests. This is where you l
- 18 identify and evaluate the code goodness to model, a 19 particular separate effects test.
l 20 The requirement is that you use the same 21 nodalization. For example, if you used three axial nodes in l
22 the downcomer of a nuclear power plant for full-scale, you f 23 must use the same nodalization in the separate effects test, 24 the reason being we obtain very good agreement with the code
\
() 25 if you use identical nodalization.
/\CEJFEDERAL REPORTERS, INC.
- X)2-347-3700 Nationwide Coverage 800-3346M6
,. . - - _ - - _ - _ - .L - . - . . . _ - - . - -. . - , _ . . . - - - - . - . -
7280 13 16 486
( ? bur 1 But this is really irrelevant if we want to apply 2 this code to a full-scale reactor because I cannot afford 3 this fine nodalization. So if you want to compare 4 something, your nodalization at Stage 4 must be the same as 5 at Stage 5.
6 If the agreement in Stage 5 is not satisfactory, 7 you have to change the nodalization and you go back to Stage 8 4 and proceed. Once you are satisfied with the agreement, 9 you proceed and moderate the capability of the codes and 10 modify separate effects tests.
11 Again, you perform these tests using the same 12 nodalization as close as possible. If the agreement is not
((O
/ 13 satisfactory, you return it again. If it is satisfactory, 14 you proceed.
15 At this point here you differentiate data which 16 are not affected by distortion from data which are affected 17 by distortion.
18 For a large break LOCA, the early PCT during the 19 blowdown is not affected by distortion. That portion of the 20 transient is well-scaled. We can use these data without any 21 distortion. There is no effect of geometry or of fluid 22 conditions for this portion of the transient. So we can use 23 it better.
24 The late PCT is affected by distortion for
) 25 reasons I mentioned. The power changes very much the ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6M6
1 d
7280 13 16 487
( )/ bur 1 surface to volume ratio.
2 This affects the upper plenum, therefore affects 3 the entrainment, the entrainment affects steam binding.
! 4 Therefore, we have to take into account the effect of 5 distortion in scale at Stage 7.
6 At Stage 9 we take into account the effect of 7 plant parameters, fuel conditions, and their combined 8 uncertainty at Stage 10.
9 This is the strategy which was developed in 1
10 February of this year.
. 11 Let me tell you one thing. We are in the process 12 of implementing it. There are still problems we have to j () 13 resolve between now and December, but what it means is I
]
14 think we shall be able to reach this stage here in December.
15 The problem we have to solve is to determine the 16 effect of scale distortion on the arithmetic. This is one 17 point we have to discuss. There is no problem in the first 18 PCT.
19 We have to ree.11y consider in more detail the
- 20 third item, as really how we combine these uncertainties.
! 21 There are several ways to do it. I will not go into the
! 22 discussion because we didn't establish.
i 23 MR. WARD: Let me ask you, Novak, I get the idea i
24 of the approach but how is it going to be implemented? !
l
() 25 You are going to expect each code owner, l
/\CE. FEDERAL REPORTERS, INC.
- 202 347 3700 Nationwide Coverage 80 4 336-6646
7280 13 16 488
(,7 bur 1 developer to go through this process, is that it?
2 MR. ZUBER: No. This will be up to NRR to 3 approve or disapprove a method for the industry. This is 4 the method we want to apply in our codes to evaluate the 5 uncertainty.
6 In this case here for TRAC BF-1 MOD-1, what is 7 the uncertainty of predicting PCT for a large break LOCA, 8 which takes into account scaling, models, nodalizations, and 9 plant conditions?
10 MR. WARD: Go through this process for this one 11 code.
12 What sort of effort are you talking about?
() 13 MR. ZUBER: Let me just come in a minute to this.
14 Let me just address it and I will come back to it.
15 (Slide.)
16 When we described this method in October of last 17 year, the idea of identifying important phenomena, we got an 18 estimate from various groups that it wot*!d take for a large 19 break LOCA one scenario. It will take two months' studies 20 to develop this ranking tables, and this appeared to us to 21 be excessive.
22 What we did, we convened a group of 20 people, 23 got them in a room, and we derived or established a table in 24 one day.
() 25 So the question is the beauty is in the eye of ACE FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coserage 800-336W>t6
7280 13 16 489 i
ss jlbur 1 the beholder. There is the example of the mathematician and 2 the engineer. There is a pretty girl, and the mathematician 3 will say I will never be able to reach because as I come 4 closer the steps get smaller and smaller and I never reach.
5 The engineer will say for all practical purposes I will 6 reach it.
7 Being an engineer, I would say we did achieve 8 this process in one day.
9 MR. WARD: That was a month of man-days?
10 MR. ZUBER: No, it was a group meeting.
11 MR. WARD: 20 man-days, yes.
12 MR. ZUBER: But 20 man-days of people who knew
( 13 the process, and it was a' unanimous agreement. It is very 14 seldom you find 20 engineers agreeing on one item, and I 15 asked them. Unanimous. It was absolutely unanimous, and 16 they came to this ranking. It can be done.
17 of course, if somebody doesn't want to do it, you 18 can always find an excuse and then ask for more money to do 19 it.
20 Now, this process, this is a Mickey Mouse 21 representation of the process. I will not go through it.
22 But to answer your question, we asked our 23 laboratories to give us an estimate for various phases of
- 24 this work. As I said, it depends on to whom you talk. We
) 25 got an assessment to go through the process we described.
l /\CE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage m)-33MM6
-. . , . . - . , . - - - , - - _ - . - - . _ , , . - - - = . - - . - , . - . . , . - , . , , . . - , , . - - . - , - - . . .
7280 13 16 490 1 Wo had a meeting with our laboratories three
()7 bur 2 weeks ago in Washington. Some tasks were assigned, and 3 yesterday they reported the cost. LASL came in with an 4 outline that this effort will cost $1,400,000. I don ' t 5 think that the time was really established, but it will 6 probably be finished sometime next year.
7 I heard other estimates I don't want to discuss 8 right here because it is a very much lesser amount.
9 Also, I did ask LASL to give me an estimate of 10 options, high, low, and medium. Evidently, this was the 11 high option. They came with another proposal of I
12 implementing this method, which could be cheaper. It would
() 13 cost probably around $600,000 to $900,000, almost half, if 14 they used the statistical method based on the Latin 15 hypercube, 16 DR. CATTON: The Latin hypercube has been a 17 failure so far.
18 MR. ZUBER: This was what was proposed by LASL in 19 lieu of this. I personally see a problem with this 20 hypercube because it makes minimum use of experimental data.
21 It relies on calculations. It does not address scaling.
22 This was one of the difficulties which the 23 community had had for the last five or six years. How do 24 you address scaling? How can you assure an engineer, a
() 25 competent engineer, that what you are doing really makes a
I ACE-FEDERAL REPORTERS, INC.
202 347-37m Nationwide Coverage 800-336-6M6
, _ ~ - . . . - . - . - . - - , - . - , _ - , . _ , - - . - _ - . - . , -- - - , - .- .,. _ _ _ _ _ - _ _ . _
7280 13 16 491 1 sense in terms of engineering?
()7 bur I 2 Then we will be going, I think next week, to 3 discuss this method with Sandia. The only thing I can state 4 today, we have made no commitmeni ..o any laboratory to 5 perform this work. We shall evaluate the different methods 6 and proposals. I think that we shall make some decision l
7 next week or week after next.
8 But what we are really striving for, number one, 9 is to have this OA document by the end of this year and to i
10 implement this method also by the end of this year. We 11 shall make our best effort to do it in a way we can afford 12 it and be able to defend it to the technical community to
() 13 show that we, a regulatory agency, can regulate itself and 14 produce good data.
15 I think I came to my conclusion. If you have any i 16 questions?
17 DR. CATTON: You might walk the people through 18 this. It may look simple to you, but I am not sure they 19 understand.
20 MR. ZUBER: Maybe I should go through the
- 21 bullets.
22 (Slide.) ,
23 This is the one Sol Levy put as a result of the
! 24 discussion we had.
() 25 First, we established the range of conditions ;
i ACE FEDERAL REPORTERS, INC.
202 347 3700 Nationwide Coserage 8m336-6646 i
,,.-,,,-,--,m-.._-.,,--_-.m -
. . _ , _ _ , , - - _ .,_._-....,,.,,.,,-_,,....._m.,_,,.,,,,ym_ , ,r._ ,,. . . . , _ , - , .
7280 13 16 492
( )1 bur 1 over which we apply the code and the locations we want to 2 have on the break at various locations and spectrums, and we 3 should do this for the large break LOCA.
4 At this point here, we shall identify only one 5 NRC code. We are testing this method on the code which is 6 best documented and best assessed for the large break LOCA.
7 This is TRAC BF-1 MOD-l. This is a frozen code. We shall 8 get a OA document by the end of this year. So we shall test 9 this method using TRAC BF-1.
10 Then the requirement is to provide the OA 11 documentation. This shall also be provided by the end of 12 this fiscal year.
13 (Slide.)
14 Step number four, if you look at your chart 15 there, we perform reactor analysis for the large break LOCA 16 and we fix the nodalization. This addresses the problem, 17 the effect of nodalization.
18 We have to provide inputs for all the cases, 19 describe the nodalizations for each key element. If the key 20 element is the downcomer or the cold leg or the upper 21 plenum, we have to describe whether it is nodalized and 22 document the results.
23 This is step number one. We first perform the 24 given nodalization calculations on the large break.
( 25 (Slide.)
ACE FEDERAL REPORTERS, INC.
202-347-3700 Ltionwide Coserage 800 33MM6
7280 13 16 493 (a,f! bur 1 Stage No. 5 in your flow chart is we perform a 2 reasonableness evaluation of the goodness of the code. We 3 do this by comparing the results of separate effect tests 4 for the important phenomena, and we compare the calculations 5 using the same nodalization in the reactor case. We have 6 separate effects tests.
7 Now, if we have already performed such 8 calculations, we shall use them. We shall also use any 9 calculations we have already performed to date. So we do 10 not want to repeat.
11 And then this is Bullet 5(c). Or if the 12 nodalization is not the same in the reactor and in the
() 13 separate effects test component, we have to show that the 14 nodalization is not important. You have to give a ratinale 15 why the nodalization in these two separate effects tests and 16 code calculations are not the same.
17 (Slide.)
18 Now, at this point here we have to identify for 19 each important component -- let's say this may be the pump, 20 the downcomer or whatever -- if we do not have enough 21 separate effects test data, they are statistically not 22 significant. Maybe we have only two data points or 23 something like that, or the agreement between the code 24 calculations for that particular nodalization are not in l
I
() 25 good agreement with the experimental data, or the separate l
l l
/\CE. FEDERAL REPORTERS, INC.02-347 370) Nationwide Cmcrage 80M3MM6
7280 13 16 494 I~'7 bur 1 effects tests do not cover the range of interest for nuclear N./
2 power plant applications, and therefore we identify this 3 also at Stage 5.
4 We do this for all the components which are 5 important, and this is Step 5 in the flow chart, assessing 6 the goodness of the code, going through these five items.
7 (Slide.)
8 This is on the Mickey Mouse graph you have there.
9 You have two graphs. One is the flow chart. The other one 10 I call the Mickey Mouso.
11 There is a number SP or SG. The numbers 12 correspond to the bullets here.
() 13 Four of the sets of experiments which do not 14 cover the entire range we may extrapolate to the full scale 15 and evaluate the uncertainty to the full scale. I don't 16 think this will be the case for a large break LOCA because 17 for a large break LOCA we shall have -- once we have data 18 from UPTF we should have data on all the components which 19 are distorted in the integral test facilities.
l 20 The components which are distorted are the upper 21 plenum, lower plenum, and downcomer and hotleg. These are 22 four sides in UPTF. From these experiments we should get 23 full-scale data and models to put in our codes.
24 So if you look optimistically in the future, what
() 25 I would expect -- and I really hope we should be able to ACE FEDERAL REPORTERS, INC.
202-347-3700 Nanonwide Coserage 80b33MM6
7280 13 16 495
()/ bur 1 show that we have a body of data which can address the use 2 of integral facilities which are well-scaled for a portion 3 of the transients. Then you use separate effects tests, 4 again full scale, so that we really can cover the entire 5 range of scaling in a rational way, and I think if you I
6 provide us the information which we didn't have to date.
. 7 (Slide.)
t i
8 Now, we do the same thing for the integral test 9 facilities. This is Item 6 on the flow charts in the Mickey 10 Mouse table.
11 We again clearly divide the transient in two 12 parts, the early PCT and the late PCT, the reason being one
( 13 is well-scaled, one is affected by distortion. We do the 14 same in nodalization, and we do then a matrix to compare the i
15 code versus data.
l 16 What we have to do now is to identify whether we i
17 have or not calculations, integral facilities and plant 18 calculations which have similar nodalizations so that we do 19 not have to repeat calculations and we would save money and l
20 time.
21 If not, we have maybe to perform some additional 22 calculations.
23 (Slide.)
'l l 24 Now, for the case that we do not have enough data
() 25 or for the case that the results in the test facility may be ACE FEDERAL REPORTERS, INC.
! 202.",47 3700 Nationwide Coverage MU-336-6M6 l
7280 13 16 496 1 affected by scale, we may do several things.
lllVbur 2 This is something we have still to go through and 3 resolve between now and December. How do we treat the 4 seconid peak?
' I would also like to say 5 There are several ways.
6 at this point here that there is an alternate approach that 7 has been provided, and I can discuss it in a minute, but
- 8 this is one where we still have to resolve the effect of 9 scale on the reflood peak. We may try to determine an
'10 uncertainty, and I think this is shown in the graph there.
11 There may be some others we shall explore to resolve Item 12 6(d). This is not resolved.
() 13 DR. CATTON: I don't understand where the lack of
- 14 resolution is. Where are you missing data?
- 15 MR. ZUBER
- No, there are two conditions. Either 16 we do not have data --
i 17 DR. CATTON: I understand, but you just said the 18 second peak and tha' this is a concern.
i i 19 MR. ZUBER: No, no, no. If you don't have data, 20 that is one concern.
21 The other concern is where the results may be i
22 affected by distortion.
) 23 DR. CATTON: But we have lots of data?
24 MR. ZUBER: That is fine. Then we have to i
() 25 resolve what is the effect of scale distortion on the test i
i ACE-FEDERAL. REPORTERS, INC.
202 347 3700 Nationwide Coverage 800 336-6M6
--L. , - -, e b+ 6 L - - - - - 4 :r-- w - -- 4 , - a
! 7280 13 16 I
497 4- ('
(,,S1 bur 1 data and on the code.
2 DR. CATTON: Okay. I just don't think there is 3 as much distortion.
4 MR. ZUBER: Look, okay, you may feel it. I also 5 may feel it. Put it this way:
6 I feel that once you obtain data from UPTF all 7 these answers will be very conservative.
i 8 DR. CATTON: With reflood, we have data for j 9 something from all the way not bigger than this bottle all 10 the way to UPTF.
11 MR. ZUBER: Okay. I said once we have it for
! 12 UPTF we shall show it. My gut feeling is, as an engineer,
() 13 we shall show the results are conservative, but I don't have l
14 to convince myself. As a regulatory agency, you have to 15 respond to a criticism. Can you do it? Can you prove it?
16 And I think we shall be able to prove it.
i 17 (Slide.)
i 18 Item 7, this is the thing that we have really to I 19 address the effect of scaling. Either this is in separate
( 20 effects facility or in integral, but as I said, I think UPTF l 21 will provide enough data at this point here that we can take i
22 care of the separate effects tests' results.
- 23 And the only thing we have to worry -- or to i
1 24 concentrate is on the second peak during reflood.
l 25 (Slide.)
i ACE. FEDERAL REPORTERS, INC.
, l 202 347-37m Nationwide Coserage 8m336-6M6
.- _ _ _ - .=. _ . - - . .- - .- - - - .
)
l 7280-13 16 498 1 Step number eight is to estimate code
()7 bur i 2 deficiencies.
3 Now, this one, first, if the correlations are not a
4 satisfactory in the code, one, if the data from separate 5 effects tests are not sufficient or the comparisons are not 6 satisfactory, if the integrated test facility is also not 7 satisfactory, you can add a delta PCT due to deficiency of 8 the code.
I This, if you want, is like a safety feature, 9
10 where you either do not have data or you do not have a good j 11 feeling about the correlations or the agreement with the 12 data is not good enough.
() 13 Then you may put an uncertainty on these data, we
. 14 hope, but we have to show it. This may not be very much.
I 15 (Slide.)
16 The question is how do you combine these i 17 uncertainties?
18 One way which was suggested at the February 19 meeting -- and we still have to discuss it -- is to have an 20 uncertainty due to code and experiments, the other one due
- 21 to scaling -- this is Items 6 and 7. This is Item 6(c),
22 Item 7, and 6(d) -- one due to code deficiency, Item 8, and 23 plant inputs, Item 9.
i 24 The reason is that there is no uncertainty due to
() 25 nodalization because this method really uses the same
{
ACE FEDERAL REponTEas, INC.
j 202 347 3700 Ntionwide Coserage MO-33MM6
7280 13 16 499 1 nodalization in assessment of separate effects tests and lllVbur 2 integral effects tests.
3 I don't really want to leave you with an 4 impression that we have really solved the problem. We have
- 5 established less than a month ago -- not a month - in 6 February a method whereby we can now do scaling and 7 nodalization. We have still to test it. We are not there, 8 but we have a role. We have to clean it and play with it.
