ML20127A978
| ML20127A978 | |
| Person / Time | |
|---|---|
| Issue date: | 06/13/1985 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1417, NUDOCS 8506210315 | |
| Download: ML20127A978 (195) | |
Text
NSf-IhO OR'Ghf.L UNITED STATES NUCLEAR REGULATORY COMMISSION O
IN THE MATTER OF:
DOCKET NO:
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON EMERGENCY CORE COOLING SYSTEMS LOCATION: ALLIANCE, OHIO PAGES:
1 134 DATE:
THURSDAY, JUNE 13, 1985 38FFe#F.co.1s9 en 0
g n,, o w, g omove from ACRSoffice ACE-FEDERAL REPORTERS, INC.
igo l CWicial Re:vrters s
444 North Ca itci street Washmgton, C 20001 62 5 850613 (202)347-3700 NATIONWIDE COVERACE
5071 00 01 1
AGBagb 1
UtJITED STATES OF A!! ERICA 2
tJUC LI:AR REGULATORY CO!!MISSIO!!
3 ADVISOKY CO!!!!ITTEE Oli REACTOR SAFEGUARDS 4
SUDCOl't!ITTEC Oli E!!I:RGEFICY CORE COOLIIIG SYSTEttS 5
6 Babcock & Uilcox 7
Alliance Research Center 8
1562 Deeson Street 9
Alliance, Ohio 10 11 Thursday, 13 June 1985 12 13 The Subconnittee convened at 8:30, David A.
- Hard,
(}
14 oreciding.
15 51Ef tBI:ES P R E S E t!T :
16 David A. Ward 17 Carlyle t'ichelson I
18 l
19 Ivan Catton 20 Virgil Schrock 21 ACRS Consultants 22 23 24 25 l
l h
5071 01 01 2
1 AGBagb 1
PROCEEDINGS 2
MR. WARD:
The meeting will come to order.
3 Our first speaker this morning is Dr. Kn ig ht, who
)
4 seemed to have been our last speaker yesterday.
5 (S l id e. )
6 DR. KNIGHT:
Thad Knight again, from Los Alamos.
7 I have also been asked to talk about plant calculation as it 8
compares to the MIST calculation that we have shown you 9
previously.
I am here to talk about it, Jim Lime did the 10 actual work relating to the plant calculation.
11
( Slide. )
12 This work is supported by the TRAC calculational 13 system number, it is not a part of the MIST support analysis
(}
14 at Los Alamos.
Harry Tovmassian is the NRC program i
15 monitor.
And we have issued a draft report on this analysis i
16 in April; it's a draf t, there is an error in it and we will 17 have to re-issue it.
18 The TRAC plant model reflects data that we i
19 received from B&W.
And the plant calculation is compared to 20 the MIST pretest-7, which'is the calculation you saw 21 yesterday from zero to 2500 seconds.
Beyond 2500 seconds, i
22 j.
we had a reactivity-feedback input error-in the code, or the 23 calculation, and so we need to repeat that portion of the 1
[
24 test.
l 25 (S l id e. )
l I
l
5071 01 02 3
1 AGBagb 1
This is labeled TRAC Plant Model Improvements.
2 Basically it's things that we did to try to get a better 3
calculation than the first calculation that we previously 3
\\-
4 inputted.
We have lowered the secondary-side water 5
inventory, trying to get down into a better estimate than we 6
got from B&W.
7 We have included the reactivity feedback.
We 8
used B&W's reactivity insertion rate to account for the 9
We got the B&W reactor coolant pump curves and we 10 have included the reactor coolant pump power to the 11 coolant.
This is not something that comes out directly 12 automatically from the model but has to be accounted for 13 through the proposed systems and power inputs.
()
14 We have done an improved vent valve model.
This 15 is a code update.
By " improved," what we mean is that 16 instead of having a two-point table we now have essentially 17 an unlimited number of points in the table.
This allows us 18 to better describe the K as a function of delta-p 19 prescription of the vent valve behavior.
(
20 The upper-head recirculation flow is included'.
l 21 It's not quite the way it really is in the plant but it does 22 provide us the desired ef fects of heating up the upper-head i
23 fluid.
1 24
{
And we got a slight lower main feedwater flow 25 rate coastdown than what we had originally.
5071 01 03 4
2 AGBagb 1
That steam generator auxiliary feedwater axial 2
spreading modeled, that just means that we used the parallel 3
channel in much the same area that we d id in our MIST 4
calculations.
S And then we have also modeled the heat structures 6
associated with the shell in the steam generator.
7 (Sl ide. )
8 I don' t want to belabor this slide very much, it t
9 is sort of a choice between showing you very little on the i
10 picture and showing you more than is reasonable to talk 11 about and I thought you might be more intere sted in seeing 12 the detailed noding, so individual cells are shown 13 throug hout the one-dimensional components.
I will show you
(}
14 the steam generators a little bit later.
i 15 Let me just say at this point that there are 144 16 cells in the 3-D vessels:
two range, two radial range, six 17 azimuthal segments and 12 levels to account for the vessel, 18 and there are 330 one-dimensional cells in the input.
19
( Sl ide. )
i 20 This shows you the vessel noding as compared to 21 the B&W reactor vessel.
And again, as I said before, there l
22 are 12 levels.
The recirculation flow that we get is up and 23 then down the guide tubes, that recirculation amounts to 24 about 10 percent of the total core flow.
25 And then we have al so included 1 think it is e
r--
e
-w e
w -
5071 01 04 5
2 AGBagb 1
approximately 2 percent leakage through the bypass --
2 through the vent valves.
That account for leakage around 3
the hot leg nozzles and other assorted leakages between the
(
4 downcomer and the upper head.
5 This figure shows you that we have allocated the 6
connections for the cold legs and hot legs accordingly and 7
that's how we get to six azimuthal segments.
8 (Slide.)
9 This shows you the steam generator noding.
It is 10 the same on both loops.
Primary is divided into two 11 parallel channels, a 10 percent channel representing the 12 wetted channel, the 90 percent representing the dry 13 channel.
The secondary is split in the same way.
()
14 There is one cross-flow connection.
The 15 downcocer is modeled and the aspirator flow coming across 16 here and then the exit into the steam line.
17 (Slide.)
18 We did one other thing that we had not done in 19 MIST and it gets us back to the original calculation where 20 it looked like we were removing perhaps more heat than B&W 21 thought we should f rom our steam generator early in the 22 transient.
23 And so what we have done is we have partitioned
{
24 the heat structures in the wetted channel, other than just a 25 straight 90 percent dry and 10 percent.
The 90 percent
5071 01 05 6
1 AGBagb 1
channel has slabs to represent 90 percent of the tubes and 2
then in the 10 percent channel, starting at the bottom we 3
have slabs that represent 10 percent of the tubes but then O
4 as we get near the top, we had some information f rom B&W 5
that talked about the spraying of the auxiliary feedwater 6
after it comes in, and so we have partitioned at the very 7
top only a small amount of the surface into the wetted 8
channel and a larger fraction into the dry channel.
And 9
it is a linear interpolation until we get down to the full 10 10 percent of the wetted channel at that elevation.
11 You'll notice that there are some very fine nodes 12 low in the steam generator and the reason that we did this 13 is that we had the model sitting there previously in which
()
14 we were trying to do steam generator dry-out type 15 calculations and we just left that noding in there.
It's 16 not necessary for this transient.
17 (Slide.)
18 Trips and Safeguard Assumptions.
The reactor 19 trip on low pressure at 13.2 megapascals occurred at.2 20 seconds into the transient.
The turbine trip and stop valve 21 closure occurs on the above trip with a half-second delay.
22 Reactor coolant trip was a two second delay.
Main feedwater i
23 trip on the same trip signal with a two second delay and 24 then a 12 second coastdown.
Auxiliary feedwater is delayed 25 40 seconds beyond the trip.
And then the high pressure
5071 01 06 7
1 AGBagb 1
injection is tripped on on a low pressure signal at 10.44 2
megapascals, 35 seconds.
3 The HPI flow is MIST-specified scaled by 817 as t\\'
4 is the auxiliary feedwater flow and these represent full 5
system capacities.
6 It is important for you to understand that this 7
plant calculation really does not look like any specific 8
plant but is sort of a generic calculation in order that we 9
can do this type of analysis when we get for the HPI and 10 auxiliary feedwater flows.
11 (Slide.)
12 Transient Sequence of Events.
Let me first point 13 you to the footnote:
the MIST event times were basically
(}
14 what I showed you yesterday in that table, shif ted by 177 15 seconds, so that the start of core power decay in the MIST 16 calculation coincided with the 3.5 percent power in the 17 plant calculation.
18 And what you see is that MIST tends to be a 19 little bit later.
The auxiliary feedwater is coming on 20 later.
The Loop A and hot leg saturates at about 80 seconds 21 later.
HPI delivery is coming on just slightly late.
22 This point we forced to be the same 23 (indicating).
And the reason why these other things are 24 late in MIST is that all these other things are tripped off 25 of this 3.5 power point.
Y 5071 01 07 8
2 AGBagb 1
The top of the hot leg voided in the plant 2
calculation, Hot Leg A, at 500 seconds versus 350 seconds in I
3 the MIST calculation.
The B loop voided at 1100 seconds in 4
the plant and 1150 in MIST.
And you are seeing the sa=e 5
sort of assy=etry between the two loops in the plant 6
calculation that we saw and the MIST calculation.
7 BCM starts earlier in the plant at 1800 seconds.
8 And the calculation that I a= going to show you is l
9 terminated at 2500 seconds because of the core power l
10 proble=s.
i j
11 MR. MICHELSON:
Is there se=e reason why it 12 starts so much earlier in the plant, you know, a physical 13 reason?
i
(}
14 DR. KNIGHT:
Why the BCM starts earlier?
i 15 MR. MICHELSON:
Yes.
16 DR. KNIGHT:
Just standing here I a= going to be
[
17 hard pressed to tell you.
I think =aybe as we go through 18 the slides =aybe it will becoce so=ewhat clearer.
It has to 19 do with the way in which the syste= voids versus when the t
20 HPI coces on and the fact that the plant continues to 21 circulate so that the cold legs are actually a little bit 22 cooler than the MIST facility, I believe.
23 MR. MICHE LSON:
But it has more to do with 24 rotation of levels than it does with te=peratures of the 25 legs, doesn' t it?
f t
i
,,...m c
.-3
4 e
t 5071 01 08 9
2 AGBagb 1
DR. KNIGHT:
Well as the temperature of the cold i
t l
2 leg goes down, the break flow goes up.