9 But I do hope we can really prove to an engineer, 10 a skeptical engineer, that we can address the effect of 11 nodalization.
12 The requirements are good documentation. I have,
, () 13 as I say, willingness on our part to work on it and to 14 succeed.
.i 15 MR. WARD: For the three codes, you know, the
- 16 three agency codes here, you don't really have the 17 documentation to even enter this process?
18 MR. ZUBER: That is right. That is the reason we 19 want to have this for our best documented code, which is 20 TRAC. It is not very good, but we have it.
21 Let me say this. We are changing our mode of i 22 operations. Up till recently, we didn't rely on best 23 estimate codes. Now, we have to rely because the vendors 24 are also going to rely.
() 25 So for many reasons the documentations of our
! ACE FEDERAL REPORTERS, INC.
l 202 347 3700 Nationwide Coserage MIX)-33MM6
4
! 7280 13 16 500 1 codes were not as good as vendors' because NRR imposed some
( )7 bur 2 standards on the vendors, and for whatever reason, our 3 documentation is not as good.
4 But as I said, our upper management made a 5 decision, made actually several decisions, and the 6 developers were informed yesterday.
7 One is that from now on no code will be released 8 unless it is fully documented, and this should prevent us 9 from paying twice for something, for the same information.
10 I think this is a great decision.
11 The second thing is that we shall install and 12 enforce a quality assurance and inspection of our codes from
() 13 our contractors. We cannot ask anything less than we demand 14 from them.
j 15 I doubt that we would ever get an argument on j
16 that.
17 And, as I said, we made this documentation 18 available. We need only a good QA document. We expect to 19 have it by the end of December of this year, and we shall i
20 use these codes to assess a method for determining TRAC.
21 Then we may follow it with RELAP or TRAC-3. But this is the 22 test case.
23 As I said, when we developed these methods, there i 24 were usually about 25 or 30 people, participants, and the
] () 25 more people sit on the problem you get more ideas and points ace FEDERAL REPORTERS, INC.
j i 202-347-37W Nationwide Coserage 800-3346M6
-_ , , . . _ . . . , . . - . . . . _ . . . . _ _ . _ . , . . . - , . _ . . . _ . _ . _ _ _ _ _ . _ _ _ , . _ . _ . _ _ . . - . . - . _ . . _ _ _ _ ~ . . . . , _ , _ _ _ _ . . . _ _ _ _ _ _ _ _ . _ _ , _ . _
7280 13 16 501 lhVbur 1 of view from different sides.
2 So if you have any comments, any criticisms, any 3 suggestions, we really welcome it because we would like to 4 implement it and test it. We are really not wedded to this 5 method or any other, but next week we shall select one of 6 the contractors who are going to do this and select methods.
7 Probably this will be done next week or the week after next.
8 DR. SCHROCK: Novak, one thing that caught my eye 9 in looking at the compendium document, Section 6, I didn't 10 find reference to the flow regime map, which I think is 11 pretty ad hoc in each of the codes.
12 In this process did you confront the question of
( 13 assessment of uncertainty due to the calculation of 14 incorrect flow regimes?
15 MR. ZUBER: This is something I think will be 16 addressed for each flow regime. It will be each be 17 addressed by the code.
18 Let me say one thing. I have very often heard i
. 19 arguments -- I don't want to say " arguments." Let me put it 20 even more bluntly -- excuses -- I cannot do this because the 21 first thing, my flow regime map is not good.
f
- 22 This is not exactly true. One reason is that we i
! 23 have flow regime data for rod bundles back in the sixties, I i
24 think. We also had a research contract. So for an i
() 25 important component like a rod bundle we have flow regimes.
1 ACE FEDERAL REl'ORTERS, INC.
! 202 347-37(C Nationwide Courage Mn336-(M6
-,-_-______,_-,__...,._______.,_.m.- . _ _ _ _ . _ , _ , _ _ _ _ _ _ _ _ , __ _ _ _ , . , . _ _ _ _ _ _ . _ _ _ _ _ .
7280 13 16 502
()/ bur 1 If a contractor didn't use it, then he has to 2 justify why he used something else. But this is done in the 3 OA document.
1
! 4 Another question which comes up is that the flow 5 regime maps which appear in the literature are usually based 6 on dimensionless maps -- J-star or something -- whereas, the 7 flow regime maps which are in our codes are based usually on l
8 alpha.
- 9 Now, I didn't write this up, but there is a one-I 10 to-one relationship. You can show for some regimes that i 11 there is a one-to-one correspondence, but based on these i
12 dimensionless groups. So if the flow regime change of alpha
( 13 is as low as .25 or .3, this can also be expressed not in i
14 terms of alpha, but in terms of dimensionless groups, 15 because for every critical thing there is a critical void i 16 fraction.
l 17 So it doesn't make a difference whether I want to 18 do this in the terms of a void fraction or I can express 19 these bounds in terms of alpha or express these bounds in i 20 terms of a dimensionless group.
21 I don't have time now, but there is nothing wrong l
i 22 having a flow regime mapping expressed in alpha, because 23 there is a critical alpha. All this is corresponding to a
- 24 critical transition. So whether you want to discuss it in
() 25 terms of dimensionless groups or alpha, it depends on the i
i ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Cmcrage Km336-6M6
7280 13 16 503 l Ebur 1 usage.
2 DR. TIEN: When we are talking about flow regime, 3 we also have to be careful whether you have the adiabatic 4 situation flow regime. If you have a condensing flow 5 regime, it could be totally different.
i 6 MR. ZUBER: Absolutely right.
- 7 Now, let me say I first put my best foot forward
! 8 and I said they are not that bad because there are important 9 components like rod bundles we have flow regimes, but we are 10 getting, I think, 14-inch data from ROSA. So we shall have 4
- 11 a span on the horizontal axis. In the upper plenum there is 12 always a structure. The upper plenum is dry, but it is 1 () 13 really not empty. There is really a geometry.
14 In that case this is the component, and I don't i 15 have a really strong feeling that our flow regimes are good, 16 but in some components I feel pretty comfortable.
17 Now, what you bring up is another important case, 18 the effect of condensation. This has to be addressed, but 19 it has to be addressed in the context of the condensation 20 model.
21 For today, our goal for December, we have to be
! 22 able to establish this is good, this we can use with j 23 confidence. Let me say the code is good for many
! 24 conditions. There is no question about it, but we cannot 1
) 25 say the code is good for anything because we have been
't ACE FEDERAL REPORTERS, INC.
l 202-347 3700 Nationwide Coserage WK 33M446
-- --, , _ _ . . . - _ . , . . _ _ _ . _ . , _ . . _ . . , . , _ . , . . . , _ _ _ . _ _ _ _ . . ~ _ _ _ _ _ - . - . _ _
~_
7280 13 16 504 1 -coping. We have to identify where there are shortcomings,
()/ bur 2 like the peak condensation in TRAC for Calvert Cliffs, so 3 we have confidence that we can use it here, but we have to 4 be very careful here, and I cannot use it here, period.
5 That is important.
6 Now, we have data. What needs to be done is to 7 put it -- -
I 8 DR. TIEN: I am not even sure we have data. I
- 9 looked into this situation lately. We find for the 10 condensing flow, two phased flow, we have some recent l
) 11 limited data about horizontal flow geometry, but for 12 vertical almost none, and you cannot draw a really good, you i
() 13 know, flow regime map for condensing two-phased flow. It is l 14 really a bad situation over there.
i
)
15 MR. ZUBER: We may have to look at how important i 16 and for which transient this is important to perform a I 17 sensitivity analysis and see how sensitive are the results 18 to what model of condensation we are using.
19 DR. CATTON: I think at least for the large break l
l 20 LOCA that can be put aside, except maybe in the cold leg.
i l 21 MR. ZUBER: Okay.
22 DR. CATTON: In establishing the pressure j
23 boundary conditions, but other than that I think the i condensation can be put aside for this first exercise.
i 24 f () 25 MR. ZUBER: That is the reason we selected it.
/\CE. FEDERAL REP 40RTERS, INC.
202-347 3700 Nationwide Coverage m)-3364M6
I j 7280 13 16 505 ;
( 7 bur 1 We know most about it, and we know the phenomena, but we 2 wanted to demonstrate that you can determine the 3 uncertainty.
f 4 MR. MICHELSON: Let me understand the process for i
5 a moment. This method will be described in the compendium?
6 MR. ZUBER: It is.
q 7 MR. MICHELSON: But until after you get through 8 doing your exercise, your test exercises, and so forth,
, 9 there is probably going to be some fine-tuning of it, or do 10 you think it is all finished and frozen now?
t 11 MR. ZUBER: We shall do this with a frozen code.
) 12 MR. MICHELSON: Maybe even change the method 13 itself.
- 14 MR. ZUBER
- Oh, yes, we have to test it, that is j 15 right.
16 MR. MICHELSON: I am just trying to determine the i
17 sequence of things here. The compendium will contain a 18 description of this method.
1 19 How are you going to revise it later to decide
! 20 what you are really going to do?
21 MR. ZUBER: Dr. Beckner addressed this question.
t l 22 He said there are sections in Part 5, which will be then
- 23 enlarged for the next version of the compendium, which would
! 24 include this method.
i
() 25 MR. MICHELSON: How soon will that next version 1 ACE FEDERAL IlEPORTERS, INC.
202 347-3700 Nationwide Cmerage 800-33MM6
- - - - - . ~ _ - - . - . - _ .- - _ - .- - ,.-- ..,.,.-- ...- - --, - -. - - - .,-,
7280 13 16 506
( ) bur 1 of the compendium be issued? j 2 MR. BECKNER: We are not exactly sure. This is a 3 draft for. comment. So even the description of plant might 4 change based on our experience as we get into it, or else 5 the comments we receive.
6 Right now, the comment period ends July 1, and I 7 would not think we would start a major rewrite until after 8 the comment period ends.
9 MR. MICHELSON: By the time the rule is adopted, 10 certainly the compendium would be in good shape?
11 MR. BECKNER: Hopefully, we will continue along 12 the path, whether in lock-step or nearly in lock-step.
() 13 MR. WARD: But your trial run here, what do you 14 call it? The CSAU methodology isn't going to be completed 15 by July 17 16 MR. ZUBER: Oh, no.
17 MR. BECKNER: The comment period ends July 1.
18 Right now the final rule, at the earliest, is in the spring.
19 So we have got to rewrite.
20 MR. WARD: I will ask the question, then. Is the 21 trial run of the CSAU methodology going to be complete by 22 next spring in order to be incorporated?
23 MR. BECKNER: I think that is Novak's schedule.
24 MR. ZUBER: That is the reason we are pushing for
() 25 December.
ACE. FEDERAL REroRTERs, INC.
202w.nm Nationwide cmerage sm-U6 6646
7280 13 16 507
/~N
(_,1bu r 1 MR. MICHELSON: I wasn't quite finished with my 2 question. The rule won't refer to this method specifically.
3 The rule will refer to the fact that you must have a means 4 of determining this.
5 The regulatory guide, will it refer to this 6 method? What will it refer to?
7 Again, you have to have something.
8 MR. BECKNER: The regulatory guide right now is 9 rather general, which says these are the types of things you 10 have to consider, like scaling. It doesn't say how to do 11 it.
12 MR. MICHELSON: Again, of course, not being in
( 13 the rule or the guide, I am not sure ACRS would comment on 14 this unless the full committee decides to pick this 15 particular item out because it is not part of any process.
16 It is just a sample of how it might be done.
17 MR. WARD: Well, it is a pretty important part, 18 whether it is formally a part.
19 MR. MICHELSON: This may not be the method used 20 at all. It might not be referred to specifically in the 21 guide, apparently.
22 MR. WARD: But the NRC is going to insist that 23 this be used for its own codes.
l 24 MR. MICHELSON: Just for its own, though.
t
( 25 MR. WARD: Or that something be used.
I l
Aci! 171 :nt:RAL lttiroRTI:Rs. INC, 4 20244707m N.ioonmde Cmcr.ye kmlM IM6 I
7280 13 16 508 1 MR. SHOTKIN: What we would like to -- we can't
()? bur 2 use the word " insist," but we use the word " encourage,"
3 because I am not sure of the legalistics that you seem to be 4 bringing up here. We look at the package as all three parts 5 coupled together -- the rule, the reg guide, and the 6 compendium.
7 Inside the compendium there will be this method.
8 We would like the ACRS to comment on all three parts of that 9 package. I don't think we would like the ACRS to say we are 10 not going to comment on one part of the package because it 11 is too loosely coupled to the other parts.
12 MR. MICilELSON: Uncoupled to the other parts.
() 13 There is nowhere in the rule, nowhere in the 14 guide that this will be mentioned, as I understand it.
15 MR. SilOTKIN: The rule refers to the reg guide.
16 The reg guide refers to the compendium. The compendium 17 contains this method.
18 MR. MICilELSON : We are going around in circles 19 all day today.
20 Is this method going to be spelled out in the reg 21 guide?
22 MR. SilOTKIN: No, it is in the compendium.
23 MR. MICilELSON: But will that reference to the 24 compendium, that particular chapter?
() 25 MR. SIIOTKIN: No.
Acti FEntinAL RiironTnns, INC.
2C347 37m Nationwide Cmcrage kab34 6M6
1 7280 13 16 509
( )/ bur 1 MR. MICHELSON: Then there is no linkage between 2 the two.
3 MR. SHOTKIN: All the reg guide will say is 4 accept certain models and data. It will also say amplify 5 the rules, demonstrate the high probability. The reg guide 6 will say demonstrate to 95 percent probability. The reg 7 guide will say what phenomena should be considered. The 8 compendium will contain or will be referred to in the reg 9 guide as the basis for demonstrating the applicability of 10 any model that the industry submits. They will have to look 11 at the methodology that we are using because that is the 12 standard we are using for our own codes, and they will have
( 13 to be aware of it at least when they come in for their 14 licensing submittals.
15 I think for that reason alone we want comments on 16 it.
17 MR. ZUBER: Let me give you an example. If they 18 have to address the problem of scaling, they may use this 19 method. They may say I have a delta PCT due to scaling of 20 so much. If they can justify this, this may be 21 satisfactory. They may devise a completely third method to 22 do it.
23 So I think they should have the freedom to select 24 and do it. It is up to NRC to evaluate what they do,
() 25 whether what they do is reasonable or not.
ace-FEDERAL REPORTERS, INC.
j 202 347.3700 Natia. wide Cmerage 80rb 336 (M6
7280 13 16 510
/~N
(_,Jbur 1 DR. CATTON: So this becomes a tool to confirm 2 that the delta PCT picked by one of the vendors is 3 adequately large?
4 MR. WARD: It could be.
5 MR. ZUBER: Not necessarily.
6 MR. WARD: Do you expect the vendors to use your 7 codes in their submittals? Will you permit them to use 8 your codes?
9 MR. SHOTKIN: Yes. Why won't they?
10 MR. MICHELSON: They will develop their own.
11 MR. WARD: No, I don't know. They can't go 12 through this routine with their codes.
() 13 MR. SHOTKIN: We already have some submittals by 14 vendors. Some have come in under SECY 83-472, and some I 15 believe are being submitted under the proposed new rule.
16 MR. WARD: The first one is just BWRs, isn't it?
17 MR. SHOTKIN: There is Westinghouse two-loop 18 PWRs. CE is coming in with something.
19 MR. WARD: I thought 83-472 just applied to --
20 MR. LAUBEN: 83-472 just adds to the Appendix K 21 required feature calculations. We took that away,83-472.
22 But as Lou was saying or as you were saying, I 23 think, GE has come in with their safer methodology, but in
, 24 addition to that Westinghouse has come in with methodology
() 25 for two-loop plants, and two of the two-loop vendors Aa FEnERAL REPORTERS, INC.
I 202 mmoo Nanonmde coserage mowuw,
7280 13 16 511
( 7 bur 1 contract with CE to come in with a methodology for those two 2 plants as well.
3 So --
4 MR. WARD: What are they using?
5 MR. LAUBEN: What they are proposing to use is 6 their modification of a German code for refloods. Their 7 modification is called CEUPER. But they will use what they 8 call their best estimate version of flash and blowdown.
9 Exxon is proposing in some time in the future to 10 come in with a best estimate methodology, and those are 11 really it.
12 DR. CATTON: Is someone proposing COBRA TRAC 7
() 13 MR. LAUBEN: Westinghouse has already submitted 14 their version of COBRA TRAC. So there is a spectrum of 15 submittals, which includes all NRC-developed codes, some 16 NRC-developed codes, and some of their own development, and 17 those which were principally developed by the vendors 18 themselves.
19 There is indeed an entire spectrun of what kind 20 of codes are being submitted or proposed '.o be submitted.
21 (Slide.)
22 MR. ZUBER: This is the last slide on schedules.
- 23 MR. MICHELSON
- Other questions or "omments?
24 DR. TIEN: I would like to say I think this is a
() 25 very good development. Of course, the detailed review, we ace FEDERAL. REvoRTERs, INC.
202 .147 17(X) Nationwide Coserage 8(W)-33MM6
1 1
i I 7280 13 16 512 lhVbur 1 didn't have time to go through that, but I think we can 1
2 still comment on the general one.
l 3 The direction is a good one. I particularly like 4 to see the scaling closely intertwined with the code 5 numerics and nodalization. That really hasn't been done l 6 before, clearly. So I personally feel this is the right i
7 direction.
l 8 MR. MICHELSON: Other comments?