So as the break flow 3
goes up you will increase the depletion rate from the
- (
- )
t 4
primary system.
l f
5 MR. MICHELSON:
This is quite a difference.
Go 6
ahead, maybe it will become apparent.
l i
7 DR. KNIGHT:
Okay.
8 (Slide.)
l 9
This shows you the pressurizer pressure and, as i
10 you can see, we are falling about one megapascal low during
[
]
l 11 most of those calculations.
Basically we never got on track l
i I
l 12 between MIST and the plant, however the shape of the curve 1
I i
13 is essentially the same af ter the MIST f acility begins.
So t
()
14 we feel good about this comparison and because of the course 15 of the transient we are not bothered by this slight j.
1
]
16 discrepancy in terms of the system pressure; the important l
i 17 thing is that both facilities appear to go through the same i
l 18 sorts of phenomena in much the same way.
i
(
19
( S l ide.' )
I l
20 MR. WARD:
Let me get what you' re saying.
You' re i
21 not bothered by it but I mean who ought to be bothered by it l
j 22 are the people that are using -- that design the MIST I
i i
23 intended to be a mock-up of the plant.
Now maybe this isn't 24 a serious discrepancy but as I see it -- I mean, what you' re
}
25 doing is telling us from the perspective of a TRAC l
i
I 5071 01 09 10 j
1 AGBagb 1
calculation how well the MIST f acility models the plant.
2 DR. KNIGHT:
I'm telling you, from the i
I 3
perspective of using the data for assessment purposes, that i
4 it is not a particularly important effect because the 5
facility ultimately goes through the same set of phenomena 6
and in much the same way in the plant facility as the MIST 7
facility, okay?
And because of that it doesn't bother me.
8 If this pressure difference had caused the plant j
9 calculation to go through a different set of phenomenon, j
i 10 then I would have been more concerned.
But I view the MIST 1
i
]
11 facility as primarily providing data to support code j
12 assessment and the code will ultimately tell you what the l
h 13 plant does.
And because of that all I need is the j
14 phenomenon to come out in about the right order.
i 15 MR. WARD:
So you are getting the phenomenon in 16 sort of the same sequence and in fact the timing is a little 17 different?
18 DR. KNIGHT:
I expect the timing to be a little' y
t i
19 bit dif ferent because the scale volume in MIST tends to be a
]
20 little bit high at the upper elevations, so it takes it 1
21 longer to drain and things like that.
22 DR. SCHROCK:
But the demonstration that you are
)
23 getting the same phenomenon is not adequately represented by J
24 this slide.
You see that in many other places.
25 DR. KNIGHT:
I have to show you many other slides-i 1
i
- -.m.--..--
5071 01 10 11 1
AGBagb 1
to confirm that.
2 You know, to get back to the question -- I don' t 3
think I gave you a very good answer earlier on why BCM 4
starts earlier in the plant calculation relative to the MIST 5
calculation.
And I think I just hit on it when I said that 6
the volume in MIST above the hot leg nozzles is overstated 7
so it takes longer for that to drain.
That's probably a 8
major f actor.
9 MR. MICHELSON:
Volume is overstated in MIST.
10 DR. KNIGHT:
Yes, the volume up above the hot leg 11 nozzles tends to be too high.
12
( Sl ide. )
13 The Intact Loop Secondary Pressure.
Again keep
(}
14 in mind that the MIST calculation is shifted 177 seconds.
15 We are starting off at the steady-state operating condition 16 on the steam ger.erator secondary.
With the turbine -- with 17 a
reactor trip the main line steam valve closes until you 18
...t the safety relief valves which limits the pressure 19 excursion and then the pressure begins to come down.
The 20 primary syntem is still flowing because of the pump 21 coastdown for a long time and you get good heat transfer j
22 between primary and secondary and we are coming out here.
23 And basically this loop is beyond -- coupling up the set 24 points between the primary and secondary.
1 25 This depressurization is primarily due to the l
l i
5071 01 11-12 1
AGBagb 1
injection of the auxiliary feedwater into the secondary 2
causing condensation of the loop.
t 3
(Slide.)
4 The broken loop steam generator secondary 5
pressure, it is the same sort of curve.
If you will recall I
6 from my MIST. calculations I showed you that the intact loop 7
steam generator depressurized significantly more than the 8
broken loop steam generator --
i i
9 MR. MICHELSON:
Excuse me.
On the previous 10 slide, this was the intact loop that you were showing and 11-then you are showing the secondary side and the pumps are 4
12 running.
i 13 DR. KNIGHT:
The pumps are coasting down.
i(}
14 MR. MICHELSON:
The pumps are coasting.
15 DR. KNIGHT:
I think if we refer back to -- I 16 thought on the table --
17 MR. MICHELSON:
When do you shut the pumps off on l
18 your calculation?
19 DR. KNIGHT:
The pump trip occurs with only a 20 couple second delay off of the reactor trip and they start 21 coasting down.
As in the MIST calculation, the intact. loop 22
. voids earlier and stagnates earlier than in -- than the 23 broken loop, okay?
/
24 We see that same comparison coming out even 25 though the pumps are coasting down.
i.
L
..~
5071 01 12 13 1
AGBagb 1
DR. CATTON:
Thad, in your handling of the hot 2
leg, the results of Jean-Pierre's study at SAI show that
'3 there is slug flow.
O 4
Have you incorporated this in your plant 5
calculations?
6 DR. KNIGHT:
No, we have not.
7 DR. CATTON:
So you basically have -- your 8
small-scale integral facility bias is built into your plant 9
calculation?
10 DR. KNIGHT:
At this point, yes.
11 DR. CATTON:
That takes away from what you're 12 trying to play out, you really should have shif ted that.
}
13 DR. KNIGHT:
I should have shifted what?
{ ()
14 DR. CATTON:
You should have accanodated what is j
15 being observed in the large pipes.
16 DR. KNIGHT:
I don' t disagree with what you' re 17 saying.
I will say though as that information comes out it 18 doesn't just automatically get dropped into the code.
[
19 The other thing is that I'm not sure that I'm 20 ever getting into a slug flow regime.
21 DR. CATTON:
But that's something then that 22 should be pointed out to give this more credibility.
If you 23 never get into the slug flow' regime then what I am saying is 24 the relevance if you don' t know, no, until you find out.
l 25 DR. KNIGHT:
I can't say that I have looked at it-l r
,,-.m
5071 01 13 14 1
AGBagb 1
but the way the thing coasts down and the void profiles that 4
2 I saw, it really looks like what it goes into is a bubbly 3
flow and then at a very high elevation and at a very low
\\
4 flow rate it transitions into a dispersed droplet and it 5
separates very nicely.
And it's the same separation 6
mechanism that the code used in the MIST calculation.
7 DR. CATTON:
I think that should be checked.
8 MR. MICHELSON:
What break size did you use?
9 DR. KNIGHT:
This is a 10-square-centimeter 1
10 break.
11 MR. MICHELSON:
Did you run a parametric study to 12 see if it was size-sensitive?
I 13 DR. KNIGHT:
No, we did not.
I'm sure that
(}
14 changing the break size will change the course of the 15 calculation slightly.
16 MR. MICHELSON:
The comparison though that you j
17 are trying to exemplify here might be sensitive to the size 18 break that you select for the calculations.
19 DR. KNIGHT:
You're right.
20 MR. MICHELSON:- I don' t know that it does.
21 DR. KNIGHT:
What we did is we took -- we already 22 had the MIST definition of the nominal test and we took that 23 definition of. the nominal test and forced it onto a generic 24 plant model.
25 MR. MICHELSON:
At least for 10 square t
i
+,,
ww,,
w wwo w
-w-w<,
m
i 5071 01 14 15 2
AGBagb 1
centimeters that comparison is reasonable.
2 DR. KNIGHT:
Yes.
l 3
I am basically through with this plot.
It is the
.O 4
same phenomenon that we saw in the MIST calculation where we i
5 get not as great a condensation depressurization on the 6
secondary side as in the intact loop and then because of the 7
loop flows the pressure rises more rapidly and gets its 8
control and then for the duration of the transient then I am i
I 9
showing that it remains under couple control to the primary 10 system.
11 (Slide.)
12 The intact loop steam generator primary level,
.i 13 again what in my opinion looks like a fairly good comparison
' ()
14 given the discrepancy in the scaling up high in the system.
15 The two curves basically track one another and there is a 16 f airly large dif ference through here but it gets wiped out.
17 MR. MICHELSON:
Just a moment, in that previous 18 slide you showed the pressures on the broken secondary side 19 are higher than the non-broken secondary side?
20 DR. KNIGHT:
Yes.
I 21 MR. MICHELSON:
What's the explanation?
22 DR. KNIGHT:
That's the same relationship that I 23 showed you in the MIST calculation, okay?
4
(~}
24 MR. MICHELSON:
Broken loop pressures build up 1
\\-
25 higher.
i ne m--
m.s-y%e 2--
y-v--
,.-c-e,v
,w-7 g
,----wpa w..g.g
-,.rq
,-----og-<---
+wg-,
e py,
- -wm-+Tr-W--
9 i
I 5071 01 15 16
[
l 1
AGBagb 1
DR. KNIGHT:
The intact loop flow stagnates i
2 earlier and therefore you lose coupling between the primary i
3 and secondary.
In the intact loop as you --
That's right.
i 5
DR. KNIGHT:
-- as you inject the auxiliary l
6 f eedwa te r, the cold auxiliary feedwater, at a high 7
elevation, the condensate level is constant.
Because there 8
is no flow, the intact loop does not couple back up and so 9
the secondary pressure remains fairly flat for a long period i
10 of time.
11 In this loop, in the broken loop, the flows l
12 continue and so we've got thermocoupling primary and i
13 secondary and yes, I get a condensation effect but it is
- ()
14 mitigated by the continued heat transfer from the primary to l
15 the secondary.
And therefore at this point the broken loop i
i 16 steam generator is at a higher pressure than the primary l
17 intact loop, i
18 MR. MICHELSON:
That doesn' t seem intuitively 19 reasonable that the broken loop is at a higher pressure than l
5 20 the unbroken loop.
21 DR. CATTON:
If the unbroken loop stagnates then
{
j 22 you are going to lose the coupling between the primary and 23 the secondary in that steam generator and it is going to b3 24 a very strong weight to the secondary side.
25 MR. MICHELSON:
Clearly as the break size gets i
l l
i l
l 5071 01 16 17 j
1 AGBagb 1
bigger there is a point at which the secondary side of the i
1 2
broken loop has a lot lower pressure than on the unbroken i
j 3
loop, clearly as you get the bigger breaks.
It must be.
l 4
DR. CATTON:
If you have a large enough break, I
l 5
sure.
I 6
MR. MICHELSON:
So now this really creates the f
i 7
problem for the operator, doesn't it?
Because he usually 8
uses pressure to measure whether or not he's got a broken I
i j
9 pipe, doesn' t he?
10 DR. CATTON:
If he does here he's going to be in I
11 t rouble.
i l
12 MR. MICHELSON:
If he does here he's going to be l
13 in real trouble but where is the crossover point at which
(}
14 you start to use --
15 DR. CATTON:
-- on the break side.
E 16 MR. MICHELSON:
It looks to me like it is a 4
17 judgment test for the operator to judge which secondary side 18 is broken.
I 19 DR. CATTON:
I think the rate of change of things 20 is telling us.
I 21 MR. MICHELSON:
Perhaps, although it doesn' t seem i
22 like the rate of change of things is all that much dif ferent 23 in this case.
24 DR. CATTON:
No, I am referring to the size of f
25 break.
i f
i I
I
t 5071 01 17 18 f
1 AGBagb 1
MR. MICHELSON:
Clearly you must do a parametric i
2 study for operator procedure purposes now to start to l
3 instruct the operators when does he change his method of 4
4 operation in selecting the broken loop.
i
{
5 DR. CATTON:
I would hope so.
6 MR. MICHELSON:
That is astounding, that the 3
I i
7 broken side is higher pressure than the unbroken side for 8
this size break.
t i
9 DR. CATTON:
The break is maintaining the flow 10 through the steam generator.
I j
11 MR. MICHELSON:
I understand.
But for a larger I
i 12 break the situation would be opposite.
j 13 DR. CATTON:
The larger break you would exhaust
(}
14 the steam generator if you use the coupling.
l 15 MR. MICHELSON:
But the break size -- they both 16 can be drifing on about the same pressure for a certain i
17 break size, obviously.
18 DR. CATTON:
Do the procedures hinge on --
l 19 MR. MICHELSON:
I haven' t even seen -- you know, i
20 that's the point of this whole business is to write l
1 21 procedures.
l 22 I always thought that the broken side would be 23 obvious.
i 24 DR. CATTON:
Well in a large break it is, but in 25 a small break you must depend -- I'm not sure what 1
l i
i i
-,.,m.
-. ~. _ _,. _ _. _ _, _..
5071 02 01 19 1
AGBagb 1
procedures depend on the small break.
2 DR. KNIGHT:
Jim Gloudemans has something on 3
that --
b'l 4
MR. JONES:
On the procedures, I guess I am a 5
little lost as to what your problem is called.
The 6
procedures for a small break LOCA will have him just feed 7
both generators all the way up and to use both generators 8
for heat removal.
9 MR. MICHELSON:
Yeah, but I thought he was going 10 to try to identify where the -- you know, what ancillary 11 line might have broken on the primary side or whatever and 12 this doesn' t help him any because you can' t tell where the 13 break is from this information.
(}
14 MR. JONES:
No, and to my knowledge the 15 procedures do not tell him to identify which line is 16 broken.
He's got actions to close all available lines that 17 he has valves on, decay heat drop lines, that kind of stuff, 13 but not to tell which side is broken.
19 MR. WARD:
Okay.
So Eventually he is going to 20 try to isolate the leak but the immediate action procedures 21 for a small break LOCA don' t have him attempting to do that.
22 MR. JONES:
Actually the immediate actions have 23 him close all these valves, if it's not in the immedia te i
24 it's in the first set of follow-up actions.
25 As far as the secondary side response, if it is i
i
5071 02 02 20 1
AGaagb 1
primary syste= saturates, he's just to feed both stea:
2 generators up to the 95 percent level as quickly as 3
possible.
t 4
MR. MICHELSON:
How does he tell this fros a 5
secondary side break?
6 MR. JONES:
If it is a s=all secondary side 7
break, he will have difficulty.
If it's a s=all secondary
^
8 side break, the dif ference would be you wouldn' t expect to 9
saturate in the prisary systec.
You shouldn't saturate the 10 pri=ary except for large secondary side breaks and you 11 definitely see dif ferences in the secondary side response 12 for those.
13 MR. MICHELSON:
You =ay not even reach saturation
()
14 on the pri=ary side with if this is a s=all hole.
15 MR. JONES:
It does, that's why you lose the 16 circulation.
4 17 MR. MICHELSON:
You don' t really know what you' ve 18 got, do you, for a while.
19 MR. JONES:
That's one of the reasons why the 20 procedures have changed to deal with sy=ptoms and not to 21 worry about what the accident particularly is.
And his 22 first action is to deal with the subcooling, then they deal 23 with excessive heat re= oval, et ce te ra.
24 MR. MICHELSON:
That's interesting to note.
25 DR. DiIGHT:
Let =e just coc=ent tha t it is the I
i l
l l
5071 02 03 21 1
AGBagb 1
pressurizer that determines which loop is going to stagnate 2
in this transient, it is not the break really -- and I'm 3
guessing because I haven't done the analysis but if you move 4
the break into the loop with the pressurizer I think you'll 5
still see that that loop stagnates first because of the 6
pressurizer dumping the hot liquid and it's a large volume 7
of hot liquid in this case into the pressurizer -- or into 8
the hot leg, it is near saturation to start with and as the 9
pressure comes down you are going to void the hot leg.
So I 10 look at it and I am not really all that surprised from our 11 results.
12 (Slide.)
13 This slide and the next one show you the broken
()
14 loop steam generator primary levels for MIST and plant.
It 15 just confirms again that the plant calculation is voiding 16 more rapidly or the levels are dropping more rapidly than in 17 MIST.
But, again, given that difference that is going to 18 show up throughout the comparison, the curves are showing 19 the same basic trends.
20 (Slide.)
21 Steam generator secondary levels, MIST and 22 plant.
I don't really have a whole lot to say about it.
23 (Slide.)
24 This is the intact loop hot leg u-bend voiding b
25 and you see that in the intact loop although the voiding
5071 02 04 22 1
AGBagb 1
begins a little bit earlier in the plant, okay, than in the 2
MIST calculation, the ultimate voiding, void fraction of one 3
is reached slightly later and I think that that is primarily 4
due to a dif ference in the flow rate between the two 5
calculations.
6 (Slide.)
7 MR. MICHELSON:
Is this at the top of the hot 8
leg, of the U-bend?
9 DR. KNIGHT:
Yes, it is at the very top.
10 Okay.
This is the broken loop hot leg U-bend 11 voiding.
In the MIST calculation we did get a blip fairly 12 early but then it went to a void fraction of one again and 13 continued to flow.
We get the voiding to begin occurring
(}
14 really in earnest at about the same time between MIST and 15 the plant, there are some differences.
We got a big 16 spillover in MIST that we don't see in the plant 17 calculation.
But again I think that the phenomenon are 18 tracking about the same, j
l 19 (Slide.)
20 Cold leg A-1 mass flow.
What you see here is --
21 and keep in mind that for this plot and the other mass flow l
22 plots not only have we shif ted MIST but we have scaled it by 23 a f actor of 817.
The. comparison of the first 177 seconds is 24 good and then the MIST plant, the MIST calculation shows a 25 fairly steady symmetric flow in the two A-loop cold legs
5071 02 05 23 1
AGBagb 1
whereas the plant reverses flow and we've got a sustained 2
negative flow in this one.
3 (Slide.)
4 If we flip to the next figure, you' ll see that we 5
have a sustained positive flow.
So that we set up very 6
early in time a circulation from one cold leg in the A loop 7
to the other cold leg to the downcomer.
8 DR. CATTON:
Thad, did you take a look at what's 9
happening in the downcomer here, by beginning to flow from 10 one cold leg directly back into the other one when you get 11 this reverse?
12 DR. KNIGHT:
I think at this time what we are 13 seeing is that it is flowing around the loop, okay, from one
(}
14 directly over into the other.
We looked at a lot of those 15 plots and they tend to merge in my mind.
16 The net flow pattern through the downcomer varies 17 in time.
There is at one point in time a gross 18 recirculation that goes like from A-1 to A-2 through the 19 downcomer to B-1 or whichever cold leg is on that side, a 20 big loop like that; at another period of time, and it is 21 circulating just between the two cold legs in the same 22 loop.
23 DR. CATTON:
How important is this type of recirculation in the downcomer to the final results?
(~}
24
\\_-
25 DR. KNIGHT:
One thing that it does is that we l
5071 02 06 24 1
AGBagb 1
no longer calculate the extremely cold temperatures in the 2
cold legs that we saw in our MIST calculation.
I don't 3
think I showed you that plot.
But we calculated cold leg
\\~
4 temperatures in MIST that looked very much like the REDBL, 5
RELAP-5 cold leg temperatures; they were extremely cold.
6 This circulatory flow in the cold legs mixes that very well.
7 DR. CATTON:
This sounds like a truly 8
multi-dimensional downcomer.
9 DR. KNIGHT:
It is.
It is behaving that way.
10 The downcomer participates in the setting up of the 11 recirculatory flow and once it is established it provides 12 part of the head necessary to sustain that flow.
13 DR. CATTON:
What role do the vent valves play
(}
14 during this period, are they opening and shutting?
Is there 15 flow out of them and into them?
16 DR. KNIGHT:
I guess what I would like to do is 17 move on to the vent valve flow.
18 MR. WARD:
Wait a minute, though.
Is that why l
19 you're not seeing the oscillation in the MIST flow, because 20 there was not really a --
l 21 DR. KNIGHT:
You know, I can't say with l
l 22 certainty.
I can't say that these oscillations are directly 23 related to oscillations in the vent valve flow.
l l
24 If we get a little bit further back into the l
25 presentation -- in fact, I think it's like the last slide --
l t
i 5071 02 07 25 l
AGBagb 1
(Slide.)
2 You'll see that this is the mass flow through the 3
vent valves and it tends to oscillate.
And what is 4
happening in the calculation is that the vont valve is 5
opening and closing, okay?