9 DR. TIEN: Maybe the schedule is too ambitious.
10 By December 31st I don't know whether you can get all the 11 things.
12 MR. ZUBER: Let me say there is one thing that I I
() 13 think we have done quite a bit of hard work. I can see that 14 we can put our arms around the first peak. I think we can 15 do this reasonably surely by December.
16 The other one we still have to do some work. So 17 optimistically, I think we can do both. For sure, I do
't 18 expect the first peak.
19 DR. CATTON: The second peak i. going to depend a 20- lot on what is found out at UPTP. I think you will be able 21, to conclude that the second peak is conservatively 22 calculated, and that in a way satisfies the December 31 23 requirement.
24 MR. ZUBER: We shall be going next month to a
() 25 UPTF meeting, and we shall try to maybe change the schedule ACE FEDERAL REponTEns, INC, 202-347-37(0 N.itionwide Coserage Mo-34u44
-. . - _ _ - . - - ~ .
l 7280 13 16 513 llhVbur 1 and put some more tests between now and this year, and you 2 can put the UPTF, which will help us --
3 DR. CATTON: I thought you had already done that.
4 MR. ZUBER: We did, but we want to do something 5 more. I think the interest of all three partners are the 6 same. So I think we can tango together on this, but thank 7 you, Professor Tien, for the comment. The proof in the 8 cooking is in the pudding, and the pudding I hope will be 9 delivered either well-cooked or raw or over-burned, 10 hopefully by December well-cooked.
11 DR. CATTON: As long as it is eatable.
12 MR. ZUBER: That is right, thank you.
() 13 MR. MICHELSON: I would like to take just a 14 minute to make a request to the consultants. If they would 15 refer to the meeting agenda schedule so they can follow what 16 I am asking for.
i
, 17 What I would like to have as a result of this 18 meeting is a letter from each consultant giving not detailed i
19 comments -- those would be perfectly all right to transmit 20 d',rectly -- but rather overview comments or general comments l
!, 21 concerning Item 3, the technical integration center, and 22 when commenting on Item 3 give particular emphasis to the
! 23 proposed B&W steam generator study. Don't overlook that in 24 your comments.
() 25- V In addition, I would like comments on Item 4, I ACE-FEDERAL REPORTERS, INC.
i 202-347 3700 Nationwide Cmerage 800-336-6M6
. _ . _ _ _ . - - , _ . _ _ , . . - - _ - _ _ ~ - . - _ - , ._ _ _ _ , _ , ,
7280 13 16 514 ,
lhVbur 1 which is the CEC proposed facility, and comments on Item 10, 2 which is the research compendium, keeping in mind of course 3 just general comments, general observations, an overview 4 type thing, not detailed comments.
5 And I would like some initial comments on Agenda i
6 Item 11, which is the code uncertainty process, again in an 7 overview sense.
8 The plan right now is to have our next meeting on 9 June 18th. I woul.1 like to receive these comments perhaps -
10 - excuse me, we have got a month -- about May 15th, if you ;
i 11 could get the comments in.
l 12 The meeting of June 18th, we would like to have l
5 ( 13 the comments because it affects the agenda. The agenda is
! 14 still somewhat open for June 18th because we already have ,
j 15 more items than we can cover in one day. We are not sure 16 whether to add a little time at that point or have carryover 17 items for the items which were tentatively scheduled.
- 18 DR. SULLIVAN
- Carl, one of the things that a
19 becomes apparent to me is that we may have spent a lot of 20 time reviewing the compendium, but we should spend a lot of 21 time reviewing the reg guide.
i 22 MR. MICHELSON: For instance, yes, I would 23 anticipate that we would discuss that reg guide. We could 24 discuss in a preliminary sense at the meeting, but we are
() 25 really planning not to dig into it unt,il about September.
ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-M46
7280 13 16 515 o
! jIbur 1 We will dig into it when the Appendix K rule comes through.
2 DR. CATTON: We really have to take a look at the 3 reg guide before we can comment, even in a general way, on 4 the compendium.
5 MR. MICHELSON: That is why I was concerned that 6 we may have to add a little bit of discussion at the June 7 meeting.
8 DR. CATTON: Maybe just general comments?
9 MR. MICHELSON: On the reg guide.
10 Urfortunately, we have got a long shopping list 11 here for June.
12 DR. CATTON: It seems to me the reg guide ought
() 13 to describe a regulatory guide. Guide means description of 14 a process.
15 MR. MICHELSON: You may wish to comment in your 16 letter what you think needs to be done, keeping in mind that 17 you otherwise would not really discuss it until about 18 September, as I see it.
19 MR. SHOTPIN: We will accept comments from the l
20 ACRS at any time; however, the comment period officially 21 closes July 1st. It would be most helpful to us to get 22 whatever comments you have in on either the reg guide or the 23 compendium by that time.
24 MR. MICHELSON: The full committee does not
() 25 intend to comment until after the comments are in, which i
l ACE FEDERAL REPORTERS, INC.
( ) 202-347-3700 Nationwide Coserage 800-336-6646
7280 13 16 516
(~h
( ,7 bur 1 would put us in September. -
2 The subcommittee might give you comments, but the 3 full committee would not look at this before the public 4 comments are in.
5 MR. SHOTKIN: The subcommittee or individual 6 consultants.
7 MR. MICHELSON: I have no problem with the 8 detailed comments going directly to you. If it is in letter 4
9 form, of course, we need to have copies, but I would like to 10 have the consultants give us a letter of overview, and that 11 should take care of it.
J 12 Now, the subcommittee members should think about
( 13 the fact that for the May meeting we have prime time one 14 hour, two or three next Saturday, I think, to discuss this 15 subject.
16 I had planned on giving the. subcommittee report, 17 indicating what we are doing. I don't know how many will be 18 left to listen by that time, but I will go through a 19 subcommittee report. I don't intend to take an hour to do I
20 it, simply because I am not sure we need that much detail at 21 this time.
22 I don't intend to bring anybody in for this type i
23 of thing on Saturday for sure. So that is the present plan, 24 is to really do nothing more than tell the full committee l r
\ 25 that we have had a meeting, that we will have another i
l ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6M6
7280 13 :6 517 A
(_j/ bur 1 meeting and the real emphasis of the meeting, and there is 2 really no good mechanism for presenting the compendium to 3 them.
4 MR. WARD: I think you ought to tell them about 5 it.
6 MR. MICHELSON: I mean in terms of going into any 7 chapters or even into this uncertainty methodology. It is 8 beyond my capabilities to do in that short time.
1 9 DR. CATTON: But I think it is the single most 10 important thing associated with this meeting. Someone has 11 got to do it.
12 MR. MICHELSON: I don't necessarily disagree with
() 13 you, but I am not sure the full committee can get involved 14 in it. I will ask them what is approximately involved and 15 ask them do they want to do it, and if they do, we will 16 schedule a full committee meeting for July.
17 DR. CATTON: Give them an opportunity.
18 MR. MICHELSON: We can even give them Chapter 4.
19 For the May meeting we will just tell them of its existence 20 and take a poll of what interest they have in looking at 21 this.
22 If I was looking at nothing else, I would look at 23 that one chapter, but I don't know what their wishes are.
24 But I am wondering, the next question now is on the part of
() 25 the subcommittee meeting as to whether they would recommend ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6M6
7280 13 16 518
()7 bur 1 this whole matter.
2 Clearly, we would recommend that they would 3 review the Appendix K and the regulatory guide. What is our 4 feeling on recommending going into the compendium, which I 5 think would involve at least three hours on the part of the 6 full committee in July if they want to comment on it.
7 At least three, maybe four hours, because you are 8 talking about this uncertainty analysis, you would then have l 9 the Staff come in and give the details. So what do you 1
l l 10 think?
l 11 I won't tell you my opinion.
12 MR. EBERSOLE: You are going to get lots of i
() 13 individual reactions. I would open it.
14 MR. MICHELSON: That is one way.
15 David?
16 MR. WARD: I think the compendium -- there is no 17 reason to agonize through that. That is history. The thing 18 that is new is what we were just hearing from Novak, and I 19 think it is an interesting approach.
20 You know, the committee bought off earlier on the 21 general concept of going to the best estimate -- or endorsed 22 the best estimate with an uncertainty allowance or 23 something. We didn't know what the hell that meant exactly.
24 Now, I think they have evolved a rather specific
() 25 methodology, which is kind of different from traditional ACE. FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6646
7280 13 16 519
( )/ bur 1 approaches. I think we ought to hear about that.
2 MR. MICHELSON: How much do you think we can get 3 enough across to them? About an hour? Because really it 4 doesn't fit the rest of the compendium.
5 My own opinion is we ought not to get into the 6 compendium at all, but I would be willing to make a 7 reservation, okay, we will go into the chapter on 8 uncertainty.
9 MR. WARD: I don't think it will be a bad idea to 10 have someone describe the compendium and its purpose.
11 MR. MICHELSON: For the full committee in May I 12 will tell them that the subcommittee would not recommend
() 13 that they get into the details df the compendium except for 14 Chapter 4, in which case we would recommend that they take a 15 look at that and, if the full committee wishes, we will make 16 the presentation to them.
17 I would imagine we would have to allow an hour 18 and a half. There is just a lot of questions.
19 MR. WARD: As a matter of fact, the committee, 20 when it encouraged the general approach, called out that 21 concern.
22 MR. MICHELSON: If they do wish to hear about it, 23 I would propose the July meeting for the presentation.
24 That gives us one more subcommittee meeting in
() 25 which to bring forth anything else before the July meeting, 1
ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-M>46
7280 13 16 520
( 7 bur 1 and that would be the last that would be presented until 2 September.
3 DR. CATTON: There are three reports, 4 particularly, of the last one of those meetings?
5 MR. MICHELSON: We will send a full package of 6 materials by mid-May to all the members for any discussion.
7 I don't anticipate any. It is really not till July, but we 8 will try to get it to them in six weeks.
9 Any other questions, comments or concerns?
, 10 (No response.)
I 11 MR. MICHELSON: We should be sure now to mark the 4
12 16th day -- pardon me -- June 18th. I assume there is no
( 13 problem.
14 If there aren't any other comments - and I see 15 no indication -- we are finished.
16 (Whereupon, at 3: 55 p.m., the subcommittee was 17 adjourned.)
i 18 i
l 19 20 21 l
22 23 24
() 25 ACE FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6M6
CERTIFICATE OF OFFICIAL REPORTER (a~s This is to certify that the attached proceedings before the UNITED STATES NUCLEAR REGULATORY COMMISSION in the matter of:
NAME OF PROCEEDING: ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON THERMAL HYDRAULIC PHENOMENA DOCKET NO.:
PLACE: IDAHO FALLS, IDAHO DATE: WEDNESDAY,' APRIL 29, 1987 were held as herein appears, and that this is the original transcript thereof for the file of the United States Nuclear Regulatory Commission.
(sigt)
Ncrf (TYPED)
DAVID L. HOFFMAN V Official Reporter ACE-FEDERAL REPORTERS, INC.
Reporter's Affiliation O
O O O- ~
^
i i
INEL WORKING CONCEPT AND COMMENTS BY SCALING REVIEW GROUP
/
\ W i awm ANDunswL J. R. WOLF l
metasseneses, 24marenarr APRIL 28, 1987 4
%b i
I i
O O O' I
OUTLINE o REQUIRED PARAMETERS o PARAMETER RELATIONSHIPS o TRADE OFFS o " WORKING CONCEPT" RECOMMENDATIONS
O O O .
PARAMETERS REQUIRED TO DEFINE SYSTEM MODEL o WORKING FLUID l
FLUID PRESSURE o LENGTH RATIO I o DIAMETER RATIO OR LENGTH RATIO / DIAMETER RATIO
! o CORE POWER FULL DECAY l
- o BASE PLANT o AVAILABLE FUNDING l
I i
"~
O PARAMETER RATIOS FOR SINGLE-PHASE NATURAL AND FORCED CIRCU NATURAL CIRCULATION Paramette Ratio $
Symbol le Forced Ci-culation le Natural Circulation ?e Natural Circulation Parameter top (defined) top (defined) tog (defined)
Length top dog (defined) dor (defined) dog (defined)
Diameter dog 2 2 d og Area a g dg d,g vn aRsn tog s on acR tor Volume o o o g
Core aT af g aT,q(defined) h "I/3 Velocity ug Q sp $ Q
-1/3 Time tg Q Q ag g 1
1 Gravity gg 1
-~i- a:. (iin)a'o..o.c, (en)(a),.,,<s (1/n)(a), <R <,R
-t ~ 4, (lin)aT..o.C.,
o (un)(g), .<, ,,, (1/a)(%),.f ,,R Core Pw ig a opg aT,, ,,, Cp p a,pg efp hfgp a,pg h ofR fgR n
1 1
Rod Diameter RD g 1 ag ag ano NMer of Rods NRg e e 3 1/3 Flow Rate 4g a,g 5 og sp 5cq oa, n a,p g ,,p h
.. -- y aM Subcoolin9 aH 9g af Subcooling aTgug afoR aToR
~~
Pug Head 4HdR tor Friction Nder Fg 1 i 1
O~
~
O O Scaled Parameters for B&W LPWR (Full Pressure Water) 200 -
10 l-'
\'Ns ',' 1500 SG tubes 18 0 -
9 - ',,-
N
[t R
= 1.0
,, sN 160 -
8 -
. ,- s s
1000 SG tubes (
_#A on
= 2.0 NN '
l 14 0
- ^* 7 7- = 1.5 ' -
x' V I
/ olume = 500 ft3 3 ,-,' da ,,, %s p3 20 - u6 -
%s~
1 e , g v
u E
' s,< f"=3.0 y 100 - E.5 -
l E' = 5.0 o >. '
'- N
" 8 500 SG tubes h,___ - - -
5 80 g4 -
Volume = 250 f t 3 nt Volume3 = ~~__. -
/
~~~-
60 *3 -
50 ft 250 SG tubes g,
~~',,,,,..._
Volume3 = ~
E = 7.5 -- -
40 -
2 -
25 ft -
~~~~~__ _____ ___ _ -- --
f-- Volu m e = 10 0 f t s
,, UMCP C ='- - - -
20 -
1
~,
-? #C~~ - I ___ - _I 100 SG tubes
~~
- [ ------ - , k=10 ,
0- 0 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 MIST 0.0 0.1 Length ratio I I I I I I l I I I I
.447 .548 .632 .707 .795 .837 .894 .949 1.0
.31 6 Velocity or time ratio I I I I I i 1 1 I I I 2.236 1.826 1.581 1.41 4 1.291 1.195 1.118 1.054 1.0 3.162 Heat flux ratio , ,, ,,3 3_3 ,
l o o o 1
i
. SELECTED SYSTEM PARAMETERS FOR EXAMPLE i
o FULL PRESSURE WATER l 0 PRIMARY SYSTEM VOLUME <250 FT3
! o NUMBER OF TUBES IN EACH STEAM
! GENERATOR <500 l
l 0 LENGTH RATIO / DIAMETER RATIO <2.0 l
l 0 CORE POWER <10 MW o BASE PLANT - OCONEE l
l l
r O -
% G 200 - 10 - . ,
. .; .. ,. ! ., s. .,/ <. ./
/ - . ,. s
/
/
,! :/-?,/ *
..'/-
/.
180 - 9 - gn
. ./ s*si i .. .-f 560'%G
..s, / i ,! .!
/
,' 14tvis'./ , .- . ,<..i
. , ',.. .,-s . <
- 1.0
- i ,/ < //
/. ,
gn ... . .i .! .
. .i . p.
160 - 8 - ...
6-7 'n iM. . ./ ,' / ./ ; . .'.-.
'.. : ..'/. .
Af.~Y.., . / ' tuy,,
&.?,/
l < p/.t !;'g .l . /
. , ,i: .7
~
.ghe'q.fg i
r' o v. s .'- - / / Q' '.i l '.' . / --
- / - / ../i .
14 0 - ^y 7 -
, ' /gg.
- i
/.[J-7'*./ - - h' '.>'. .c?/:-?
9 ' /. ',' . -,-
$61 ping = 500 ft*
1 .g^./ =.R.W n
v .-
.- / ..,'. /. '. .
.'? ~ ~- . ~; s y 120 . - - -
-r ' .
u6 ,- -
.' . -" , - '..'.'*' M..
, : ..?. ' ' :.-/' // . !) .- .-
,2, e ,- ./ .
u s
.4 4, /. . ' ,,. f'd y.n.' , .
- E.5 i/, tt'j
'. s, ./ i.' - s y 100 - -
/ .
.5 0 a >. ..- ..'. , -< .! . ' - . ..' ./ Hs~/ . ..
- '50~0:5.G ' tub.e~s..
a o '.. ' '
/
Q.'m i . m- . ' .
-3 80 - o
- u. a e4 -
- m. -"'M
. -.? -
- i/
, s . a- ,' -
.- : i. ,- .
. . .' /.-?
= 250 f t 3 M Volume = __ i.4 ~ - -80lb.pte - - ' "' '
60 o3 -
50 f ts ,..... -
250 SG tubes tn Volume3 = '
~~
E = 7.5 - --
40 -
2 -
25 f t --
- ~~~~........ --- --- - -
UMCP
=r- ~~~~~
f olume V
.[,
= 100 ft 3
20 -
1 M _
t m..