Those flow surgos are also 6
reflected in the core inlet flow and when you got them, you i
7 get some change in the downcomer levels.
8 (Slide.)
9 The intact cold log in the loop B mass flow.
And 10 the reason why we indicated it that way is that we've got a 11 discrepancy between the MIST facility model and the plant 12 model; in one of the calculations, the break is in B-1, in i
13 the other it's in B-2.
And so we have compared the broken
(}
14 cold leg to the broken cold log and the intact cold log in 15 that loop to the intact cold 109 It's the same basic i
16 pattern except that in the MIST calculation the loop flow 17 was flowing in the other direction.
18 MR. MICHELSON:
Is the break at the dischargo of 19 the pump?
20 DR. KNIGHT:
Yes.
I can' t really tell you why it i
21 flows one way in the plant and the other way in MIST, it 22 looks like to me it is a very subtle difference in the 23 calculation required to got the MIST calculation flowing in 24 the other direction.
}
25 (Slido.)
i l
\\
I
a 5071 02 08 26
]
1 AGBagb 1
The other -- this is the broken cold log in loop 2
B and it i s the same basic pattern, as I said before.
The 3
interesting thing is that the code is indicating that MIST O
4 and the plant show the same basic flow ps.ttorns though, in l
I 5
terms of increases and decreases.
There is a little blip 6
here that we don't soo in plant but all in all wo think it 7
is a reasonably good comparison of phenomenon betwoon tho 8
two facilities.
And the direction of the flow, the 9
direction of the flow in that loop I don't think is really i
10 important.
11 MR. WARD:
It is interosting that it really is 12 picking up tho difference in the way the MIST has tho
)
13 downcomor and vont valvos modelod and the way it would bo in 1 (}
14 the plant.
Let's soo, are thoro any plans whoro you aro 15 doing this sort of a calculation for the other integral 4
l 16 facilitios?
The University of Maryland, wo've got that nico 17 vont valve mock-up there and I just wondor how --
18 DR. KNIGitT:
What I have boon trying to do is 19 drive tho interfacility comparison tost to a common 20 definition which at this point I would like to soo 21 correspond to the MIST nominal tost, becauso I've got a lot 22 of calculations alroady dono for the MIST nominal.
If wo do i
23 additional plant calculations, like Bill Docknor indicated 24 to you yesterday, and wo run somo moro, you know, assymotric j
25 casos, then one of those other tests is fino.
.,m-. - - -
vr
---,--y
-T---+v-i-
--ir--
sc--
---s--
5071 02 09 27 1
AGBagb 1
Those calculations are like an order of magnitude 2
more expensive than the MIST calculation because of the 3-D 3
vessel and I don' t want to have to do any more of those than O
4 I absolutely have to; that is, I would prefer not to have to 6
i 5
do one por facility.
1 6
MR. WARD:
I guess to do the University of 7
Maryland modeling it would have to be with a 3-D vossol, I 8
guess.
9 DR. KNIGitT:
Right.
We have a University of l
10 Maryland deck that wo are working on now that is i
11 throo-dimensional.
}
l 12 MR. MICilE LSON :
llow did you model the conductanco
)
13 of the vont valvo arrangement on MIST?
()
14 DR. KNIGitT:
You mean the flow conductanco?
l 15 MR. MICilELSON:
Yes.
16 DR. CATTON:
With difficulty.
17 MR. MICilELSON:
With great dif ficulty.
18 DR. KNIGitT Actually it's from a user standpoint l
19 it's easier in MIST than it is here.
20 MR. MICl!ELSON :
Is that right?
i 21 When it is used.
22 DR. KN IGi!T:
Because the downcomor and the upper
{
23 plonum woro separated with a pipo running in-botwoon them.
24 Wo stuck a pipo in-betwoon with a valvo componont and the 25 valvo opens and closos by somo sotpoints --
i i
I i
1 1
1 l
t
5071 02 10 28
]
1 AGBagb 1
MR. MICHELSON:
There is a fluid impedance valve 2
through that pathway --
lO 3
DR. KNIGHT:
And hopefully we have gotten that i
i 4
impedance right.
1 5
MR. MICHELSON:
And how does that impedance for l
l 6
MIST calculations compare with the impedance used in the 7
real plant calculation?
8 DR. KNIGHT:
Jim Gloudemans can correct me, but I t
l 9
think we are like a factor of 10 higher.
10 MR. MICHELSON:
With these strong spikes, that 11 indicates very high instantaneous velocities.
In the real 12 world, that system will take a high instantaneous velocity 1
13 but if it is a pipe it will just damp it right out.
14 DR. KNIGHT:
Keep in mind that that's a fairly
(}
15 long time scale and the flows are accelerating at a i
j 16 reasonable rate, given the resistance that they' re seeing.
l 17 What is happening in the plant that is not 18 happening in the MIST facility is that the valve is swinging j
19 open and closed, now it's not going all the way closed but i
20 it is swinging open, relieves the upper plenum pressure and i
21 swings closed.
In MIST the control is set up such that it 22 opens full at an eighth of a psi, closes full at almost 23 zero --
i 24 MR. MICHELSON:
Which would infer to me very high 25 instantaneous velocities when it is open in MIST.
l i
l l
[
1 1
l 5071 02 11 29 2
AGBagb 1
DR. KNIGHT:
in MIST.
I 1
2 MR. MICHELSON:
-- or a higrar instantaneous 1
i 3
velocity.
l j
4 DR. KNIGHT:
No, i t's got a larger frictional i
j 5
dropdown and so the -- you are able to maintain your i
6 pressure difference between the upper plenum and the top of 1
l 7
the downcomer at a value greater than the setpoint for i
j 8
closing that valve.
4 i
{
9 DR. CATTON:
-- for a longer time.
I 10 DR. KNIGHT:
In the calculation that we showed i
j 11 you, once it opened, it stayed open.
l 12 flR. MICHELSON:
In MIST, once it's open it stays j
13 open and that's why you don' t see so many spikes on the f (])
14 calculation, because it is just an accurate' filter, and then I
1 l
15 you put a finite conductance -- and I don' t know how you
}
16 modeled the thing --
l 17 DR. CATTON:
Is this the nominal case?
I j
18 DR. KNIGHT:
This is the nominal case.
19 DR. CATTON:
And in the nominal case you are
. i l
20 seeing the multi-dimensional characteristics of the annulus.
l 21 DR. KNIGHT:
Yes.
i i
j 22 DR. CATTON:
I really think you need to run the 23 assymetric case.
i l
24 DR. KNIGHT:
Ivan, at this point I agree with you i
N
(
25 and we've been talking with the NRC about doing that.
You l
i
,y, r,.,,,
..-~.--,--,-w-m
,m,,-
-.--.,-mm..-,&
w,
..m-r--w-,w-n---,
--.,,- --- -+-, w
-n e,,m---
--.ee
.n,
-,.w.-.
,,._.7 7
5071 02 12 30 1
AGBagb 1
know, if you have a suggestion out of the list of MIST tests 2
in the matrix at this point that you think should be a 3
candidate for running, I would like for you to give us what 4
your choice for the tests would be.
5 DR. CATTON:
My choice would be the one that most 6
exaggerates the assymetric behavior.
7 DR. KNIGHT:
1 hear those words and those are the 8
words that I want to go through and make my selection.
I 9
just, you know, for diversity of opinion I would like to see I
10 what your selection would be.
If it agrees with mine, 11 then we're in good shape.
If not, then I need to have a 12 reason for the next meeting.
1 13 MR. WARD:
Let's see, would you remind me, what 14 is the nominal test here?
}
15 DR. KNIGHT:
The nominal test in MIST is a 10 16 square centimeter for pump' discharge break.
17 MR. WARD:
And what would give you more --
18 DR. KNIGHT:
What would give you more assymetry?
l 19 It could be the assymetric control of the steam generators.
20 A larger break size might give you more assymetry.
You l
21 know, it's hard to just sit down and look at the list of 22 tests, at least for me, and a priori pick the one that is 23 going to be the most assymetric.
24 DR. CATTON:
But you are in a better position to l
25 do it than any of the rest of us, running the code and l
l l
~
t 5071 02 13 31 t
2 AG3ag b 1
looking at all the nitty-gritty.
Whether you choose a large 2
break or the one stea= generator being quite dif ferent than 3
the other; I don' t know wnich would drive it harder.
r 4
MR. WARD:
And see this is really the issue that 5
was -- really a concern that the MIST design and, you know, 5
the realities of the designing secething like MIST and why 7
there was a lot of interest in a facility, experi= ental 3
nodel like the University of Maryland has developed.
So I 9
think, you kncv, we don' t want to -- I sean, this is exactly i
10 the area where the analysis kind of ought to be going, it 11 see=s to
=e.
I 12 DR. KNIGHT:
I think that, given the questions I
13 that have been raised about the code, a considerable part of 14 the analysis time needs to go into the code assessment and 15 that is looking at tests as related to the codes.
16 MR. WARD:
This is the experinental facility 17 assess =ent and since the data for any expert: ental 18 f acilities are the, you know, pins for the code.....
2 19 DR. KNIGHT:
Yes.
20 (Slide.)
21 This is just showing a cocparison of break flow.
22 You get dif ferences because of dif ferences in the 23 facilities.
We are getting sone voiding at these spikes out 24 into the cold leg of the plant that we have not seen in the l
25 MIST calculation.
But all in all it is -- again, given the
5071 iO3 01 32 1
AGBagb 1
disparity among the systems, a fairly good comparison, I 2
think.
3 (Slide.)
4 Core Inlet Mass Flow.
I don't want to say a 5
whole lot other than the oscillations that you see here 6
coincide with the oscillations on the vent valve flow that 7
we have already shown you.
8 (Slide.)
9 Conclusions and Recommendations.
MIST and plant 10 transients show similar behavior. The pressures and loop 11 voiding compare reasonably well, there seems to be an offset 12 in pressures in the plant calculation but we don't think of 13 that as particularly important.
The cold leg temperatures
()
14 dif fer because of mass flow differences and plant cold-leg 15 flow circulation.
Loop assymetry occurs in both 16 transients in much'the same way.
I was really impressed 17 with how similar the two calculations were.
18-We need to correct the calculation out beyond the 19 2500 second point and we expect that when we do we will see 20 this system refilling much as it does in MIST.
We expect 21 that it may occur somewhat earlier than it does in MIST.
22 Always a push here continue to support -- I 23 think that is supposed to be MIST program -- with full-scale
()
24 plant calculations with both 3-D and 1-D vessel modeling.