7_____
M --- w ~ -
%g = 10,N, t
.0 100 SG tub *s O- 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 MIST Length ratio I I I I I l l l i I I
.31 6 .447 .548 .632 .707 .795 .837 .894 .949 1.0 Velocity or time ratio I l l l l l l l i I I 3.162 2.236 1.826 1.581 1.41 4 1.291 1.195 1.118 1.054 1.0 Heat flux ratio WRR 8615-1 B
J
- s. . - ,
200 - to - <, '
../ g...s. - " s ./ /
..~/*~ .
. ' .l<
180 -
9 - .
. k. ' .-.- / '",/s
/-
h=1.0 , 7
.n~// /
, </ <
. / i ./ s/ .i. , .
.,Q ' . /.
- -7c/.+~
.. ~
g r.. , ? /.~- /././ ./ .
./ .
160 - _
ctuue,
'g.' ', %.' y6.r:*s',.i' %)s' ./ .pnza'&?./
./
. .ec:'. .- ~.g ' g. %~. :. .,s. %',z,s/ / .- .
/ '\/
y0 - my . . - - .
500 g
':f,s./:.:.
C.
,,':.r , ;./. - .
n g .. -
/
W,.,
- re
,s-7G/_ ,.
?
t120
- e
. ~.. . ;.,s-y / ' <,h,j A.J.'0'
- /
/ .
-' . / .'
g
,* ?
g ~~Q',
, 't; a.!g"g - s.- '
n 100
- a. 5 .
,'.M..Q...
g _
y
< . s .
f , 50D 5 0't d ~
ha . .
. ' . . w. . e.s '
~.Y7 E. .n s . -
.?$
U ..~.'- -
M-a gQ ~age4 - .
.. ., ' ' ' ./ /
A 4. \.'/'. y ? -i
,,v (.'~'s., 5 ' N-...S m' i d/
- /
,'..- / . 250 f t' N3 j
. . s~
ns.
~M//
s ' - %
Volume = ' ..
60 m3 _
50 ft" ..' .. ' '. ' ... . '-
250 SG tubes
- ' - Ia g;" = 7'5 40 -
2 -
YI 2$* s -b '* 3 ..- '~
--- --- -- --- - -----,- - --- _I"m2 1 0
~
UMCP -_
-- - ~
20 -
..= ?h-1 ,
-- ~
- _ _ _ _% t M- ---- " **
u ,
i , ' ' '
0- 0 o,g 0.0 oj 0.2 0.3 o,4 05 0.8 1.0 MIST Lengt ratio I i ; ' I I l
'9 .837 .894 .g49 1.0 e ocity f me rat o I ; >
3.162 2.236 1.826 1 581 1 195 1.118 1.054 j,o
, Heat fi ratio WRR8615-1 C
J 00 ~ 'O ~
/,.// ///,/ /-f/' ,////] // /.: /,i//,/ _
" / .- '/// '
/,
18 0 - 9 -
t / / y/ / // .67'/ //./ / / ,i .'/ / / / ./.
/ / /
/ .: / / // / - /< / ~i . // 7/ ,.fSGo%g
/ .< / // /.
/ / .'. 14lg if*' = 1.0//,//./ -f 9/// / / / / / / / /, /. .'//-",/..- .. s.p.-5 '.i ///.'#-
160 -
8 -
7 ,/ ,c/,./ .. ,i.%. /i // ,./ . / // // ./ ,'/ :/;.
,i , ,
.. . / <
/
l , f2/ ,/ .:.' ,/ / ' '[y.z!'h k. / $.' <' / ,,hf ~<~' // [.4,h IS' tubes n7
/ / / i 7 m./[.e'[/h'/ {l/,.'. /7 14 0 -
, /-
l n $ ,.
' ,-' h h.'1h',//,l ' '.[/. .:'.'.l.~.?./ Y N.,i ;
/ / ,l /.h*'&,'y500ft8
- - u6 v .
<, '/ ./ '. . '. :.///~ .<g
' / ' <'
</
1 120 '/ - '< .-
i .- ! '/ i i . .
/ .'/ ..' ./ /' ,/ /
i /..'/ /'/ . , /,a/r:/ i ./ f Ax/o / '. / / '../ ////
i w e - : . / . .
s i
. n -;
y 100 - E.5 . \ '.' i .. -*.*'.. / /// Y 'i i ./ . ,. . / i s',. - /
.- _'t ' l5'.9.//.
a a.
- g.4
.>(, ~....
x...'. .. m;~ .'/
///i, .< ' /,./ ' igo'n.fy'y.'
i A j_;f;.:
/ . / . ./l'./
. - /.
g 80 a
. ,., ' ' . 'N .
...~
. s., ..N.
\ . .. ~ , .' ' / .
// / / ./ .. / /b. ,. e./ ',
- ! i s., -
i//
w .. ..' ., .
5
. N .' ' . [ . ' ~ ~'.. .250-SG.J.0Bss ' , .[ \. \ y~. $ h ': * . %; .!$.4 volume8 = ' .[' ,. s. N . \ 0@,N 9 $ 250 fi 60 g3 -
50 ft .
s.
,'s.
N 's .
~Z"' ' J..-
N ' . ;,_'.. ' . .'. .'#'.s,.':.d '-f -
' ' - . . .7~h~~= .' 2 'f~'--- -
40 -
2 - Volume =
.'-1- *
~~
. ~ ' ' -' ~ - . ' ~ .
'-L 25 ft8 .
2'
~ ~ ' ' ' ' ~ - . ' . '--' - ~ - - .' ' ' ' _' l ' '.'::= ~ . i'."'"-"':
~ -- : :-..T' f V610mej 100 fta
- (g..~.. .'_.'ry ~.-., -, ~ . . " . ' '-
_.','y~ ~. . ..
~_. . X,-
4
~
20 -
1 3 '-- ---.
~g. - ' .~.. ,
... f~ - . , .. .. s ~ = -- -
.,,'-~%'E g t" e' ~g gg400
-=-- - ;_
_'~ _ . ..~.w.- _-- m . C ~~- 'ASG tubes
= - ' - ' ~ ~ ' - #- '
0- 0 i
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 MIST i Length ratio l I I I I I I I I I I
.31 6 .447 .548 .632 .707 .795 .837 .894 .949 1.0 Velocity or time ratio i l I I I I I i l l I I
- 3.162 2.236 1.826 1.581 1.41 4 1.291 1.195 1.118 1.054 1.0
! Heat flux ratio ,,,,,,,_,,
200 -
10 -
. .;// ,s,-f. i . /. :/ . .: / . . / ,' s
.Q gh * ,.' ; i.'.' / ,,. , * ! ;
t' /*/ 4' ,"s ,.1 l
/ <//./ ./,/ -
180 -
9 -
i,. . x/ / ! ,.6.v. / i.i, i / ..s ':..?
- < s <
= 1.0 ,. ./ / %. s?,, ./ ' < ' . - ./ : .' / ., f ' ' // .'./,/ .
. . f '
i . .. . .
. / 6 ?', .
' .: .//.'.i './ ':i :
. ' . 4/ .?
N r; / /
160 8
/' // .'i' /: ./ -?i. '.</ ./ X000v$'G' tubes i . /
/ /<
- <Q;/g
<. / : .:. /
i:/, 7 /i' <..(5 .
si q .
/ - .' <' .'s ..// .
s //. '.>w 14 0 ^y 7
-f .//* sX.p' su V.! a
. . : i ;.
1 ,.- ,
- '. '. s'./;Q.f .w/ / -1' : .:
- y . % Q * *. _. ' No1 pin.i.= 500 f t'
. ,.'.~ l . ,.
n , . . .
v y1 110 ~g6 -
,/,',. ! . .
./ .?., / , ', . .-
- ' : : .- . .: .. ' : / / ;-
E / * ./:; . . - .,
f'.se3.0' / ,'
u '
'///
y 100 a
a.
- E.5 n
gy
. ' .% s..:. ' . ' . ' -
~. .;..c..
~
.~.. .
/ /
m'
./
- / /! .'
/;s. .
'/ / .
/:
'. r
i.:/[t'=i .5.0 .' /
.i : i.
. . /a s t
/s '/. ' '/ '.: ... ./ '/- +/.-?
./,/ .GD'0 $.G.t. ub.is
=s 80 - u ..- . . . ' ....,~.'.M /
- '- . ' </< /
e4 -
i/ . p.- c.
a 4 250 fi 8 p
' . .--' ~
M Volume = .' .- s
./l. ' '.1/
A 6.'
60 *3 -
50fts .,.,'. ..>-<.s.
~...w.'.-y' .' .~ . _~~250-SG'.J.Ubss- .
~. *7~'~~L~
. .,~
.s.
3 .
_~~~
. ~
.,~.... .. .
~ . _ ' ~:.'~- - ... -. .__'.~.::'_......-. .
.~
- .' --~... . .
- Volume, =
40
'.,.,,...~i.~.~..~..
2 .1 . ~
25 t# . . . ... - -
~~~ .
,,:':;d' :-- ..g =~. .. - --
" '. C -. _ .. ~v W.!am+5 100 ff*
20 g ~~ g, ;;
.;. g y? A 1 _. ;,-
_[___.[ ~~~._~X.
_ -[2 .h*t0
', M, M :--- -
] _
4OOgG tubes 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 MIST Length rollo I I I I l l l I i l l
.31 6 .447 .548 .632 .707 .795 .837 .894 .949 1.0 Velocity or time ratio I I I I I I I I l I I 3.162 2.236 1.826 1.581 1.41 4 1.291 1.195 1.118 1.054 1.0 Heat flux rotto wanssis-t o
O~
O O
" WORKING CONCEPT" ;
i WILL PROVIDE BASIS FOR
. o ADDITIONAL ANALYSIS 0 NRC AND INDUSTRY CRITIQUE l
0 COST ESTIMATING l
o CONCEPTUAL DESIGN I
1 l
O~
~
O O BASIS FOR DECISION
! o FACILITY OBJECTIVES 4
MODEL/ CODE DEVELOPMENT l
MODEL/ CODE ASSESSMENT l
TRANSIENT IDENTIFICATION i --
TRANSIENT SIMULATION RECOVERY PROCEDURES DEMONSTRATION OPERATOR TRAINING RESPONSE TO PLANT EVENTS l 0 CONCEPT ABILITY TO SIMULATE IMPORTANT PHENOMENA o CURRENTLY AVAILABLE DATA BASE o AVAILABLE FACILITIES o WEAKNESSES IN PAST OR CURRENT FACILITIES o AVAILABLE DOLLARS o AREAS OF GREATEST CODE UNCERTAINTY m- e
i o o o-1 l
I PROPOSED WORKING CONCEPT o FULL PRESSURE WATER I o LENGTH RATIO = .375 I
0 LENGTH RATIO / DIAMETER l! RATIO = 2.0 o SCALED DECAY POWER 1
j o B&W OCONEE PLANT i
I .. __ _ _ _ _ _ _ _ _ _ _ _ _ _
O O O' Scaled Parameters for B&W LPWR (Full Pressure Water) 200 - 10 -
\ ',-
,.- K,. 1500 SG tubes 18 0 -
9 - \y ' ,,, -
1 b=1.0 da s ...
! 160 -
8 -
N ...
! ,-,' N 1000 SG tubes 14 0 ka-2 w
'# A=1.5 Volume = 500 ft 8 il ^* 7 - ,-
2 ,-,' da ~~s p0 j
2 12 - m6 e ER0EOSED.
'.. .. g
~~%
v f
! _ WORKING _CONCEEI . -fi = 3 0 la j
y' 100 - a5 o >.
' ,x - N
% = 5.0 5
80 - 8ae4 -
- s N
% 500 _...
SG tubesh___.
.! w
- N__ V l M Volume = '_.. / olume
= 250 ft*
'a 3 250 SG tubes 60 -
50 ft* ,__. t, da = 7.5 Volume = -
40 -
2 -
25 f t* .- -
,,________..____- -- -- - -- - 8
! , _ j uMCP _____._.
- " _ _ _ _ _ _ f olumeV = 100 ft ,
l 20 -
1
__ _-7j ~ ~ _ - - g __I.
M --- w
, Lt =10,N t m ._____-- % 4
.0 100 SG tubes <
-p , , ,
0- 0 j 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 MIST i
Length ratio i
I I I I I I t
i I I I I
.31 6 .447 .548 .632 .707 .795 .837 .894 .949 1.0 l Velocity or time ratio i
i I I I I I I I l l 1 1 l
i 3.162 2.236 1.826 1.581 1.41 4 1.291 1.195 1.118 1.054 1.0 i
! Heat flux ratio wanss15-1A i
l
. 3 O~
~
O O 1
l l
i FULL PRESSURE WATER 1
i o MINIMIZES ANALYSIS REQUIRED TO DRAW i INFERENCE FROM MODEL DATA TO PLANT i
l 0 PROVIDES BEST SIMULATION OF LOCAL
) PHENOMENA I
J o PROVIDES UNDISTORTED TRANSITION FROM SINGLE-PHASE TO TWO-PHASE OPERATION i
l
~
O O O
- Summary of Babcock and Wilcox Scaling Concept Rankings l
4 i
l Facility Overall Rank Percent Top Rated Percent Bottom Ratec FHFPW - Ideal 2.06 57.7 42.7 4
FHFPW - Actual 1.72 63.9 23.6 RHRPW 2.35 20.6 56.2
)
RHFPW 1.86 38.1 23.6 LFPW 1.87 49.5 38.2 i
RHFPF 1.94 39.2 29.2 CCK00 01 i
O O O~
! LENGTH RATIO / DIAMETER RATIO l
l 0 CURRENT INTEGRAL DATA ARE FROM ONE- .
, DIMENSIONAL FACILITIES l
l l LOFT -
LR/DR~ 7.5 1 SEMISCALE - LR/DR = 41.3 ROSA-IV -
LR/DR= 6.9 MIST - L R/DR = 28.6 o MULTI-DIMENSIONAL MODELS ARE BEING INCLUDED IN CODES.
o IDEALLY, LR/DR = 1 FOR BEST SIMULATION OF MULTI-DIMENSIONAL EFFECTS s
O O O LINEAR SCALING o WITH 10 MW POWER CEILING MAX LR = .10 FOR FULL POWER SCALING l CONCERNS WITH 'R = .10
! o HEAT FLUX RATIO = 3.16 i
0 TIME RATIO = .32 i
! o QUESTIONABLE DRIVING HEAD FOR NC 0 T00 SHORT FOR PHENOMENA DEVELOPMENT l
o T00 CLOSELY COUPLED FOR FLOW l INSTRUMENTATION o DIFFICULT LIQUID LEVEL MEASUREMENTS l
l
l O O O~
! CONSENSUS OF CONSULTANTS I
l I o APPROACH LINEAR SCALING l LR/DR = 2.0 MAX l
l i o BIGGER IS BETTER i
l o GIVE UP FULL POWER FOR SIZE l
l
! O O O i ;
l -
l ~. :. -
CONSIDERATIONS ' '
i -
4 s~
! APPRl)XIMATELY 3 MW ARE AVAILABLE AT '
SEMISCALE WITHOUT ADDITIONAL UPGRADES 4
) .'
i OPTIONS o LR = .2 WITH LR/DR = 1.0
)
o LR = .3 WITH LR/DR = 1.5 ,
i o LR = .375 WITH LR /DR = 2.0 l
1
- O O O i
! WORKING CONCEPT PARAMETERS l
! PRIMARY SYSTEM PRESSURE 2170 PSI
! SECONDARY SYSTEM PRESSURE 925 PSI i PRIMARY SYSTEM VOLUME 150.2 FT3 i HOT LEG DIAMETER 6.75 IN.
! COLD LEG DIAMETER 5.25 IN.
! HEATER R0D LENGTH 54 IN.
CORE POWER 2.76 MW
- VESSEL INSIDE DIAMETER 32.0 IN.
! DOWNCOMER GAP WIDTH 1.82 IN.
I NO. OF SG TUBES /SG 546 i SG TUBE LENGTH 19.54 FT l SG SHELL ID 25.86 IN.
.6124 I VELOCITY RATIO l TIME RATIO .6124 HEAT FLUX RATIO 1.633 l
1
COMMENTS BY SCALING REVIEW GROUP i
o METHOD OF WEIGHTING PHENOMENA IN SCALING REPORT HAD NO j EFFECT ON RANK OF SCALING CONCEPTS o LINEAR SCALING CONCEPT CONCEPT WILL BEST SATISFY DATA NEEDS DUE TO GE0 METRICAL SIMILITUDE 1/10 LINEAR SCALE PHYSICALLY T00 SMALL TIME DISTORTIONS LACK OF NC DRIVING HEAD l
EXCESSIVE CORE HEAT FLUX INSTRUMENTATION ,
l 0 FULL SCALED POWER VS. SIZE I --
FULL POWER REQUIREMENTS PUT SEVERE CONSTRAINTS ON
! DESIGN l --
ONLY CERTAIN TRANSIENTS (LBLOCA) REQUIRE FULL POWER l
l LARGER SIZE FACILITY OPERATING AT DECAY HEAT WILL
- PRODUCE DESIRED PHENOMENA 1
l O O O'
3' TRAC VOD' A4J MOJ2 l A 3R SEbTATON TO THE -
, BY : D b s S ULES 1
LOS ALAMOS sdTONAL LABORATORY
O O O
- 03m EC- VES
~~
VOJ' VA \ ENANCE WA NTAN A TROZEs VERSON 07 T-E RAC37'/kO31COJEsSU3POROr j ,
E USs RC 3ROGRAVS - CAP,23/3J, V S , ROSA- V, A 33Es D X <, E C.
l l ACCE3- A3_E C- ANGES TO l- E rROZEs VERSON Or VOJ1 ERROR CORREC-~OsS l USER CONVEslEbCE FEATURES s W kOJELS THAT ARE 03 ONS LosAlamos
- o. o o- -
S~A LS O SH 3RO3_EVS
- A33ROX VK E_Y 700 SE3ARATE USER COsTAC- S s 3AS-~ 17 V O N FIS.