25 The reason that I say with both 3-D and 1-D
5071 03 02 33 1
AGBagb 1
vessel modeling is that we would like to back out at 2
full-scale what the multi-dimensional effect is on the 3
calculation.
And that's something that we have not done but
()
4 we have talked to the NRC about doing.
5 That's it.
t 6
DR. CATTON:
Is this draft report available?
7 DR. KNIGHT:
Yes, it is.
It was transmitted to 8
the NRC -- I guess Bob has it -- I was going to say members 9
of the PMG but I don't think they all have it. I've got a 10 copy if you would like it.
11 DR. CATTON:
I don't want to be given it here.
12 DR. KNIGHT:
My personal preference is that I 13 would prefer not to give you a copy of it until we correct 14 the top of the transient after 2500 seconds, because the O
15 power went back up and --
16 DR. CATTON:
Even if I promise not to look at it 17 beyond 2500?
18 DR. KNIGHT:
I know you, Ivan.
19 We will see to it that you get a copy of the 20 report when it comes out.
21 MR. MICHELSON:
In the MIST there isn't a safety 22 injection feature, is there, in this calculation?
23 DR. KNIGHT:
What do you mean safety injection?
24 MR. MICHELSON:
Normally if the pressure gets O
\\l 25 down to 600 pounds or so in the primary side you have safety
5071 03 03 34 1
AGBagb 1
injection.
2 MR. WARD:
The core flood tanks.
3 DR. KNIGHT:
The core flood tanks are in the
()
4 model.
5 MR. MICHELSON:
The flood tanks are in the model.
6 DR. KNIGHT:
Yes, we have not run the calculation 7
or any calculation to the point at which they have injected.
8 MR. MICHELSON:
They are set up below 600 pounds.
9 DR. KNIGHT:
I think they are set up at 400, is 10 that right?
11 MR. JONES:
They're set at 600.
12 MR. MICHELSON:
They will start to show up as you 13 go beyond 2500 seconds.
14 DR. KNIGHT:
If the pressure goes down there, 15 yes.
16 AR. CARTER:
I have just one question:
the MIST 17 calculation is the same one you showed yesterday where the 18 downcomer is a single node?
19 DR. KNIGHT:
Yes.
20 MR. CARTER:
Okay.
21 DR. KNIGHT:
No multi-dimensional analysis of the 22 MIST calculation.
23 MR. JONES:
Carl, as far as the flood tanks, I am 24 not sure that we would expect flood tanks actually for the 25 nominal case.
One of the cases in the MIST test case is the l
5071 03 04 35 1
AGBagb 1
50 square centimeter break, which is expected to get to the 2
core flood tanks.
3 MR. MICHELSON:
You said they were set at 6007 (O
_j 4
MR. JONES:
Yes, 650 psia.
5 MR. WARD:
Let's see, Thad, Randy Carter's 6
question about the MIST model having just a single node for 7
the downcomer; would you expect to see anything different if 8
you modeled that in more detail?
9 DR. KNIGHT:
The main thing that I see right now 10 in the plant calculations is that I've got the intact loop 11 cold leg circulating.
In the MIST calculation they did not 12 circulate until very late in the transient.
It's hard to 13 understand a circulation in those two cold legs because 14 there is a common pressure point at each end of the two cold O
15 legs at the intact loop.
We get a circulation because of a 16 sudden drop in pressure and a little bit of voids -- well we 17 have set up an oscillation in there that finally gets a 18 little bit of cold water into the pump suction in the MIST 19 calculation.
20 They are not ouite exactly symmetric and once 21 they go into the -- when I say " exactly symmetric," I mean 22 out five or six digits.
When I get a little bit of cold 23 water into the pump suction and it is unbalanced, then the 24 oscillation starts in the MIST whereas in the plant O~
25 calculation because of the downcomer not providing a common
5071 03 05 36 l
1 AGBagb 1
pressure point for each cold leg, I get a circulation set up 2
there very early.
3 MR. WARD:
I see that.
But I'm trying to J ()
4 understand -- do you think that that is truly a difference 5
in the characteristics of the plant in the MIST experiment?
6 DR. KNIGHT:
Yes.
7 MR. WARD:
-- not just in the way that you noded 8
the calculation.
9 DR. KNIGHT:
Yes.
10 MR. WARD:
You can't be sure, though.
11 DR. KNIGHT:
I believe that the MIST facility i
l 12 will behave in a much more one-dimensional fashion than the 13 plant.
It's just because the downcomer by necessity is very 14 small and in the plant it is very big.
(
15 MR. MICHELSON:
In the plant there are some 1
16 anti-rotational devices on the reactor coolant pumps, a 17 reverse flow device and so forth.
Are such features 18 included in your MIST reactor or MIST pumps?
19 DR. KNIGHT:
Right now I don't know if the pumps 20 themselves are designed with an anti-rotational or 21 anti-reverse-rotation device, but the calculations that we 22 have shown you so far, the rotor is locked.
23 MR. MICHELSON:
Now in the MIST case will the 24 rotor be locked?
25 DR. KNIGHT:
As long as we are running the test l
4
5071 03 06 37 1
AGBagb 1
without pump operation.
2 MR. CARTER:
Just to clarify that, the MIST pumps 3
are in a case where we lock the rotors.
()
4 MR. MICHELSON:
You can lock the rotors on them.
5 MR. CARTER:
In ours there is pretty much like a 6
disk brake that locks the rotors for both forward and 7
reverse.
8 MR. MICHELSON:
You are reasonably satisfied that 9
the fluid conductance after that is properly scaled somehow?
10 MR. CARTER:
We have a fairly high heat loss in 11 that area, that is included in the calculations.
We also 12 have a fairly high heat storage structure there -- as you 13 saw the flanges, that also is modeled --
14 MR. MICHELSON:
The fluid impedance is what I am O
15 talking about.
16 MR. CARTER:
Right.
The lock rotor resistance is 17 there, the irrecoverable losses are scaled correctly.
18 DR. SCHROCK:
Thad, I just wanted to comment 19 again about this primary to secondary modeling.
You have 20 seized on something that is terribly simplistic and it looks 21 like you are kind of frozen on it and I am looking for 22 something that is better.
I am not convinced that you 23 aren't going to come up with some surprises eventually as a 24 consequence of this.
O 25 DR. KNIGHT:
I can't disagree.
5071 03 07 38 i
3 AGBagb 1
DR. SCHROCK:
It seems to me that you do have 2
possibilities of modeling in a better way.
What I'm saying 3
is I think some effort should be put into more
()
4 aophistication in modeling, that rather than taking a simple 5
prescription and freezing it into the code and esnentially 6
saying there doesn't seem to be a reason to do better --
7 maybe you haven't said that but it almost sounds that way --
8 DR. KNIGHT:
We did the best that we could with 9
what we have, with the information available and the state 10 of the code, okay?
In my opinion, for us to do a better 11 analysis, a more predictive analysis will require some 12 work in a multi-dimensional area of primary to secondary.
13 If you're talking about adding into the code 14 models like B&W has added into RELAP-5, we have looked at O
15 that, we may do that, but those were subject to the same 16 somewhat arbitrary prescription as what we have done here.
,i 17 We haven't at this point hardwired anything into 18 the code in terms of that split.
The split here was 10 j
19 percent wet and 90 percent dry for the plant calculation.
20 In MIST there are three tubes that are wet and 16 that are 21 dry and that works out to be about 15 porcent wet.
22 DR. CATTON:
What about doing something like BWR 23 track on the upper core support plates where CCFL would 24 determine how much flow through those two support plates?
25 That would determine how much spreading thoro is at each
5071 03 08 39 2
AGBagb 1
level?
2 DR. KNIGHT:
We are looking at adding a CCFL 3
model to the code, which would be similar to the BWR track,
()
4 it may not be exactly.
5 DR. CATTON:
It should be similar, it's the same I
6 process, dumping the cold water in to the top of the BWR and 7
it gets down to the core via CCFL.
8 DR. KNIGHT:
We need to do something like that.
9 The multi-dimensional analysis, in my opinion, has the 10 capability of predicting the spreading correctly.
The other 11 thing you've got to be aware of is that when you go to a 12 multi-dimensional analysis, particularly if you put a lot of 13 high nodes in it so that it can be predictive in terms of 14 spreading of the auxiliary feedwater, how many of the tubes 15 are wetted, that analysis is going to be expensive.
Okay?
16 Something that perhaps really ought to be looked 17 at in terms of separate effects type calculations is where 18 you try to develop a predictive capability that can 19 determine the wetted versus unwetted regions and then 20 hardwire that into your input for a given transient.
21 MR. WARD:
Is there any way, Virgil, Ivan, to 22 find out how to get a better handle on how important this 23 is?
I guess I'm sort of sensitivity simpleminded.
The TRAC 24 would be run, trying to figure out what a key transient 25 is and make sensitivity analysis, but that is probably
I 5071 03 09 40 1
AGBagb l-pretty clumsy and expensive using TRAC.
Is there any other 2
way to do it?
3 DR. CATTON:
I asked Thad the question yesterday
()
4 as to how important this was and he said very so I think the 5
sensivity --
6 MR. WARD:
How do you know?
I don't know how you 7
know how important it is.
8 DR. CATTON:
He said it yesterday.
Maybe he'll 9
say it today.
10 DR. KNIGHT:
It is very important.
11 MR. WARD:
Now back that up.
12 DR. KNIGHT:
I don't have the slides to show you 13 but the original OTIS calculations that we did -- which were 14 conducted rather naively on our part -- consisted of a C) 15 single channel steam generator model, it is a single loop 16 test facility, a single channel steam generator model, high 17 elevation.
The high elevation minimizes the effect of the 18 cross-flow but the single channel model we did not calculate 19 the pressure coming out, the primary system pressure coming i-20 out on the same plateaus that the data showed.
And it was 21 only when we went to the two channel'model, three wetted, 16 22 non-wetted, that we got the code to come out on the correct 23 pressure plateau.
24 And if you will look back at the presentation 25 from yesterday, there were pressure plots for OTIS test
_~
_. _ _ _.. _ ~ _. _ _,
5071 03 10 41 1
AGBagb 1
2201, if I run a single channel model, you know, just draw a 2
curve that fares in substantially lower than that and it is
~
3 on the order of a megapascal, is my recollection, and that
()
4 lower pressure leads back into your auxiliary feedwater and 5
a variety of other things, it changes the details of the 6
transients, the course of the transients slightly.