- 97.1% 07 USER 3RO3LEvlS RESOLVED.
- 2.9% 07 USER 3RO3_ EMS Us RESOLVEJ.
- Us RESO_VEJ 3RO3_EWS ARE E F ER CURRENTLY Us JER NVES~~GK ON OR WAF NG TOR VORE Is TORk ATOs.
- VOST SGs TCANT USER COslCERNS ARE 3
CK ALOGUEJ s O COV U _R JK A 3ASE.
(CURRENTLY 270 " TROUBLE 70RWS")
LosAlamos
VOJ' V A \ ~~EN ANCE
- ~ 70 SE3ARATE U3]K E JESTS FAVE 3EEb
~~
DEVE_03ED TOR V A \ EslAbCE Or k031.
l
- TF ESE U3] ATE JESTS REPRESEs- OVER 22,500 L \ ES 07 U3JK ES, COk V Eb- CARJS, b EW l 0R v0J IIED FORTRAsl S A EV EslTS. l
- TF ESE U3JATES TO kO31 sCLUJE :
ERROR CORRECTOsS (25%)
~
2.
- 3. bUSER COslVEN W kOJELS OR 03 TOSS ENCE (20%,) (30%)
- 4. COJE CLEAsU3 (10%) '
- 5. W 3 ROVED PORTAB _TY (15%)
LosAlamos
O~
~
O O
-~
RECEN v 3ROVEV Es- S 3
~~O ~~RAC '/VO J' SELF-INITIAUZATION MULTIPLE SOURCE TERMS IN THE 3-D VESSEL
! CCFL MODEL SEPARATOR MODEL MOD 2 l (IN ADDITION) 3D-2 STEP NUMERICS WITH SOME VECTORIZATION IN 3-D
! IMPROVED CORE VOID MODEL LosAlamos
i o o o ~
SE_ N ~~ A _ ZA-~ O N A STANDARD SET OF CONTROLLERS WHICH MAKE IT EASY TO RUN THE CODE AND ACHIEVE A GIVEN STEADY S~~ ATE.
j MULTIPLE SOURCE TERMS PERMITS SEVERAL PIPES TO BE CONNECTED TO A GIVEN 3-D MESH CEL_ IN THE VESSEL LosAlamos
~ '
O O O j CC _v0]E_
SUMMARY
- CCFL MODEL IMPLEMENTED INTO TRAC-PF1/ MOD 1 WITH EC 12.8. -
- ASSESSMENT AGAINST SATURATED FLOODING DATA i SHOWS EXCFI I FNT AGREEMENT WITH THE DATA.
- ASSESSMENT AGAINST SUBCOOLED FLOODING DATA
- ALSO SHOWS VERY GOOD AGREEMENT. CORRELATION i CONSTANTS DEVELOPED FROM DATA FOR SATURATED CONDITIONS CAN BE USED.
- REVIEWED IN GERMANY (AUGUST 1986)
I
- GRS MADE A PRESENTATION AT OCTOBER 1986 COORDINATION MEETING, SHOWING GOOD l AGREEMENT WITH KARLSTEIN DATA.
LosAlamos
~ ~
.SER REQ REVEN S CORRELE04 DESCR 3 sG -E COUNIa CURREs- 7 OW _
3EF AVIOR TOR A S3EC IC GEOV ETRY Or s IERES- .
- CORRELATION ASSUV EJ ~O BE Or TF E FORM l Hg N+m*i@=C l
L W- ER E, H=J* O R <*, JE3EsD NG Os ~F E FACTOR 3 1
- USER SU3P_ ES V = SLO 3E 07 LOOJ NG CURVE, C = A3SCSSA N ERCE3-~ l 3 = FACTOR FROV 0 - O ' l l
- TF S A__OWS USER TO SPEC Y E~HER J* OR <*
SCA_ sG OR ~O USE AN s I .NEDIATE SCALibG JESCRI3EJ BY 3ANKO F.
LosAlamos
O MODEL ASSESSMENT
SUMMARY
-Bankoff, Tankin, Yuen, Hsieh " Counter Current Flow Of Air / Water And Steam / Water Through A Horizontal Perforated Plate", Int. J. Of Heat And Mass Transfer, Vol.24, No.8,1981
-15 Hole P!6te 0.0105m Diameter Saturated Water and Steam injection Pressure = 1 Bar
-Boundary Conditions
+ Initial Flow Above Flooding Point W8 input Steam Flow O
W: Input Liquid FIow time
-TRAC Model I puf Il ut e Lev m o
... Tie-plate
[1 (1 Steam Input O _
LosAlamos
O O O'
\
s EC-~Os VOJ7 ASSESSV Es~~-SU3 COO FD LQL J
- Some faciity as wi", so urated conciions a inau" floocing correaion cons" ants are 11e same as usec for satura"ec concitions since
~~RAC cacuates "1e e"fec~ o" concensaion directly
. Boundary conditions:
1 Bar '
~l = 281 <
Three licuid in'ecion ra"es Seam 'ow ramaec sepwise
- Weea and duma poin"s caserved LosAlamos
O '
o o- -
Assessmern 0:. Ehe CC7L Yocs.
For Saturated Conctions i i i 2.0 i i
%a x
x c
N
- -n 1.5 g l
m e
P-i 2
e 0
- l 1.0 co C 0 TRAC-PF1/ Modi 12.8 2e o o nata a
D A A TRAC-PF1/ Modi 12.6 i i i i i 0.5 1.2 1.0 0.8 0.6 0.4 0.2 0.0 Downwarc. :Liquic. Yass Flux, : Y
o O i i
.e i i i i i i i -
4 -
O e o
,y
.$ '5 0 4 ,
g4 4 d
@ h g
'O
- o$
N O - aa ha 4 0 e
^H4 w ~
j ~$ $ 4 oQ h
~
4
~N co
-eH xx 0 C 33 4 -
o Il 0 ll <a e .9 e
O p- 4 II - O "O
Ol.L <' <e e o si <o , y e _
g a#
U . t g _
g .-
t
- v ll a -
a om jg 40 4 O
0
.E u
<a e - oD e 0 o< o
- 4 e .E
)3 l M & O e4 8
\ e a$ <p o
<o
.4
- o P
~a o <1 e4 <a 0 h
OA
<K1 O
i i i i i 90 i i i o O O O O O o (gy)81y8 g 'uoqoa[ul umaqs
O CCFL MODEL ASSESSMENT AGAINST KARLSTEIN DATA SATURATED CONDITIONS AT 4 BAR H. Glaeser,GRS-Munich
~
1.0 i 4
g 0.9 -
e o DATA -
TRAC, SINGLE CHANNEL INPUT, i
0.8 - NO CCFL MODEL _
cn C 3 TRAC, SINGLE CHANNEL INPUT, d WITH CCFL MODEL g o.7 F.
- O s 0.8 -
! y -
- - - - .- -- - - -. _ _ _ _ _ _ __j
- O 0.5 -
l J l 1 I g 0.4 -
l l
2
<c 0.3 -
w l
$ 0.2 - -
0.1 -
1 0.0 1.0 2.0 O ' WATER DOWNFLOW RATE (kg/s)]'##
o o o~
SE3AM~OR COV 3 ON Es~~
l Separates incoming stream into 2 streams of user
! specified qualities 1
n i Motivated by the needs of modeling UPTF
- steam / water separators LANL mandate to use INEL TRAC-BWR separator model a This model received peer review and testing at INEL j
Final version added to TRAC-MOD 1 on March 25,1987
! offer testing at Los Alamos
! L.osAlamos
O O O CALCU_ATION 0 CARRYOVER A43 CARRYUbDER OPTION 1:
Ideal separator - constant values of XCO and XCU user input or default values. .
OPTION 2:
Mechanistic separator model. The performance of the separator is determined from present inlet a model developed by GE inlet conditions for their style steamusing/ water separators.
Geometry dependent. can use default input or user specified geometric input.
OPTION 3:
! User supplied performance data, supplied in tabular form, of XCO = f(mass flow rate, quality) i l LosAlamos
l COE VO 3 bOJEL l
INTRODUCTION / PURPOSE l
l
- TRAC-PF1/ MOD 1 results typically show too much water l accumulation in the bottom of the core, and very little water above the quench front. The amount of l
entrained liquid is under predicted.
1
- The MOD 1 core interfacial shear model employs an I entrainment sharpener which restricts the amount of liquid leaving the quench front.
- The purpose of our development effort for MOD 2:
1))To eliminate the interface sharpener 2 To improve during reflood
- 3) To improve the prediction of the void fraction throughout the core
- 4) To investigate several correlations or models in attempting 1, 2, and 3 LOSAlamos r i e
lO RESJLTS
- During the development effort, four different l models were investigated:
, Williams /Liles model Anderson /1shii model (TRAC-BWR)
Murao/lguchi model ref: " Experimental Modeling of Core Hydrodynamics during Reflood Phase of LOCA", J of Nuc. Sci
- and Tech,19(8),pp.613-627, Aug.1982.
! 4) EPRI model
- ref: "A Full-Range Drift-Flux Correlation for
- Vertical Flows (revision 1)", EPRI NP-3989
-SR, Sept.1986. Order from Research Reports
'O
- center, Box 50490, Palo Alto, Calif. 94303, j (415) 965-4081 l
- The EPRI model was chosen for final implementation.
l The model has previously been assessed against a wide range of pressures, flows and geometries. It also provided i
very good results for the CCTF Run 14 test case.
- The EPRI model was used to estimate the interfacid shear explicitly, no iteration was required.
l l
!O
! LosAlamos :
L_..--.--_-- - . _ - . - - - - ~ _ .- -
CCTF RUN-14 CORE DELTA-P COMPARISON TRAC-PF1/ MOD 1 vs. MOD 2 O o.s1-1.22 m s000 . . . . . . . .
5000- .. _ ___ _;._--
4000- -
. .s 1:;
n 'ti'i n b eyn. M ::. i'i;f.
3000- ..' - T
- g ,,..../
. : 2000-
- . /, - MOD 1
, 3 ..-
1000- Jy -
!?
0! '--*---t- -
-200 O
0 2S 50 75 too 12 5 15 0 17S 200 22S TIME (s)
d at a TRAC s000 . . . . . . . .
S000- -
4000- -
, fi.?ite!'.~j'*\i'i'.!'!)? .
' ~
l ,,
I MOD 2 l : .-
J@
p 2000- -
0 .' l c -
1000- Jy -
l :
i :
I 0'
t- 3 -
O -1000 , , , , , , , ,
0 2S 50 75 10 0 12 5 15 0 175 200 22S Tirn. (s) 102 AlBTl)D2
CCTF RUN-14 CORE DELTA-P COMPARISON TRAC-PF1/ MOD 1 vs. MOD 2 O 1.22 - 1.83 m 5000 , , , , , , , ,
i l
4000- -
1 3000- - -
e 6
$ 2000- -
I
.. h MOD 1
"'- ^"~ ~
0n.......e......=...' -
-1000 , , , , , , , .
0 25 SO 7S 10 0 12 5 15 0 f75 200 225 data
- Twt (s)
TRAC 5000 , , , , , , , ,
4000- .
3000- .
Z- ,
g 2000- .
MOD 2
\
E 1000- ..***
n .
0, i...........,4,,,, .
1O --<
0 2S 50 75 10 0 12 S 15 0 175 200 225 b93 bbMDb
-w- -rma~e- - -- w-- , -w.,,,w,s -, - ec wy-- - - - - , , - - - - - - -------,--------9-------------,,- - - - - -
- y. - + -
CCTF RUN-14 CORE DELTA-P COMPARISON TRAC-PF1/ MOD 1 vs. MOD 2 0 1.83 - 2.44 m :
i
, 2500 , , , , , , , ,
I i
2000- -
1500- .
e e, -
h 1000- -
E , d MOD 1 300 ,
. r / a. v'_
E
.I O: - -- :- -":"' : -
O -%0 V ,
4 , , , , , , ,
- O 23 30 73 iOO us e0 ns 200 22s - - - - -
data i
WE(s) TRAC 2500 , , , , , , , ,
I 2000 . .
1500- -
Z l
2000 .
MOD 2
,00
, ,,. ... ~- ^ ,~,,,.%/=A.dy.
q O'
^
O -20 ,
0 25 50 75 10 0 12 5 15 0 US 200 225 bDD .e- bll)!.)b
CCTF RUN-14 CORE DELTA-P COMPARISON TRAC-PF1/ MOD 1 vs. MOD 2 0 2.44 - 3.o5 m
. =0 , , , , , , , ,
800-
.:.)*A;d .
- :( p. . .
s00- . .:
., ,s..
1 -
400- l .
& l J MOD 1 NO- *
. i
.V 0 : :
8
- - : w .
,, .................?
i
-200 ,
O 25 SO IS ido 1$$ 15 0 NS 250 22S Twt (s) ..... data 5 00 , , , , , , , ,
TRAC 800- .
. /4,=1. .***
e00- .., .'
t
- ,.o -
400- [ ,. .
l ./ MOD 2 200~ '
')
0, -
0 .
i
-200- .
O O 2'S S0 IS 15 0 t$5 15 0 NS 250 225
- I') L D3 AltrijDB i
\
l CCTF RUN-14 CORE DELTA-P COMPARISON TRAC-PF1/ MOD 1 vs. MOD 2 i
O 3.os . 3.se m M00 , , , , , , , ,
e00- .:
fa.M! .s-
- .n '
,,...,,./
600- . ,,.*
,,.o
$ 400- f t
= :
. u
- i MOD 1 8
I NO - : -
t-
.' =- = = ~'-
0u. . . . . ; . . . . . s. . =
O -200 0 25 SO 75 10 0 12 5 15 0 f75 200 225
, Twt (s) ------- data TRAC M00 , , , , , , , ,
800-
...f w ,qi"
.c -
=
ls':*
s00- ,,
..~-
[ 400-i
[ -
MOD 2 i
f - .
0, i............. ...d -
~200- .
O -.. . , . . . . . .
0 2S 50 7S 10 0 125 ISO f75 200 225 DD $} M b b
' ~
O N .O EPRI CORE IF SHEAR MODEL, TRAC-PF1 MOD 2 -
ROD AT A Z ELEVATION = 0.61 AND 1.105 800 , , , ,
750- -
. o TE22Y11 700- ; .
l \, a CTE22Y11
~.
650- ,
, + TE22Y12
^ f f 6 l :
8 600_. j j _
x 2 W12 2 : :
e.
- n. 550-
- E / :
e .l9 :
, i, 500-- . / l :
-l '
450- ! ,
,e
.i --. --_ -' - t
- 0
. 400, 350 , , , ,
0 10 0 200 300 400 500
. Time (s) LOS ALAMOS
S V V ARY i
- New interfacial shear model for the core region has been implemented in TRAC-PF1/ MOD 2
)
- Assessment against CCTF Run 14 data complete i e Significant improvement in the data comparison has i been achieved
- Comparable transients were calculated about five il times faster with MOD 2 than with MOD 1 l
- Future assessment will include CCTF runs 54,77, and j LOFT large and small break LOCAs i
- Final results to be presented at the upcoming
!l Coordination meeting i LosAlamos
O O O' l HBR MOD 1/ MOD 2 Comparison SGTR RJ-4/10
- TOTAL CRAY CPU TIME 4000 . . , , , i 1
i-
-500 3500- -
1 MOD 1 1 '
MOD 2 j 3000- -
- 400
?
I
]
e E
2500- -
?o I .
F %
i 3
-300 3 O 2000- -
g
{
1 O
! .0., o
.O {
1500 - -
-200 0 l 1000- ,- '~ -
1 -
[ -10 0 500- - - ~~~~~~~~ -
,e m* ____ ....-
,ye e6--
0 , , , , , , 0 0 50 100 15 0 200 250 300 350 2 TIME (s) l
II j f)l,1'?7
\
O WATER HAMMER I
Thirty fluid systems in the four reactor designs have the potential for gater hammer events
\
l
- Ten of the thirty fluid systems perform safety functions a
o Somi interesting observations 4
- subsystems designed to prevent water hammer become an initiator
- O
- severe accident management instructions may lead to circumstances favoring water hammer
- off-normal plant configurations can aggravate potential water hammer initiators
- removal of restraints, pipe whip protection and seismic snubbers will place
! new emphasis on normal piping system load limits I
l l
O t .
y,.-,...-...v -. . , , ,,,,_y. . . - +.y.. ,z., _ , _.-.
O 58 EVENTS DURING THE 1981-1985 PERIOD 804 Resulted in Pipe Support Damage hangers, anchors and snubl>cs 174 Resulted in Component Damage i
piping, pumpw and valves 104 Resulted in Reactor Trip 0.02 scrams /yr/ plant 7% Resulted in Plant Shutdown 23% Caused no Damage l
i I
l l O .
W
- , , _ - _ = - . - - _ _ - . . _ _ . ..- . - ,-. .- --- _.. --- - _ - .-- - - -. -
_ _ _ _ _ . . - - ---_.m _ _ _ - _ _ _
O WATER HABBIER FREQUENCY LER + INPO Data Show 0.2/yr BWRs 0.1/yr PWRs i
Backroom wisdom is that only 1 in 10 water hammer events is reported.
O If a water hammer yields its maximum pressure pulse, a failure will occur.
The statistics of a water hammer are not well defined.
It is not clear that water hammer is properly incorporated into risk studies.
I I
I O
O m -- s - -~ - - -- ,,--.e.- e- ,-- w.- --~----- -- - - - ---------~~,o w-,,-mew --- ~< w , - -m m w ,,m-nn-w<e- - -
vg-w -- - - ---w-- --
-w---
A SURVEY (25 UTILITIES) 30% are addressing water hammer 4% have an engineer assigned 30% incorporate water hammer into operator training 20% have check valve programs in place O 20% look on check valves as active components The NRC Pronouncement That Water Hammer Is Not A l
Safety Issue Has Lead To Utility Lack of Interest t
O .
9 O
. , , - , _ - . _ . . - - - - . - - - , _ _ _ _ .. , _ , . . . , . , . - _ . . ._.--._m., , , __.. _.-
O NUREG-0582 Requires a water hammer to be catostrophic for it to be considered.
Data base is biased towards events that shouldn't happen.
No agreement among system designers as to what should be considered normal or abnormal so water hammer is usually not considered.
O Plant engineers must "fix" many problems during pre-op phase and only those " missed" show up in the data base.
A question for licensing: what is preferable, fixing it after or preventing it?
Is a " design basis water hammer" needed for piping system designers?
O 1
1
(T-
'O Action Prior to How did we get into the event the problern?
pornp start 16 procedure 24 water valve cycled 10 cornponent failure 12 st earn valve cycled 9 inadequate cornponent 3 turbine start 8 operator error 3 unknown 15 unanticipated response 3 unknown 13 Pre-water hararner Mechanisrn causing conditions wat er harorner st earn in water 1ine 17 st earn bubble col 1 apse 8 condensate in st earn line 16 stearn water entraprnent 20 air in water 1ine 10 void 1ine fi11 12 single phase 6 water coluron separation 2 water in stearn line 3 classic water harnrner 6 unknown 6 unknown 10 Overall cause of Correct ive roeasures to per ' harnroer event s* water hararner events vadrai n i st rat i ve 2 procedure change
=__. --------
32 cornponent 11 operator training 8 design 21 rnodify supports 5 operator 12 rnodify controls 2 procedure 23 rnodif y piping 9
. unknown 11 unknown 10
- sorne event s have roult iple causes Water Hararner ever.ts surnrnary SYSTEM BWRs PWRs Total RHR 8 5 13 Feedwater 3 8 11 turbine stearn supply 5 7 12 ECCS 5 3 8 Main stearn supply 3 1 4
BWR RWCU 2 2 l Service water 1 2 3 1
Other 4 58 l
l i .
l ..
}
l'
il l l l O
E L
I I
R I O
F H P
O HS C C I B R S U A Y R S S E L E R S A N C A 7 E N K A N S 8 R A C E L 9 E O M L 1 S A B T O A C C N F O
C O .
U E 9 EL D t I O 2 i
P H FC M N A L 0I A E C D I i
T I S I I R l S L E L P l I L R U A I L I P A D A W R NE E R D
Y P H M
O L
_ C A M
R E
l i
T O
ll
O O O STATUS OF E'CS RUL.E EVISION
- NOTICE OF PROPOSED RULEMAKING PUBLISHED IN FEDERAL REGISTER ON MARCH 3, 1987. PUBLIC C0ftENT PERIOD THROUGH JULY 1, 1987.
- DRAFT REGULATORY GUIDE, "BEST-ESTIMATE CALCULATIONS OF ENERGENCY CORE COOLING SYSTEMS PERFORMANCE," PUBLISHED FOR PUBLIC C0ftENT ON APRIL 2,1987.
- NEARLY COMPLETED DRAFT OF ECCS RESEARCH REPORTS PLACED IN PUBLIC DOCLPENT ROOM ON APRIL 2,1987.
COMPLETED DRAFT FOR C0ltHTT PUBLISIED AS NUREG-1230.
O O O
. COMPE!0llM 0F ECCS RESFJtRCH FOR REALISTIC LOCA ANALYSIS 1
i A AU l FORWARD l ACKNOWLEDGENTS ACRONYMS EXECUTIVE SltNARY '
- 1. INTPEDUCTION
- 2. ECCS LICENSING 2.1 STATEMNT OF PROBLEM 2.2 PURPOSE OF TE ECCS COMPENDIlN 2.3 HISTORICAL PEPSECTIVE 2.4 RECENT ACTIONS TO REVISE LICENSING APPROACH l 2.5 BASIS FOR REVISION OF EXISTING ECCS RULE
! 3.1 AERICAN PHYSICAL SOCIETY REVIEW
, 3.2 EXRRIENTAL FACILITIES AND CAPABILITIES j 3.3 COMPUTATIONAL CAPABILITIES
3.4 CONCLUSION
S AND PROGNOSIS i
l I
O O O
- 4. USNRC RESPONSE - R8D KTl10DOLOGY 4.1 OVERVIEW 4.2 RE0VIPED T/H MODELING CAPABILITIES FOR LWRS 4.3 BEST ESTIfMTE COMPUTER CODES 4.4 CODE SCALING, APPLICABILITY AND UNCERTAINTY EVALUATION ETliODOLOGY 4.5 INSTRif9frATION 4.6 EXPERIMENTAL PROGRAMS 4
- 5. BEST-ESTIMATE CODE DEVEL0ffENT PROGPMS AND RESULTS 5.1 ASSESSENT ElliODOLOGY 5.2 CODE UNCERTAINTY M0 SCALING 5.3 BEST-ESTIMATE CODES
- 6. LOCA/ECCS PHEN 0E NA, PREDICTIVE P0DELS AND DATA BASE 6.1 CRITICAL FLOW 6.2 DEPRESSURIZATION AND BLOWDOWN CORE HEAT TRANSFER 6.3 PWR ECC BYPASS 6.4 FYR REFLOOD 6.5 BWR REFILL 6.6 FLLY BLOCKA&
O O O 6.7 ALTERNATE ECCS 6.8 REACTOR C00LAhT Plf1P EFFECTS 6.9 SMALL PREAK LOCA COE 11EATUP
! 6.10 U-TUBE STEAM ENERATOR PERF0 WANE 6.11 ONT-THROUGH STEAM EEPATOR KRFORMANCE 6.12 DECAY liEAT
. 6.13 KTAL-WATER REACTION 6.14 FUEL R0D PERF0 FAN E 6.15 CONT 00L ROD PERF0Pf1AN&
- 7. PROBABILISTIC RISK ASSESSENT RESULTS A!0 KRSKCTIVES 7.1 INTP0 DUCTION / BACKGROUND ON PPA 7.2 HISTORICAL PERSPECTIVE OF LOCA AND ECCS IMPORTANE IN PRAS 7.3 POTENTIAL PLANT CHANGES AND THEIR POSSIBLE RISK 7.4 EXAMPLE IMPACTS ON PLANT RISK 7.5 POTENTIAL IMPACTS ON ACCIDENT SEQUEN& CONSEQUENCES
- 8. ECCS RULE REVISION 8.1 CONSERVATISMS IN EXISTING ECCS RULE 8.2 EFFECT OF PROPOSED RULE 8.3 PROPOSED RULE BASED ON RESEARDI APPE10lX A - EXKRINNTAL FACILITIES APPENDIX B - EPRI R8D CONTRIBUTIONS TO THE TE0iNICAL BASIS FOR REVISION OF ECCS RULE APPENDIX C - INSTRlKNTATION DEVELOPKNT AND RESULTS
O O O PAJOR REVISIONS TO ECCS COMPENDIUM SINCE.LAST ACRS REVIEW PREVIOUS CHAPTERS 4 AND 7 ("LOCA/ECCS PHENOWNA AND ISSIES" AND "EXPERIENTAL TEST PROCPAMS At0 RESULTS") HAVE BEEN COMBINED IN A NEW CHAPTER 6 KR RECOWENDATIONS OF THE SUBCOPNITTEE.
INTRODUCTORY MATERIAL IN CHAPTERS 1 AND 2 HAS EEEN REORGANIZED.
PREVIOUS CHAPTER 6 ON INSTPL'FBITATION HAS BEEN MADE AN APPENDIX AND REWRITTEN.
CHAPTER 4 ("USNRC KSPONSE - R8D METHODOLOGY") HAS BEEN EXTENSIVELY EDITED AND EXPANDED TO INCLLDE CODE UNCERTAINTY AND SCALING METHODS. CHAPTER 5 (" PEST E5flMATE CODES") HAS A i SECTION PESERVED FOR THE RESULTS OF THE CODE UNCERTAltTIY WOPK ONCE COMPLETED.
APPENDICES CN FACILITIES AND EPRI WORK HAVE BEEN INCLUDED EXTENSIVE EDITIf6 THIOUGHOUT 4
FANY DETAILED COPTBITS OF ACRS CONSULTANTS CONSIDERED.
i
)
i 4
O O O SECIFIC ACRS COMTNTS/ CONCERN _S ADDESSING TEST DATA UNCERTAINTY TIE C0lTENDIlti INCLUDES A SECTION ON CWE UNCERTAINTY AND SCALING, OF WHICH HOW TO INCLUDE TEST DATA UNCERTAINTY IS A PART. EPRI HAS ALSO FUNDED WORK ON OUALIFICATION OF SPECIFIC DATA.
WE PLAN A TASK TO EXTEND THE EPRI NORK TO EXAMINE, TO T11E EXTENT POSSIBLE, THE DATA THAT IS EC00TB'DED IN THE PEGULATORY GUIDE.
"ItTFXING" 0F TIE COPPENDIUM WITH ECf9 RULE AND REGULATORY GUIDE. THERE IS NO C0mDN F0FFAT rHKEEN THE RULE, REG GUIDE AND C0K _m0M. THESE AE DIFFERENT DOCitENTS SERVING RELATED, BlR DIFFEPENT PURPOSES. IT IS VERY IMPORTANT, HOWEVER, THAT TE THREE DOCLPENTS AE CONSISTENT AND DO NOT CMTPADICT EACH OTHER. WE HAVE TRIED TO BE ALERT TO THE POSSIBILITY OF CONTPADICTIONS AND INCONSISENCIES FROM TE START AND HAVE CORRECTED ANY THAT HAVE BEEN F0lflD. WE WELCLTE BEING ALERTED TO ANY PROBLEMS THAT WE MAY HAVE MISSED.
PLANS FOR ADDITIONAL PEER REVIEW. TE DOClFENT WILL BE FORMALLY RELEASED FCR PUBLIC CCMENT APO lE EXPECT THAT INTERESTED GROUPS SUCH AS TE ACRS, EPRI, AND TE VENDORS WILL PROVIDE SIGNIFICANT REVIEW, E WILL ALSO FORMALLY REQUEST REVIEW FROM T E ICAP MEMBERS. CONSIDEPATION IS ALSO BEING MADE FOR REQUESTING A FORMAL REVIEW FROM AN INDEPENDENT GROUP.
4
O O O CONCLUSIONS AND PROGNOSIS
- THE EFORT IS THREE INCKS Tl!ICK (SINGLE SPACED, TFO-SIDED COPIES), CONSISTS OF ABOUT 1300 SINGLE SPACED TYPED PAGES AND FIGURES, WAS PEPARED USING 1500-2000 HOURS OF NRC STAFF TI E AND $500K l 0F LAB SUPFORT. CONCLUSION - CUANTITY IS SUFFICIENT.
REPORT WAS WRITTEN BY A LARGE NLNBER OF AlIIliORS AND EDITED BY SEVERAL DIFFERENT E0PLE. AT TI ES Tile PEPORT liAS SUFFERED FROM THE "TOO MANY COOKS" SYNDRGE AND HAS BEEN TOO LAPGE FOR A SINGLE ERSON TO GIVE TliE NECESSARY ATENTION IT NEEDS. WE IIAVE TRIED TO REK DY llilS PROBLEM AND BELIEVE WE IIAVE BEEN SOEM1AT SUCCESSFUL.
- TliE PRODUCT TilAT YOU HAVE IS FAR FROM KRFECT, BlIT WE BELIEVE IT PEPRESENTS A SIGNIFICANT COMPILATION OF VERY INCRTANT INFORMATION AND IS PEADY TO BE EVIEWED BY OllTSIDE CRITICS.
WE EXPECT AND DESIE SIGNIFICANT CRITICISM AND PLAN TO MAKE TIE BEST USE OF T111S CRITICISM POSSIBLE TO PROVIDE A CUALITY FINAL REPORT THAT DOCLPENTS TIE LARGE PODY OF ECCS ESEARCH RRFORWD SINCE 1975.
l
. ,~.
I i
CODE SCALING, APPLICABILITY 4
! AiO UNCERTAINIY EVALUATION s
. PRESENTED BY NOVAK ZUBER, USNRC l
.1 TO ACRS K ETING AT INEL
- APRIL 28-29, 1987 l
l
)
i l
l l
i 1 CODE SCALING t ...
O O o- -
NTIIODOLOGY RESILTED FROM A CONTINUING EFFORT T11R0091 A SERIES OF TliEE NETING:
- IST E ETING (0CTOBER 9-10, 1986) - DESCRIBED A ET110D TO ADDESS
- 2f0 E ETING (JANUARY 5-6, 1987) - APPLIED TE ElliOD TO LBLOCA AND IDENTIFIED WHEE SCALING EFFECT ARE Ilf0RTANT
- 3RD EETING (FEBRUARY 26-27,1987) - DEVELOPED AN lNCERTAINIY EIl1000 LOGY PURPOSE OF PRESENTATION - TO DESCRIBE RESULTS OF 3RD KETING OBJECTIVE OF 3RD K ETING:
- 1. TO DEVELOP AND ESTABLISH A CONSENSUS ON A Ell 0D TO DETEW.INE LNCERTAINTIES IN BE CODES.
- 2. TO IDENTIFY CALCULATIONS AND EXPERIENTS NEED TO DEFDNSTRATE PROPOSED Ell 10D BY TE END OF 1987 CRIJERIA PROR) SED ETliOD MUST BE:
- 1. TE0filCALLY ACCEPTABLE (DEFENSIBLE)
- 2. REALISTIC AND FEASIBLE i
2 CODE SCALING 04/24/87
O O O 1
CUESTIONS THAT NEED TO BE ADDRESSED
- 1. HAS TIE CODE TIE CAPABILITY TO SCALE UP PilEN0ENA OBSERVED IN SMALL SCALE TEST FACILITIES TO FULL SIZE NFP?
i I
- 2. HAS TlE CODE TIE CAPABILITY TO ADDRESS A PARTICULAR SCENARIO OR A SET OF SCENARIOS FOR A GIVEN PLANT DESIGN? AND l
- 3. WHAT IS TIE UNCERTAIN 1Y WITH MIICil TIE CODE CALCULATES IITORTANT PARAFETERS, SAY TE PEAK CLADDING TDiPERATURE (PCT), IN A FULL SCALE NPP?
i 3 CODE DOClPENTATION 04/24/87
O O O' ROLE OF SCALING
- 1. _R_ELEVANCE SCALING IS CENTRAL TO:
A. TEST FACILITY DESIGN B. TEST SECIFICATION C. CODE ASSESSMENT l .
l 2. EFFECTS OF SCALE DISTORTION ON i
A. TEST ESlJLTS (PHENDENA, MAGNITUDE, TIMING)
L B. CODE ASSESSENT:
- " TUNING" OR ADJtJSTING " DIALS" NODALIZATION i
- 3. EFFECTS OF " SCALED-0UT" PROCESSES i
! A " SCALED-00T" PROCESS IS Olff FROM TEST RESULTS Af0 FRm CODE ASSESSENT i
l 04/24/87 4 CODE DOCLtefTATION
9 e
O -
~
t ,hi dbi
~
- 1 ,
l :
m
_ ir=L . _ _ _
a i 8
b - ,
o i 11 1 1111 !! j my* l, e ,
n o
a a
I yh'I m
_ 11 gI jw c ,
m
~
~
l' l1 g.
Ill O
SCALING EIH0DOLOGY
- 1. POWER TO VOLLE SCALING PRESERVES:
TIE FLUID DISTRIBUTION ENERGY DISTRIBUTION VELOCITIES ACCELERATIONS LENGTH DISTORTS:
AREAS DOWNCOKR PLENA ,
j STRUCTURAL HEAT TRANSFER HEAT LOSSES 1
i CONSEQUENTLY:
1 SOE PHASES OF TE TRANSIENT ARE ELL SIPULATED OTERS, MAY BE AFFECTED BY SCALE DIST0RTIONS 04/24/87 17 CODE DOCLENTATION
' ~ '
O O O I
- CODE CHARACTERISTICS
- 1. PRE 1975 ERA: RELAP3, RELAP 4 -
CAPABILITY: llEM,EVET,llVET FIELD EQUATIONS 3 l CLOSURE EQUATIONS
- PARAEIERS
- 21
- 2. POST 1975 ERA TRAC,RELAPS CAPABILITY
- IRIN FIELD EQUATIONS 6-7 CLOSURE EQUATIONS PARAfETERS: LISTED 175 lN.ISTED?
i i
)
04/24/87 6 CODE DOClPENTATION l
O O O CODE 01ARACTERISTICS CONT'D
- 1. ALL INTEGRAL TESTS AND PDST OF TIE SEPARATE EFFECT TESTS HAVE BEEN CONDUCTED IN SCALED-DOWN TEST FACILITIES Wi!IG1 IIAVE SCALE DISTORTION. CONSEQUENTLY, IT IS NECESSARY TO ASSESS TliE EFFECTS OF SU0i DISTORTION ON PROCESSES AND/0R PARAETERS OF IhTEPEST TO A NPP ACCIDENT SCENARIO OR SET OF SCENARIOS.