7 Interestingly that calculation refilled at about 8
the right time, because the calculation ran out seven hours 9
and it refilled.
10 DR. CATTON:
But if it is very important, that 11 seems to me that should point you in the direction of 12 developing a model to accomodate this important aspect that 13 you have found.
And that's not being done, at least I don't 14 see that it's being done.
15 DR. KNIGHT:
If my approach in terms of looking 16 at it from a multi-dimensional standpoint is the correct way 17 to do it, then yes, we are effectively working in that 18 direction to provide that capability.
19 What you need within the context of our code is 20 the absolute minimal requirement or generalized heat 21 structures.
We have a limited generalized heat structure 22 capability in the TRAC PFl code, TRAC PFl/ MODI code as far 23 as the steam generator secondary is concerned.
24 What we need is a generalized heat structure O
25 capability that allows us to thermally heat transfer-wise i
5071 03 11 42 1 -AGBagb 1
connect cells in a vessel components and cells in another 2
component.
We are working on that capability, we have an 3
in-house version of the TRAC MPA code.
()
4 The other part that is desirable from this 5
process is to have the substantive merits in the vessel 6
component at work and that is not planned at the present 7
time.
The substantive merits are in order to be able to run 8
these transients in a reasonable length of time.
Because 9
when you go to a lot of detail on the steam generator 10 secondary in order to be predictive, I think the running 11 times will go up significantly.
12 DR. CATTON:
We are talking about two different 13 things:
one is physics and the other is what you think you 14 need to do to acco=odate that.
O 15 DR. KNIGHT:
Sure.
16 DR. CATTON:
What do you need in order to be sure 17 you accomodate the physics?
It see=s to me you don't have 18 it yet, you don't have the data for one thing.
19 DR. KNIGHT:
What I need to be able to accomodate 20 the physics is some confidence that the code -- and it may-21 be a new model in the code -- that the code would calculate 22 correctly the flooding at these spacer grids on the 23 secondary side and I don't have that.
24 MR. WARD:
This discussion seems to be kind of 25 disconnected from the discussion of what the SAI auxiliary
5071 03 12 43 1
AGBagb 1
feedwater experiment is going to accomplish.
We didn't seem 2
to hear much about that yesterday.
3 DR. CATTON:
The SAI experiment is the same steam
()
4 generator that they have in MIST and, as a result, it is 5
very small in diameter.
So I'm not sure that the results 6
from it will be very helpful.
7 What Thad needs to be able to do is to track this 8
thing down --
9 MR. WARD:
But even that's better than the model, 10 that is at least a rea];<crld or a little steam generator.
11 DR. KNIGHT:
It's real little, though, because 12 it only has I think one support plate across, doesn't it.
13 If you are going to be looking at the flooding, you need to 14 look at several levels.
O 15 MR. WARD:
-- the spreading.
16 DR. CATTON:
If they are ever going to get the 17 heat -- the mass centers correctly they are going to have to 18 do that.
If it's important.
I don't see that separate 19 effects test being carried out anywhere at the present time.
20 MR. WARD:
I guess I don't have an understanding 21 of what EPRI -- of what the whole program is going to 22 accomplish with that SAI AFW test.
Can somebody explain 23 that?
24 MR. BECKNER:
Let me try to address a couple of O
25 questions.
It's my understanding the SAI test has been
5071 03 13 44 1
AGBagb 1
looking at flooding in effect in a MIST geometry.
So that 2
will hopefully help us understand MIST.
Now there is a 3
bigger question about what do we do about full scale and
()
4 what Ivan is pointing out is we do not have a first 5
principle model to predict auxiliary feedwater spray.
Thad 6
right now is just trying to work on the mechanism to put 7
some model in there.
8 DR. CATTON:
From somewhere.
MR. WARD:
Yes, he seems to be suggesting the 9
l l
10 first principle model might not be workable for TRAC.
11 MR. BECKNER:
Right, until he gets the structure 12 of the code changed.
So that's the first problem he's 13 working on that he is structuring here.
14 The second problem is in a first principle model, f
t 15 Ivan is right, we do not have a program to do that.
We, 16 right now, are relying on plant data which is good in the 17 sense that you don't have the scaling problem but it's bad 18 in the sense that, like all plant data, it is kind of hard 19 to interpret sometimes.
I think it is a matter of 20 priorities.
It is not going to be a simple or a cheap thing 21 to do if it looks like we need it.
Thad says that even 22 approached from a calculational standpoint it will be 23 expensive.
So it's something we have to look at.
24 DR. SCHROCK:
You probably should do that, 25.
shouldn't you, should try-the calculation?
1
5071 03 14 45 1
AGBagb 1
MR. BECKNER:
Again it's a matter of priorities, 2
whether or not we can afford it and the adequacy of this 3
plant data that we have.
()
4 MR. WARD:
Well I come back to the question then 5
how do you get a handle on how important it is in the course 6
of any particular transient?
I mean Thad pointed out for 7
the simple comparison, off and on, for one transient it 8
seemed to give dramatically different results but I'm not 9
sure that that, you know, gives you a real good handle on 10 its total importance.
11 DR. KNIGHT:
I think you are right and what you 12 are getting at is the need to do sensitivity studies to look 13 at that particular question.
14 MR. WARD:
Yes.
O 15 DR. KNIGHT:
And we have not done that and at the 16 current time it is not within our scope of work to look at 17 that.
18 If you really want to get a number-out that tells 19.
you how important this is, whether it's on-off or, in the 20 plant case, 8 percent versus 10 percent versus 12 percent 21 or, you know, it may be 5 and 10 and 15 percent, you've got 22.
to run sensitivity studies.
And the calculations right now 23 are fairly long running until we get something done for 2k comparison.
O 25 MR. WARD:
It seems to be a weakness in all the I
l
I 5071 03 15 46 1
AGBagb 1
experimental models, it is a weakness in the calculation 2
models and we just don't seem to know how important it is.
3 But there is some reason to believe that it could be
()
4 important.
5 MR. JONES:
If I could make one comment then --
6 and I guess when you get down to importance you get into 7
various degrees of how accurate you need to predict the 8
experiment -- but if you put in the perception of a core 9
safety standpoint, all right, I believe -- I wasn't present 10 at the meeting that the committee had where they discussed 11 the II.K.3.30 reviews -- it was my understanding that that 12 problem presented the results of the sensitivity study that 13 we performed with RELAP where we had 100 percent wetting and 14 zero percent wetting and the bottom lines on those were it 15 was still safe from a core standpoint; certainly the course 16 of the transient was significantly different between the 17 two.
And as far as a benchmark of the CRAFT 2 code, that 18 laid in-between the two which has a roughly 10 percent 19 wetting case, so it was from that that we felt pretty 20 comfortable from an overall safety standpoint.
21 Certainly define the structure, I agree with the 22 Committee, I think what needs to be done is along the lines 23 of what Thad said, I think the plant data clearly indicates 24 that the amount of wetting and spreading that you get is O
25 fairly limited, it is on the order of 10 percent, and I
l 5071 04 01.
47 l 'AGBagb 1
think any sensitivity study should be probably restricted to 2
a 5 to 15 percent wetting pattern to see any particular 3
differences.
We don't have those studies at this time,
()
4 however.
5 DR. CATTON:
I've heard some comments though that 6
some people feel by the time it hits one of those tube 7
sheets it is going to just spread over the whole tube 8
sheet.
I don't know if I believe that.
i l
9 MR. JONES:
Again the data that we have from the l
10 plant, it was. data taken at Oconee back -- way back, some in 11 the early Seventies --
12 DR. CATTON:
Was it a dry steam generator that 13 was overheated?
14 MR. JONES:
Not a dry generator.
I can't 15 remember the details of the test.
It was presented to the 16 Committee back in --
i j
17 MR. WARD:
It was a couple years ago but as I j
18 recall it was a different configuration for the feedwater l
19 inlet, wasn't it?
r i
20 MR. JONES:
Not for Oconee.
1 21 MR. WARD:
Okay.
Oconee did have the new 22
' configuration.
23 MR. JONES:
It may have been different a little 4
i 24 bit -- Oconee had the separate nozzles.
(
25 MR. MICHELSON:
The separators, though -.the f
1 i
I e
+-
r
-f.v
-....-w&,-.g
.-m wv.
+vmm
,,,.-,we,,,
s.+w,e
_g.+ea-,
,,...-#---w.w,,,, +,,
5071 04 02 48 2
AGBagb 1
support plates, rather, were I thought significantly 2
different on the steam generators than the one in your MIST 3
test facility.
4 MR. JONES:
I would argue that the differences in 5
the MIST facility, the support plates are s=all, it is =cre 6
in structure.
They are drilled a little bit different but 7
the representative flow areas were preserved, to =y 8
knowledge.
9 MR. MICHELSON:
I was thinking in ter=s of the 10 tube denting experience that have teen observed on a nt=ber 11 of these plants.
And I thought that this tube denting in 12 part was coming from the quite a bit of different spacer 13 design.
And of course that area is becocing an area for 14 deposition of crud and whatever and it is causing denting.
15 And by the ti=e you have a practical stea= generator that 16 has been running for a while, I think the configuration 17
=ight be significantly different unless there's no crud 18 buildup in the upper part, which is even more critical.
19 MR. WARD:
Carl, that would drive you in the 20 direction of getting = ore spreading, is that what you're 21 saying?
22 MR. MICHELSON:
Does B&W see =nch tube denting in 23 the upper part of the steam generator?
24 MR. JONES:
I'll turn that over to B&W.
25 MR. MICHELSON:
I would think not but I don't
5071 04 03 49 1
AGBagb 1
know because, you know, it is a undoubtably vapor phase up 2
there.
1 3
MR. TURNER:
We have almost no information on
()
4 that, we don't have the people here to address it but I 5
don't believe I can recall much discussion about denting.
6 We can look at that later.
7 MR. MICHELSON:
Yes, I don't think you're getting 8
it much like Westinghouse sees because theirs, of course, is 9
submerged.
And if the spacer plates are nice and open like 10 you're using in MIST then I don't know how you get that 11 much spreading.
You know, the countercurrent flow idea is 12 fine but the water is only around the edges to begin with.
13 It doesn't get very far into the tube bank.
14 MR. JONES:
The plant data clearly.showed O
15 spreading.
I mean, it is very clear, Carl.