- 2. BE CODES RELY UPON A LARGE N W BER OF E WIRICAL CORRELATIONS AND PARA NTERS. SU0i EWIRICAL PELATIONS AR OFTEN " TUNED" TO IMPROVE TitEIR AGRED1ENT WITH CERTAIN DATA SETS (OBTAINED VERY OFTEN FR W SCALED DOWN FACILITIES). CONSEQUENTLY, IT IS IW ORTANT AND NECESSARY FIRST TO IDENTIFY THE PARATTERS MIIGI HAVE BEEN " TUNED" AND THEN TO ASSESS THE EFFECTS OF SUCH 4 " TUNING" ON NPP CALClLATION.
- 3. A CODE CAN 11 AVE " COMPENSATING ERRORS" BECAUSE SEVERAL PARAEIERS MAY HAVE BEEN " TUNED" TO PRODUCE AGREEWNT WITli CERTAIN DATA SETS. CONSEQUENTLY, IT IS NECESSARY AND IMPORTANT TO DETERMINE W11 ETHER OR NOT SUG1 CWPENSATING ERRORS EXIST AND IF S0, TO IDENTIFY SCENARIOS FOR MIIOi TilESE COWENSATING ERRORS MAY PRODUCE NON-CONSERVATIVE RESULTS.
l i
04/24/87 7 CODE DOClFENTATION
~
O O o-
- CODF CHARACTERISTICS CONT'D
- 4. E CODES AE COMPARED TO A LARGE BANI' 0F SEPARATE EFECTS AND IN1EGRAL EFFECTS TEST DATA.
)
j TIEY WILL GIVE REASONABLE AGREERNT WITH A NMEER OF TESTS BlIT NOT NECESSARILY WITH ALL OF Tile DATA. CONSEQUENTLY, IT IS IMPORTANT TO IDENTIFY TE REASONS FOR SU0i DISCREPANCIES IF
- j. lliE CODE IS TO BE APPLIED WITil CONFIDENCE TO A LARGE VARIETY OF SCENARIOS AND PLANT CONDITIONS.
- 5. BE CODES WILL SIW SCE SENSITIVITY TO N0 DING AND SOME OF TEIR EPPIRICAL RELATIONS HAD TO E SIPPLIFIED OR ADJUSTED TO IPPROVE THEIR NLFERICAL STABILITY. CONSE0VENTLY, TO PERFORM
- 6. BE CODES MAY HAVE SOE C0 RELATIONS OR PARAlEIERS WHIOl AE NOT SUPPORTED BY ANY EXPERIENTAL DATA OR ARE BASED ON DATA WHICil DO NOT COVER TIE RANGE OF INTEREST TO NPP APPLICATIONS.
CONSEQUEh1LY, IT IS VERY IFFORTAhT TO IDENTIFY SUCH C0PRELATI0tG AND PARANTERS AND TO ASSESS THEIR EFFECTS ON NPP SAFEIY ANALYSES OF A PARTICULAR SCENARIO OR OF A SET OF SCENARIOS.
8 CODE DOCWENTATION
! 04/24/87
O. O O i
1 CODE CAPABILITY CODE CAPABILITY TO CALCULATE A SCENARIO FOR A NPP DESIGN, ESTS ON FOUR ELEPENTS:
- 1. FIELD (CONSERVATION) ECUATIONS:
MilCHPROVIDECODECAPABILITYTOADDRESSGLOBALPROCESSES
- 2. CLOSURE EQUATIONS:
milch PROVIE CODE CAPABILITY TO MODEL AND SCALE PARTICULAR PROCESSES
- 3. NLPERICS:
HIICH PROVIDE CODE CAPABILITY TO KRFORM EFFICIENT APO REllAPLE CALCULATIONS
- 4. STRUCTURE AND N0DALIZATION:
MIIOl ADDRESS CODE CAPABILITY T0 WDEL PLAU GETETRY AND PERFORM EFFICIENT AND ACCURATE PLANT CALCULATIONS.
04/24/87 9 CON DOCLPEMATION
O O O 1
ILPACT OF CLOSURE ELATIONS CONSEQUENRY, IN ORDER TO ASSESS WlEIliER OR NOT A CODE CAN BE APPLIED TO CALCULATE A POSTULATED SCENARIO (OR A SET OF SCENARIOS FOR A SECIFIC NPP, IT IS ECESSARY:
- 1. TO DETERMINE WlEll4ER CR NOT THE CLOSUE PELATIONS ARE ADE0VATE T0 NDEL RENOPD!A AND PROCESSES IMPORTAhT TO T11E SPECIFIC SCENARIO AND NPP DESIGNS.
- 2. TO DETERMINE WiETiiER OR NOT TllE CLOSUI ELATIONS HAVE TiiE CAPABILITY TO SCALE-UP PROCESSES FORM TEST FACILITY TO CONDITIONS OF INTEEST TO NPP APPLICATIONS.
- 3. 10 IDENTIFY CLOSUE RELATIONS (CORRELATIONS, PARAETERS, CONSTANTS) MIICH HAVE BEEN
,mn ,
- 4. TO DETEFMINE MiElliER OR NOT THIS " TUNING" COULD liAVE BEEN AFFECTED BY POTD.TIAL SCALE DISTORTION OF TiiE FACILITY.
- 5. TO ASSESS TIE EFFECT OF " TUNING" 0N RJLL SCALE PLANT CALCULATIONS.
- 6. TO DETERMINE WEll1ER OR NOT TiiE CODE IIAS C0ffENSATING ERRORS AND IF S0, TO IDENTIFY SCENARIOS FOR MIICH TE CODE MAY OR WILL GENERATE NON-CONSERVATIVE ESULTS.
04/24/87 10 CODE DOCLPENTATION
' ~
i O O .
O 00DE SCALING, APPLICABILITY AND LNCERTAINTY EVALUATION BASIC EQUIR9ENT: COPFLETE DOClPENTATION l
CODE MANUAL j MDDEL AND CORRELATIONS OUALITY ASSURANCE DOClfENT USER GUIDE 4
ASSESSE NT REPORTS i
l J
Cli/24/87 11 CODE DOClfENTATION
O O O OBJECTIVES AND EQUIREENTS OF A (M DOCLPD!T I Tile CA DOCitENT HAS TliPEE OBJECTIVES, TllAT IS:
- 1. TO PROVIDE DETAILED INFORMATION ON (TE QllALITY OF) CLOSUE EQUATIONS, TliAT IS, ON C0 RELATION
! M)DELS AND/0R CRITERIA USED IN TliE CODE
- 2. TO DESCRIBE IDI TliESE CLOSUE RELATIONS AE CODED IN TE PROGRAM AND ASSUE TilAT WHAT IS LISTED IN TllE CODE MANUAL IS INDEED WHAT TliE CODE USES
- 3. TO PROVIDE A TECHNICAL RATIONALE AND JUSTIFICATION FOR USING TIESE CLOSURE RELATIONS f
(AS CODED IN TllE PROGRAM) IN Ti1E RANGE OF INIEREST TO NPP SAFET/ EVALUATIONS.
i
~F 04/24/87 12 DDDE DOCLPENTATION
i -
9 37 9 9 A
Y CBJECTIVES AND REQUIREENTS OF A QA DOCUKNT .
~
CONSEQUENTLY, FOR CORPELATIONS, MODELS, CRITERIA OR CONSTANTS USED IN TE CODE, TE QA DOCLENT PUST .
~
- 1. PROVIDE INFOR% TION ON: . , .
A.~ -ITS ORIGINAL SOURCE
~
~
B. Ils DATA BASE
~
C. ITS ACCURACY D. ITS AFFLICABILITY TO NPP CONDITIONS _
- 4. DCSCRIBED ANY MODIFICATION REQUIED TO OVERCTE COPRITATIONAL DIFFICULTIES
- 5. PROVIDE AN ASSESSENT OF EFFECTS DUE TO IFPLEENTATION (ITEM 3) AND/0R DUE TO MODIFICATIONS (ITEM 4) ON CODE OVERALL APPLICABILITY AND ACCURACY ~
w , e 1
04/24/87 13 CODE DOCIENTATION 4
~
O. O O EQUIREMENTS OF CODE A3MNTNT PEPORTS TO GAIN CONFIDENCE IN TIE PREDICTIVE CAPABILITY OF A EE CODE WlEN APPLIED TO NPP ACCIDEhT SCENARIO IT IS NECESSARY FOR CODE ASSESSEhT EPORTS
- 1. TO ASSESS CODE CAPABILITY AND WANTIFY ITS ACCURACY TO CALCULATE VARIOUS PARAETERS OF INiutsi i SUCH AS: CLAD TEMPERATUE, INLET AND OUTLET FLOWS FOR VARIOUS COPPONENTS, PRESSURE DROPS, LIQJID INVENTORY DISTRIElITION, TEPPERATUE DISTRIBlfil0NS, ETC.,
- 2. TO DETERMIE WHETIER OR NOT TIE CALCULATED ESULTS AE DUE TO COMPENSATING ERRORS,
- 3. TO ASSESS WlETER OR NOT TE CALCULATED RESULTS APE SELF-CONSISTENT AND PRESENT A C0HESIVE SET OF INFORMATION THAT IS TEGNICALLY RATIONAL AND ACCEPTABLE,
- 4. TO ASSESS WETTER OR NOT TE TIMING 0F EVENTS CALCULATED BY A CODE AE IN AGREEEhT WITH EXPERIENTAL DATA, AND
- 5. TO EXPLAIN ANY UNEXECTED OR AT FIRST GLANCE, STRANGE RESULT CALCULATED BY TE CODE. THIS IS PARTICULARLY IMPORTANT MEN EXPhilFB.TAL EASURBDITS ARE NOT AVAILABLE TO GIVE CREDENCE T0 l
i CALCULATED ESULTS. IN SUCH CASES, RATIONAL TECHNICAL EXPLANATIONS WILL G0 A LONG WAY TOWARDS GENERATING CREDIBILITY AND CONFIDENCE IN TE CODE.
k I
I 011/24/87 14 CODE DOCLENTATION
O '
O O '
i 1
j EQUIREfMS OF CODE ASSESSENT REPORTS (CONT'D) k FURTERMORE, MIENEVER TIEE IS A DISAGREEENT PEIWEEN CALCULATED ESULTS AND EXPERIENTAL DATA IT IS NECESSARY:
i
?
! 6. TO IDENTIFY Af0 EXPLAIN TE CAUSE FOR DISCPEPANCY, TilAT IS, TO IDENTIFY AE DISCUSS TE i DEFICIENCY IN TiiE CODE (OR IF NECESSARY, TO DISCUSS T11E INACCURACY OF EXPERIN NTAL EASUREENTS),
i j 7. TO ADDESS TE QUESTION OF HOW Iif0RTANT TE CODE DEFICIENCY IS TO OVERALL ESULTS, THAT I IS, TO PARAETERS AND ISSUES OF INTEREST,
)
i j 8. TO EXPLAIN Wily THIS CODE DEFICIENCY l%Y NOT HAVE AN Ilf0RTANT EFFECT ON TE PARTICULAR f,
SCENARIO,OR j 9. TO DISCUSS M1AT CHANGES S1100LD BE MADE TO CODE P0DELS Al0 CORRELATIONS IN ORDER TO OBTAIN j BETIER AGREE?DfT SHOULD TllE DISCEPANCY, TilAT IS, IllE CODE DEFICIENCY, llAVE A SIGNIFICANT IFPACT ON OVERALL ESULTS.
I i
I 15 CODE DOClf0(TATION 04/24/87
~
O O O i
WITH RESPECT TO THE CODE INPUT PDDEL AND SENSITIVITY STUDIES (IF PERF0PFED), IT IS ECESSARY FOR CODE ASSESSENT REPORTS:
- 10. TO PROVIDE A N00ALIZATION DIAGRAM ALONG Willi A DISCUSSION OF TE N0DALIZATION RATIONALE,
- 11. TO SRCIFY AND DISCUSS THE BOUNDARY AND INITIAL CONDITIONS AS WELL AS THE OPERATIONAL CONDITIONS FOR TE CALCULATION,
- 12. TO DISCUSS MODIFICATIONS TO THE INPUT MODEL (NODALIZATION, E0UNDARY, INITIAL AND/0R OPERATIONAL CONDITIONS) RESULTING FORM SENSITIVITY STUDIES (IF CONDUCTED),
- 13. TO PRESENT AND DISCUSS RESULTS OF SENSITIVITY STUDIES (IF PERFORED) ON CLOSURE EQUATIONS OR OTER PARAETERS,
- 14. TO PROVIDE GUIDELINES FOR PERFORMING SIMILAR ANALYSES.
+
Gl!/24/87 16 CODE DOClfENTATION
~
O O O
_ CODE ASSESSENT TWO POSSIBLE APPR0 AGES
- 1. BOTTOM-UP A. EXAMIES AND ASSESS EVERY SINGLE M) DEL IN TE CODE IN A UNIFORM FASHION B. ASSESSES EFFECTS OF EVERY SINGLE MDDEL (SENSITIVITY STl0Y) ON CALCULATED ESULTS
- 2. TOP-DOW A. IDENTIFIES MACROSCOPIC PEN 0ENA WillCH HAVE A SIGNIFICANT IWLUENCE ON TIE OVERALL RESULTS FOR A DISTANCE CLASS OF SCENARIOS B. ASSESSES SIGNIFICANT PIOGENA AGAlf6T TEST DATA C. PERFORMS A SENSITIVITY STUDY l
l 3 CODE SCALING 04/24/87
~ ~
O O O TE PROBLEM l
- 1. BOTTOM-UP A.PPROAG A. TASK OF EXAMINING EACH MODEL INDIVIDUALLY AND IN COMBINATION WITH OTER MODELS IS ENORP00S.
- SANDIA CLAIMS THAT FOR IPFORTANT PARAETERS 2(n+1)2 CALCULATIONS AE EQUIRED TO COPFLETE SENSITIVITY STUDY B. FOR ELAP4, SANDIA IDENTIFIED 21 POTENTIALLY iPPORTANT PARAFETERS AND CARRIED OUT i
200 SEPARATE RUNS f
f C. TRAC AND RELAP5 HAVE OVER 175 POTENTIALLY SIGNIFICANT PARA ETERS
- 2. TOP-DOWN APPROACH i
A. IPPORTANT PHEN 0ENA NEED TO IDENTIFY AND RANK THEM FOR EACH CLASS OF SCENARIOS IT HAS NOT BEEN DONE BUT IT CAN BE DONE B. SCALING NEED TO DEMONSTRATE SCALE-UP CAPABILIlY OF TE CODE FORM TEST FACILITY TO FULL SCALE NPP l
04/24/87
O. O O SYNTHESIS i
BolTOM-UP Am TOP-DWN APPR0 AGES HAVE BOTil ADVANTAES AND DISADVANTAGES
- 1. BOTTOM-UP EXAMIES AE ASSESSES EFFECTS OF EACH CORRELATION BUT THIS CAlWOT BE DOE E CAUSE OF LIMITATIONS Ilf0 SED BY COST AND TI E .
NOTE TE NINER OF PARAETERS IN BE CODES.
- 2. TOP-DWN EXAMINES AND ASSESSES EFRCTS OF lif0RTANT PARAETERS BUT MAY MISS EFFECTS OF THOSE MlICllllAVE NOT BE CONSIDERED.
NOTE EFFECTS OF COPPENSATING ERRORS, OF "HOW MADE" CORRELATIONS /PARAETERS ETC.
IT WOULD BE DESIRAELE TO DESIGN A ETHOD MlIQI WOULD COMBINE POSITIVE ASPECTS OF B0lli APPROACHES GOTTOM-UP AND TOP-DOWN) mille MINIMIZING TE NEGATIVE DES.
5 CODE SCALING 04/21/87 1
O O o-CSAU EVALUATION ETH.0_D i
MAIN FEATURES:
- 1. TE CSAU ETHOD CONSIDERS ONLY PROCESS TilAT AE IPFORTANT TO TE SECIFIC SCENARI OF SCENARIOS). CONSEQENTLY, IT INCLUDES THE PROCESS IDENTIFICATION AND RANKING ACTIVITY OF STA E 1.