Because I think 16 the initial penetrations were on. the order of 2 or 3_ percent 17 and by the time they had dropped through several of these 18 support plates, they were -- it was out to around 10.
19 MR. MICHELSON:
It's hard to see how it spreads-20 much with as open spacing as you have on MIST and if it's 21 that open in the real world and if there is no deposition --
22-which perhaps is indicated by no denting, I don't see how 23 you can spread across the sheath.
24 MR. JONES:
I'think arguably the SAI tests --
D
- k-25 which I'm not exactly familiar with, with the detail -- one u _ _______._._____.__
8 5071 04 04 50 1
AGBagb 1
of its intents is to look at that sort of countercurrent
{
2 flow holdup, CCFL problem, at the tube support plate.
3 That's one of the objectives, they are trying to para =atize
()
4 that to find out what happens.
5 MR. MICHELSON:
There is another problem -- not 6
problem but another factor and that is I thought you had a 7
large gap between the support plate and the vessel.
And the 8
water coming in on auxiliary feedwater is coming in at the 9
edges and is not really even being sprayed much, the 10 calculations indicate it is almost like a dribbling or a 11 pouring in, not spraying in a pattern.
12 MR. JONES:
All I'can add on that is the plant 13 data did show the sprey, it did show the penetration through 1
14 several tubes.
15 MR. MICHELSON:
That was =y understanding.
16 MR. JONES:
There was separate effects data which I
17 tried to para =atize the amount of initial wetting as a 18 function of flow rate.. And in fact the two different 19 nozzle configurations that were out there at the time of the i
20 initial penetration precent-wise was about the same; it 21 varied linearly or more or less linearly with flow rate.
22 MR. MICHELSON:
The nozzle spraying was-only 23 going in, I thought, two or three tubes deep.
24 MR. JONES:
I think'at the very high flow rates
' O 25 it could go in at about four or five.
But you are correct, 1
l 1
I l
l 5071 04 05 51 1
AGBagb 1
the initial penetration, the maximum I think that I remember 2
seeing of representative flow rates is probably no more than 3
4 to 5 percent of the tubes, because it goes in and it
()
4 spreads a little bit but it is on that order and then it 5
falls through the support plates, it does spread out a 6
little bit and the data tended to indicate a spreading of 7
roughly 10 percent.
8 I will agree with you the data is difficult to 9
interpret, it was only that the instruments were along one 10 row of tubes going into the bundle and they had some 11 thermocouples in there and you could definitely tell the 12 distinction between wet and dry.
Now how far it spread to 13 the sides, there wasn't information and it was inferred that 14 the spreading patterns into the bundle would equal more or O
15 less the spreading pattern out.
And it is an assumption.
16 I think a sensitivity study, looking at varying a reasonable 17 percentage of wetting would be appropriate.
18 MR. MICHELSON:
It may be that the steam is 19 transporting rather sizable water droplets to other tubes 20 that weren't seeing the spray at all and that it was running 21 down those tubes and that's what looked like a spreading of 22 the water.
23 DR. SCHROCK:
Of course it has to run down in 24 order to generate steam in the lower levels.
25 MR. MICHELSON:
Oh yes, but it is generating the
5071 04 06 52 1
AGBagb 1
steam along the periphery, the predominant water spray is.
2 DR. CATTON:
How hot was the steam generator when 3
they ran these tests you are referring to?
()
4 MR. JONES:
As I remember it, Ivan, they were 5
tests that were run at about 800 psi in the secondary side.
6 It was pretty much a hot generator.
7 DR. CATTON:
How much superheat of that pressure 8
were in the tubes when you started to wet them?
9 MR. JONES:
There was some superheat, I just 10 can't remember it now because there were some tests looking 11 at the desuperheating and how long it took to desuperheat 12 the steam, and the inference from these tests was by the 13 time it hits that first support plate or just shortly 14 thereafter it was desuperheated.
15 DR. CATTON:
The amount of superheat is going to 16 have a big impact on the CCFL.
I would think that along 17 with the sensitivity studies somebody might do a few hand 18 cainulations just to try to see how long the fluid would be 19 held up.
20 That probably, in my view, would be more useful 21 than just the straight sensitivity study.
22 MR. MICHELSON:
The key question, of course, is 23 does it really make a big difference?
24 DR. CATTON:
But he says it does.
25 MR. MICHELSON:
Well he thinks so.
l 5071 04 07 53 1
AGBagb 1
MR. WARD:
We don't know.
He knows that black 2
and white makes a difference.
3 DR. CATTON:
Well the straight 10 percent versus
()
4 that truncated 10 percent.
5 Is that right, Thad?
The straight 10 percent 6
wetted versus the truncated 10 percent, is there a 7
difference between the two?
8 DR. KNIGHT:
Are we talking about the OTIS 9
calculation I mentioned or in the plant?
10 DR. CATTON:
The calculation you did where you 11 reduced it from 10 percent wetted down to 4 percent wetted 12 or something.
Was there a difference between those two?
13 DR. KNIGHT:
We did not run a differece.
14 DR. CATTON:
Oh, then you're r.ight.
O 15 DR. KNIGHT:
The plant calculations was input 16 with 4 percent wetted at the auxiliary feedwater location 17 and then it went down to 10 percent I think, four levels 18 below that.
19 MR. BECKNER:
I think Dave's assumption is 20 correct, black and white is a big difference, we don't know 21 about shades of gray.
And that's the sensitivity study that 22 might be appropriate to run.
23 DR. CATTON:
I think that, combined with a little 24 bit of hand calculation, might bring it all together.
O 25 MR. BECKNER:
I agree for MIST.
For plant, I'm 1
5071 04 08 54 1
AGBagb 1
not sure you're going to have one initial CCFL.
That may be 2
more difficult.
For MIST, I think you could use one.
3 DR. CATTON:
I think each one of those tube r
4 sheets acts kind of like a core support plate in a BWR --
5 MR. BECKNER:
Right, but it is a very large one, 6
it is a very, very large one and you may have like what we 7
saw in BWR's where --
8 DR. CATTON:
-- you flood the whole thing.
9 MR. BECKNER:
-- no, where we couldn't flood a 10 large grid by heat bypass because it was so large.
11 MR. MICHELSON:
But the heat source configuration 12 there is totally different for a BWR from a physical 13 perspective.
14 DR. CATTON:
The physics is the same.
15 MR. MICHELSON:
Yeah, if you could localize the 16 fluid down to a small --
17 DR. CATTON:
But they have, because they've built 18 models for the upper region of the EWR.
19 They have the models, they just need to I think 20 apply them in this other case.
21 MR. WARD:
The physics is the same at the F 22 equals MA level.
23 DR. CATTON:
Yes, it's the same at the CCFL 24 level.
25 MR. WARD:
Our next -- Let's see, as I understand
l 5071 04 09 55 l
1 AGBagb 1
it, Richard Lee tells me that what would have been covered 2
in le has already been covered, so I think our next 3
presentation would be by Lou Shotkin.
And Lou, if you are
()
4 not offended, we will take a break before we proceed, is 5
that all right?
6 MR. SHOTKIN:
Fine.
7 (Recess.)
8 MR. WARD:
Let's proceed.
9 (Slide.)
10 MR. SHOTKIN:
I will be talking about what was 11 put on the agenda and it is called Assessment of Codes:
How 12 and When?
I will be using slides that the Subcommittee has 13 seen before, we haven't gone over much detail on any of them 14-and I will refer to codes in the MIST program.
O 15 (Slide.)
16.
This was one of the slides which was seen before 17 and this I can use to' address the assessment of the two 18 codes that are being used on the program for RELAP5 code and 19 the TRAC PWR code.
20 I think, if I can interpret, what you are 21 interested in is how the MIST result is going to be used to 22 assess the codes.
The point I would like to use here is 23 that there is data from many other facilities that are being 24 used to assess the same models in both TRAC and KELAP that 25 are then also being used on the MIST program.
5071 04 10 56 1
AGBagb 1
The only purpose of this slide is to bring in our 2
international code assessment program, which was mentioned 3
yesterday as being a valid source of assessment for both
()
4 TRAC and RELAP for models that are also going to be used on 5
the MIST program.
And if you agree with that, then I can go 6
to the next slide.
7 That's all this slide shows, that there's several 8
facilities besides MIST, there's ROSA IV, Semi-Scale, LOFT, 9
and so forth.
CCFL models being tested on 2D/3D, you'll 10 hear more about it next week.
They are all going into TRAC 11 and RELAP, there's University research and then there is the 12 inter' national code assessment program, so don't forget those 13 as being sources of assessment.
14 MR. WARD:
I guess the only trouble with that is O
15 as we are hearing here and as you know we have been hearing 16 the unique characteristics of the B&W plant don't really get 17
-- all of them don't really get treated in these other 18 assessments.
I mean there aren't any -- I mean, the thing 19 you were just talking about, the once-through steam 20 generator just isn't dealt with in any of these 21 international code assessments.
22 MR. SHOTKIN:
That's correct.
We do need the 23 MIST data to finish the assessment for the B&W plants.
But 24 we are also using data from many others.
\\J 25 (Slide.)
5071 04 11 57 1
AGBagb 1
This is the overall plan for at least the 2
schedule of the assessment.
I just refer to the TRAC PWR 3
and to the RELAP5 lines which show from fiscal '84 and
()
4 through fiscal 1990 and, if we are ever permitted to go with 5
such a long-range plan, what we would expect is that our 6
final frozen version of TRAC PWR and RELAPS would be 7
released in 1990.
8 The first frozen version was released over here 9
at the end of fiscal
'84, actually the beginning of fiscal 10
'85, the last planned version with TRAC PFl/ MOD 1 and 11 then RELAPS/ MOD 2 were released at the end of calendar year 12
'84 and those codes were frozen.
13 The definition of a frozen code in that no model 14 improvements are allowed in that frozen version, just error O
15 corrections and user-oriented corrections and that is 16 because we want to define the uncertainty of these code 17 versions on our international code assessment program and 18 our domestic code assessment program.
So that when 19 questions come up like the one that came up today, how 20 important is the modeling of the auxiliary feedwater on 21 plant safety analysis for B&W plants, we would like to be 22 able to answer that question in a quantitative way rather 23 than with hand-waving arguments.
I 24 So that's the reason for having the frozen
-C )
25 version is that every calculation that gets done by several i
l
l
-5071 04 12 58 1
AGBagb 1
organizations can all be used to generate a statistical data 2
base to define the code uncertainty.