- 2. DATA Fim SEPARATE EFFECTS TEST (SET) FACILITATES AE USED IN TE CSAU ETHOD T0:
A. EVALUATE THE CAPABILITY AND ASSOCIATED UNCERTAINTY OF Tile CODE TO PODEL PROCE IPPORTANT TO TIE SCENARIO, l
l B. ASSESS CODE CAPABILITY AIO ASSOCIATED UNCERTAINTY TO SCALE-UP TESE PROCESSES l
TO FlLL SCALE). THIS UNCERTAINTY DlE TO SCALE, IS USED (IN STAGE 9) TO EVALUATE TIE TOTAL LNCERTAINTY. CONSEEENTLY, SET FACILITIES AE USED (IN STAGE 5) TO ADDESS Af0 ESOLVE QUESTIONS CONCERED WITH SCALING
- 3. TIE CSAU ETH00 ADDESSES llE QESTION CONCERED WITH EXTRAPOLATING (UP TO FULL S RESlLTS BASED ON DATA FORM INTEGRAL EFFECTS TESTS (IET) FACILITIES (WillCH BY ECESSITY, INCLl0E SOE SCALE DISTORTIONS AS DISCUSSED IN SECTION 4.4.2)
- 4. TE END PRODUCT OF TIE CSAU ETHOD IS A STAlDENT ON TlE TOTAL UNCERTAINTY (STAGE CALCULATE A PARTICLLAR PARAFETER, SAY PCT, FOR TE SECIFIED SCENARIOS (0R CLASS OF SCENARIOS)
IT TAKES INTO ACCOUNT THE EFFECTS OF N0DALIZATION, F0DELS (CLOSUE AND FOR A SECIFIED PPP.
ELATIONS), INPLIT A!0 SCALE (TilR00C41 STAGE 7). TE PROCEDUE (CF STAGE 9) TO EVALUATE TE TOTAL LINCERTAINTY IS SCENARIO DEPEM)ENT.
1 l
CODE SCALING 6
04/24/87 ,
O O O i
1 i
l i
ADDITIONAL FEATUES
- 5. DIAGNOSTIC ANALYSIS TO IDENTIFY POTENTIAL LIMITATIONS AM) EVALUATE CODE CAPABILITY
- 6. QUALITATIVE EVALUATION OF CODE APPLICABILITY TO SPECIFIED SCENARIO (s).
- 7. EFFICIENT CODE ASSESSENT TEST MATRIX i
- 8. DEIECil0N OF COPPENSATING ERRORS
}
l 9. QUALITATIVE EVALUATION OF CODE TUNING EFFECTS
)
i
)
i 1
- l I
7 CODE SCALING 04/24/87
Scenano Specification II Process / Phenomena j identification and Rankmg Code Documentation Code Documentation System Components l System Manual and QA Manual and QA 1 P 1 f T 1 f 3 Diagnostic Analys's 2 1 I 1f /%
/ \ u
,2 r
System Components System 37 7
3 3 Potentief 1 f II Potential No N ? to Model and 7 C to Modef and g Sca;e Scale Code Documentation Yes Yes 4
- liser Guide and Developmental Assessment ,
a If II I SET Matrin 4 IET Matrix 4 g]y SET # m m 2 i 0s
^
Selection and
- D g - Selection and '
1 3 j Justification l Justificat on Yes 7 8 7 Scale Y'S ,Yes Se,i, II II 4 Distortion S'.9"'I'.C ortion j lI 1f j(
No No 5 Code Assessment 5 0 SET: Up to Full Scale 6 6 IET Reduced Scales 3 7 7 Compen58t'n9 ? 4 Compensateg Accuracy. Capatnhtes.
Accuracy, Capat>htes, Yes Deficiencies, Nodalization E"O' S Yes Errors Otficiencies, Nodabration.
Scahng, Uncertainty Scahng. Uncertainty 3 7 ) 7 Assess Assess 1f 1F 7 1 f 1f 1 I 4 Toint Uncertamty to 9 Calcu8 ate Specific Scenario and for Specific NPP Figure 4-57 Flow Diagram for the CSAU Evaluation Method O O O
i.
! O O O i
l
' APPLICATION TO LBLOCA l 1. DEVEL0ED PROCESS IDENTIFICATION RAM (ING TABLE (PIRT).
- 2. BLOK)0 W PHASE:
j TEST CONDITIONS FOR TE PRESENTLY AVAILABLE DATA RELEVANT TO BLOWDOW AE PROTOT(PICAL
! 3. REFILL /REFLOOD PHASES: PRESENT FACILITIES BASED ON POWER TO V0 Life SCALING HAVE SCALE ,
D Y CT 1 S EC B ASS B I MPOR TO
! EFILL Af0 REFLOOD. TE SLEJECT DISTORTIONS IN TE CURRENT SUBSCALE FACILITIES AE EXPECTED TO BE CONSERVATIVE WITH ESRCT TO PCT, BUT PUST BE C0fFIRPED.
l i
k
. 4. DEVELOPED A ETHOD TO EVAltlATE TOTAL UNCERTAINTY.
i l
)
b; -
8 CODE SCALING
! 04/24/87
1 f
f C.a.n.i.e.aK.ac 1P g J . .s....... m. c .
I f 3
c ..o........
... . c ...... ,
F 1 4
.~ ,, -
o..s.n.... .c. ... ... ... . .. ~ . . . o.
1 P 1f g 4 t .n. J..ts., 7 c..h,.,..nc.. . .. :
. ... c. S.E.T . ...
J b i f
,,,. n,is. .a .
. . . . cy ..
o o g
.s : car;;gyg'=
n i r v-
- o. .in... c,..he _
.7 ..
If h* **
... .e.u.e, e.c7 .
g- -
t ...
i r u,. ,cv 1 P 7
.....u-,. ,,,o.l,7ta,. :
" g i
c
. . v- . : ';l; ;'a ,*,:*
u o
...u.., ; ....... : oa.,.,,,o.
a....
1 f 10
.ca. vaw saan ,.
1 i
Figure 4.g1 Plow Diagram for Determining Total Uncertainty of Code to Calculate PCT for e PWR Large Greek LOCA i
l l e
O O O -
SUMMARY
OF METHOD:
APCT = APCTICODE + EXPERIMENTI + APCT ISCALINGl + APCT(DEFICIENCIES) + APCTilNPUTI APCT = APCTIGEl + APCTl7A or 50) + APCTISI + APCTIS) l j 4. NPP CALCULATION I FIRST PEAK l LATE MAX. PEAK 2
b a
s i
j TIME l
1 l SETS NPP CALCS. IET.S LATE PCT E- 8- INCLUDING EXPERIMENTS 6E. IET'S EARLY PCT SE.
I IMPORTANT PROCESS FROM SETS 2APCT 2aPCT CODE CODE i
/
/m%% f %
- APCT U
+
EXPERT *4ENT U
+
EX PE Rl'AE NT l ,
A A 4
({ (DEF.)
g[N l MEAS. TIME MEAS. PCT ME AS. PCT
! l y u o
' NPP CALCS.
SETS IET*S PCT (LATE ONLY) i 5F. IMPOR1 APIT PROCESS 7A. INCL. SET SCALE EFFECTS 9. 80UNOS FOR IETS 6D.
I e5 #* NPP WITH -
P / %' ' SET I D S ALE
""~~~ '
- Lj NPP WITHOUT **
ti u
i [ \*#r- g INPUT UNCERT.
o y
o n
a 1
SCALE 1.0 -
TIME MEAS PCT SCALE 10 l Figure 4-82 Results of Various Activities Leading to Total Uncertainty of a BE Code to Calculate PCT for a PWR Large Break LOCA Scenario
' t ' - -
s l
i I
i l
)
,I I
SUGGESTED METHODOLOGY
! 1. ESTABLISH RANGE OF CONDITIONS OVER WHICH CODES i ARE TO BE APPLIED-ULTIMATELY OVER BREAK SPECTRUM
- AT VARIOUS LOCATIONS. INITIALLY ONLY FOR LBLOCA.
i i
1
- 2. IDENTIFY IWQ NRC BE CODES AND FREEZE. DEFINE, ONE SPECIFIC OR GENERIC PWR FOR TESTING OF METHODOLOGY.
l 3. PROVIDE HIGH QUALITY CODE DOCUMENTATION TO INCLUDE A QA DOCUMENT.
l-
\
1 l
l l
~
=
O
~
O '
O=
SUGGESTED METHODOLOGY (CONT.)
- 11. PERFORM REACTOR ANALYSIS WITH CODES FOR ONE PWR TO IDENTIFY LOCA WITH HIGHEST EARLY PCT AND LOCA WITH HIGHEST LATE PCT OVER SPECTRUM -
0F BREAK SIZES AND LOCATIONS
- PROVIDE INPUTS FOR ALL CASES
- DESCRIBE N0DALIZATION USED FOR EACH KEY COMPONENT l 1
- DOCUMENT RESULTS S0 THEY CAN BE REPRODUCED BY AN INDEPENT GUALIFIED USER l
O ' '
T O~
l l
SUGGESTED METHODOLOGY (CONT.)
i
- 5. PERFORM REASONABLENESS TEST OF CODES FOR EVERY KEY COMPONENT THROUGH COMPARSION TO SEPERATE EFFECTS TESTS SA. COMPARE PRINCIPAL COMPONENTS OF CODES TO AVAILABLE COMPONENT SET'S SB. MUST USE SAME COMP 0NENT N0DALIZATION AS i IN REACTOR ANALYSIS SC. PREVIOUS COMPARISONS CAN BE USED IF DONE WITH SAME N0DALIZATION AS REACTOR, OR REACTOR N0DALIZATION CAN BE CHANGED TO MATCH COMPARISONS, OR SHOW N0DALIZATION UNIMPORTANT. DOCUMENT ALL RESULTS.
l
~
O.
~
O - O SUGGESTED METHODOLOGY (CONT.)
- 5. PERFORM REASONABLENESS TEST OF CODES FOR EVERY KEY COMPONENT THROUGH COMPARSION TO SEPERATE EFFECTS TESTS SA. COMPARE PRINCIPAL COMPONENTS OF CODES TO AVAILABLE COMPONENT SET'S 5B. MUST USE SAME COMP 0NENT N0DALIZATION AS IN REACTOR ANALYSIS SC. PREVIOUS COMPARISONS CAN BE USED IF DONE WITH SAME N0DALIZATION AS REACTOR, OR REACTOR N0DALIZATION CAN BE CHANGED TO MATCH COMPARISONS, OR SHOW N0DALIZATION UNIMPORTANT. DOCUMENT ALL RESULTS.
4 O ' - -
T O' l
f i
SUGGESTED METHODOLOGY (CONT.)
}
j SD. IDENTIFY ANY KEY COMPONENTS FOR WHICH:
l - SET'S NOT STATISTICALLY SIGNIFICANT j - SET'S AND ANALYSES DO NOT AGREE
! (LARGE UNCERTAINTIES) t '
l - SET'S DO NOT C0VER ENTIRE SCALE RANGE SE. PERFORM COMPARISONS WITHIN A SPECIFIC IDENTIFIED MATRIX OF SET'S. JUSTIFY SELECTION FROM MATRIX.
I i
I i
1 i
o
~
' 0"~
' ~
l O i
}
}
i SbGGESTED METHODOLOGY (CONT.)
I
{ SF. FOR THOSE SET'S WHICH DO NOT COVER j ENTIRE SCALE, PLOT UNCERTAINTY VS. SCALE, j EXTRAPOLATE TO FULL SCALE, AND ESTIMATE HOW MUCH UNCERTAINTY INCREASES OR DECREASES
) AT FULL SCALE.
i 5G. CONSIDER EXPERIMENTAL UNCERTAINTIES IN f MAKING COMPARISONS l
Q "}' 0 i
i SUGGESTED METHODOLOGY (CONT.) '
i i
- 6. COMPARE CODE PREDICTIONS OF PCT TO MEASURED PCT IN INTEGRAL TESTS 6A. SUBDIVIDE MEASURED PCT'S INTO EARLY PCT'S (BLOWDOWN) i AND DELAYED PCT'S (REFILL /REFLOOD). ONE TEST CAN
- HAVE ONLY ONE OR BOTH. IF MORE THAN ONE DELAYED
- PCT, USE HIGHEST.
6B. MUST USE SAME N0DALIZATION IN KEY COMPONENTS l AS IN REACTOR ANALYSIS (SEE STEP SB AND SC).
l 6C. PERFORM COMPARISONS WITHIN A SPECIFIC IDENTIFIED l MATRIX OF IT'S. JUSTIFY SELECTION FROM MATRIX.
l l
I ie
O~
O Q" SUGGESTED METHODOLOGY (CONT.)
6D. IDENTIFY THOSE INTEGRAL TESTS FOR WHICH LIMITED SCALING DATA AVAILABLE. PLOT PCT (CALCULATED)-PCT (MEASURED) VS. SCALE AND EXTRAPOLATE TO FULL SCALE.
NOTE: EARLY PCT HAS DATA OVER FULL SCALE AND SCALING EFFECTS NEED NOT BE CONSIDERED.
l
. r +^=
SUGGESTED METil0D0 LOGY (CONT.)
6E. PLOT CALCULATED VS. MEASURED PCT FOR AND DELEYED PCT'S FOR IT'S. DRAW B0UNDING OR 95 % CONFIDENCE 2
LINES.THESELINESSl100LDGIVE2(0 CODE +
EXPERIMENT FOR DELAYED PCT, TiiEY MAY NOT INCORPORATE ALL SCAllHG EFFECTS.
T T eo f G k o 5 .
y o $ eo o j *e & ,e* ,
y o y 0 9 U * *- e
, ; g 5 Y,
- 2 m
a g?ARLY PCT (MEhsuntD) LMTE PCT (hEpsung0) l
. 1:
SUGGESTED METHODOLOGY (CONT.)
- 7. ASSESS ADDITIONAL UNCERTAINTY 7A. USE 5F RESULTS AND APPLY REACTOR PREDICTIONS TO CALCULATE CHANGE IN PCT UNCERTAINTY DUE TO INCREASE / DECREASE IN UNCERTAINTY ESTIMATES (5F) FOR EACH KEY COMPONENT. COMBINE ALGEBRAICALLY OR ABSOLUTELY TO PRODUCE AN ADDITIONAL APCT (SCALING) 7B. USE 6D TO CALCULATE APCT (SCALING) 7C. USE 7A AND 7B TO DEVELOP APCT (SCALING).
USE LARGEST IF PLUS, SMALLEST IF MINUS.
s ~% - -
SUGGESTED METHODOLOGY (CONT.)
- 8. ESTIMATE APCT DUE TO CODE DEFICIENCIES
- COMPONENTS WITHOUT ENOUGH SET'S
- COMPONENTS FOR WHICH SET'S AND ANALYSES ARE SIGNIFICANTLY DIFFERENT APPLY JUDGMENT OR EXCERCISE REACTOR ANALYSES WIT!!
INCREASED UNCERTAINTIES IN THESE COMPONENTS AND FORM APCT (DEFICIENCIES)
- 9. USE REACTOR ANALYSES TO CALCULATE APCT DUE TO UNCERTAINTY IN INPUT PARAMETERS (FUEL, DECAY HEAT, ETC.)
i l
l
- -- - ... . . . s . . .. .. , , _
~
~,-
'- - =- + - - - .., . .- . .
1 i
SUMMARY
OF METHOD:
i APCT = APCTICODE + EXPERIMENT) + APCT ISCALING) + APCTlDEFICIENCIESI + APCT191
+ APCTilNPUTI -
APCT = APCT64El + APCTl7A ce 60) . APCital I
I
- 4. NPP CALCULATION FIRST PEAK LATE MAX. PEAK I
e i
i TIME l
1 ins uTE PCT i
8- INCLUD N X REMENTS E- IMPORTA*d PROCESS FROM SETS 2APCT 2APCT CODE CODE
/ = s% ,~ _ EXPERIMENT U EXPERIMENT
, /
/ APCT IDEF.I U
6 l ,
1 u
%/
I
) TIME i
MEAS.
i If I
II If IET'S PCT ILATE ONLYi f SETS NPP CALCS*
- 9. BOUNOS FOR lET'S 6D.
I SF.! IMPORTANT PROCESS 7A. INCL. SET SCALE EFFECTS i
I NPP WITH T
- p p 5 -# INPUT UNCER'T. g LE APCT P / %
[ """a==
NPP WITHOUT d o / \* #7 """ N g INPUT UNCERT. y
( r o .
s i
8 SCALE 1.0 ME AS. PCT f SCALE 1.0 TIME 1
i i
Figure 4-62 Results of Various Activities Leading to Total Uncertainty of a BE Code to
, Calculate PCT for a PWR Large Break LOCA Scenario
O ' "
o' o' 4
FUTURE ACTIVIES APRIL, 1987 - TECHNICAL WORKING GROUP MEETING TO DETERMINE ALREADY PERFORMED
- APPLICABLE CALCULATIONS AND REVIEW
- ASSESSMENT MATRIX l JUNE, 1987 - PRESENT METil0DOLOGY AND PLANS FOR TESTING TO CSNI PWG-2 CODES TASK GROUP i
JULY, 1987 - PRESENT METH0 LOGY AND PLANS FOR TESTING TO ACRS ll AUGUST,1987 - PRESENT METHODOLOGY AND PLANS FOR TESTING TO SELECTED GROUP OF PEERS BY OCT. 1. 1987 - TEST AND DEMONSTRATE THE METHODOLOGY l FOR LBLOCA USING ONE SELECTED BE CODE
! sy DEc'.31, 198 7 -
PR* VIDE COMPLETE ceDr DocunsNTArrow ines. venus.',
Assrmswr nrmers l Mapunt,COMPLETE
- PReviot usres ouses, ga poconsur,*
DocuaggpTATisp or Test Apo DEM*NtTAMTsen age opetMTAtu1Y MET 90beLeG1
i i
! .