3 MR. WARD:
Let me ask you:
the version of RELAP5 4 ()
4 that the B&W Owners' Group is using, I mean there is more f
5
.than one version of RELAP5 to date and I get -- theirs is 6
tuned to their needs, in other words, tuned to the B&W 7
plant, I gather, is that right?
l 8
9 I
10 l
11
]
12 l
13 e
14
()
15 i
16 17-18 19
-20 21
'22 23 24 0
25
5071 05 01 59 1
AGBeb 1
MR. SHOTKIN:
They check the models that they put 2
in the REDBL code, the ones that I remember--
3 MR. BECKNER:
They aren't using the REDBL any
()
4 more, are they?
l 5
MR. SHOTKIN:
As I say, the models that they used 6
in the REDBL code, primarily the feedwater--
There was one 7
other.
And they put those models into RELAP 5, Mod 2, and j
8 they are using that version of the code on the MIST program.
9 Now you say that is a different version than what 10 we are giving out under the Code Assessment Program, and 11 that is correct, and there is not much we can do about that.
l 12 MR. BECKNER:
Lou, they used, though, RELAP 5, l
13 Mod 2, Version 36, and I think there might be a.1 now.
l 14 They use as their basis the same version,-that 36 that is 15 used under the ICAP, and it's frozen.
16 MR. SHOTKIN:
Right.
17 MR. BECKNER:
The point is that they put their 18 own models in, and we can't dictate to anyone ' hat they do w
19 or don't do with the code, or what models they put on.
On 20 the international scene, when we sign our agreements, we I
21 request that all the official calculations be done with a i
22 frozen version.
23 If somebody is desperate and wants to put in a i
24
'model that they feel is going.to solve'all their problems, O
25 then there is not much we can do about that except we ask
(
5071 05 02 60 1
AGBeb 1
them to compare that result with the frozen version and we 2
would be interested to see it, but we still would want the 3
frozen version.
l'h
(,)
4 MR. WARD:
But if your frozen version hasn't 5
picked up some of the unique characteristics of the B&W 6
plant, it is not perhaps particularly useful to make that 7
comparison.
8 MR. SHOTKIN:
Our basis for choosing a frozen 9
version is that we have worked on these codes for now over a i
10 decade, and we do believe that we have not the best models 11 but pretty good, representative models for almost all the 12
. phenomena.
And when you run into a problem--
Let me just 13 do the auxiliary feedwater.
14 When you run into that some people, perhaps the O
15 code developers or the code assessors, might say the only 16 way I can do that is to really put in a multidimensional 17 model of the steam generator secondary.
And if you think 18 very carefully about that, you have to ask where is the data 19 going to come from to assess that?
What is the running 20 time?
How important is that?
21 You can also handle that by taking the existing 22 models and putting in two parallel pipes, as was done, and l
23 just force the flow distribution to go down and then do a 24 sensitivity study.
And it seems to me that could be done
. /'}
25 with a frozen version of the code as input, after being I
l l
l'
l 5071 05 03 61 3
AGBeb 1
done, to input without going the route of going through very 2
extensive modeling changes.
3 And I think that what we're saying is that from
()
4 here on, what we would like to do is try to handle most of 5
that through the existing code versions, do sensitivity 6
studies, and then after that, if we really feel that 7
something has to go in there, then we would put it in.
But 8
it wouldn't get in--
9 The next frozen version is planned for fiscal 10
'87, and then the final frozen version would be done in 11 1990.
12 DR. SCHROCK:
But anything that is going to go 13 into the '87 frozen version needs to be under development in 14
'86.
O-15 MR. SHOTKIN:
That's correct.
There are things 16 that are going in right now.
It is just what.gets used for 17 official calculations.
18 The other thing I would like to point out is that 19 during the last decade when we were pushing the~ code 20 developers to come out with new versions, they couldn't do 21 it faster than about 18 to 24 months, so keeping it frozen 22 for 24 to 30 months is not really that much different than 23 what we were doing in the past.
It is just when you tell 24 people about it, they tend to get more upset.
25 MR. WARD:
When you talk about freezing a PWR
5071 05 C4 62 2
AGEeb 1
code, either TFAC cr FE.IAF, that is really a WR 2
Westinghouse or Cc=bustic: Er.ginee ring, and it =.ay be that 3
eventually veu vill have to ecc:.e to crirs with whether there s
4 is going to be a second FWR versicc.
5 You said you think you can hacdle it by this 6
other approach, the sensitivity apprcach, but if, for 7
exanple, in this whole IST progran, you kncv, you find that 8
there are such inpertant differences in the behaviers of the 9
two types, I can just see where that potentially cc:1d 10 develop.
11 MR. SEOTKIN:
We vill be alert to that.
We had 12 to put in a separate versicn of the code for bciling vater 13 reaeters, and if we have to do it fer EEW desic s-If, fer 14 eranple, the MIST prcera= leads to uttinc. in se r.any y
15 specialized =cdels fer ELW, then ve vill prcbably have to 16 s lit off separate versiens.
17 MR. WAFS:
Okay.
I think that's all I*n 18 interested in.
19 MR. SEOTKIN:
We veuld rather de it that vay, but 20 it could be.
21 (Slide.)
22 At the last reeting here the questien care np, 23 what's ccr future plans for TFAC and FJ.lAF?
We discussed 24 this with the Subecc=ittee and what I*n doing is nere fcr 25 FMG cr the eccle who were at the last PMG neerine..
I will r
5071 05 05 63 1
AGBeb 1
just go over it for you again.
2 Since the last PMG meeting we have had to make a 3
choice for internal NRC budgetary reasons between TRAC PNR
,(,)
4 and RELAP 5 as the code of choice for future plans by NRC.
5 And the choice that we made was TRAC PWR, and here is a list 6
of the reasons.
7 We felt that it is a mature code.
It has been 8
extensively assessed.
These are absolute statements for l
9 TRAC.
There are also relative statements between TRAC and 10 RELAP.
11 It has a three-D capability.
It has more 12 advanced numerics.
The SET method is working already, and 13 the one-D components, and it is -- at least in the three-D 14 components it is not yet ready for production use but it 15 is there.
16 There was no discussion on the BWR code.
We were 17 going to continue TRAC BWR so that was another reason to 18 continue TRAC PWRs.
It is a companion code.
19 We went through this twice before and we chose 20 TRAC back ir 1978, and again in 1984.
We feel it can 21 address all the perceived thermohydraulic safety issues, so 22 we are going to continue plans for the desktop plant 23 analyzer.
24 What we are going to do is fund its use at fm Q
25 Idaho.
We are already using it at Idaho, but to try to
5071 05 06 64 1
AGBeb 1
encourage them to use it more.
And this is part of the 2
concept of having Idaho as our thermohydraulic technical j
3 integration center.
l rm
(_)
4 DR. SCHROCK:
When you made a similar decision a 5
couple of years ago then NRR said if you don't continue to 6
maintain RELAP 5, they would.
Isn't that correct?
Is 7
that--
8 MR. SHOTKIN:
They may have made a--
I never 9
heard that strong a statement.
10 DR. SCHROCK:
I think t heard that in one of our 11 meetings.
12 MR. SHOTKIN:
What they requested us to do was 13 maintain both codes.
What they have agreed to is this 14 choice.
This is an NRC choice, and that includes NRR.
15 DR. SCHROCK:
It includes NRR now?
16 MR. SHOTKIN:
Yes.
17 MR. JONES:
I won't say we have officially 18 endorsed the choice as I am not aware of any paper that has 19 gone back and forth.
We have caved in, so to speak, on our 20 strong statements on RELAP 5.
21 The plan they have is to continue to maintain 22 RELAP 5 anyway.
It is just not going to be further 23 developed, which was our primary concern, because our 24 contractors use it a lot.
('-)
25 MR. WARD:
Lou, do any industry groups, vencors
l 1
5071 05 07 65 1
AGBeb 1
or owners' groups or anything, fuel suppliers, use TRAC 7 2
MR. SHOTKIN:
Let me--
Thad, if I'm saying 3
anything wrong, let me know.
()
4 I have just heard from Los Alamos that while 5
certainly GE uses TRAC BWR -- that you know,--
6 MR. WARD:
They do.
7 MR. SHOTKIN:
-- as a benchmark for their safety 8
code, I've heard that Exxon is planning to use TRAC, but I 9
just heard that from Los Alamos, and I don't know what that 10 means.
11 Do you have anything more on that?
12
- MR. KNIGHT:
We have had a couple of days of 13 discussion with the people from Exxon looking to develop 14 some kind of new licensing-type tool.
Whether or not they O
15 made the decision to use TRAC directly, I'm not sure, but 16 they have made the decision to acquire TRAC PF-1, Mod 1, and j
17 work with it.
They seem to be gearing up a fairly 18 substantial effort in that direction.
i 19 MR. SHOTKIN:
That's sensible.
20 MR. WARD:
All right.
i 21 MR. SHOTKIN:
Now B&W of course uses RELAP 5.
22 Westinghouse is thinking of using or they are using the 23 COBRA TRAC code.
And when I say "use," you will have to ask i
24 them what that means in terms of licensing.
What it means 25 to me is that they have got that code from us; they have i
5071 05 08' 66 1-AGBeb l'
told us that they're using it.
Now whether you ever see any 2-calculations that they have done with it I don't know, so I 3
can't speak from a regulatory point of view.
-()
4 And as far as the utilities, I would say that 5
there are -- maybe if we discount the B&W Owners' Group, 6
there may be 10 or 12 utilities that use RELAP 5.
7 There is one utility that is interested in TRAC 8
BWR, but again I don't know what for.
I l
9 I don't think any utilities use TRAC PWR.
10 The foreigners, their governments--
Well, the 1
i 11 major European and Japanese are more interested in TRAC PWR l
12 because it helps them on these joint programs..And many of 13 the smaller European' companies use RELAP.
i
! ()
14 (Slide.)
15 our plans for RELtP 5 are to continue improvement 4
16 and maintenance support at a reduced level than what we had
+
17 planned for; that is, decrease direct NRC support.
And what l
18 we will try to do is get-increased user support, which would i
I
-19 come from a domestic users' group that Idaho is now forming 4
_20 for RELAP 5.
4 I
21 The ICAP, the International Code Assessment e
I I
j
-22 Program, does have its own funding for model improvements i
23 and we wo'uld funnel that money, some of that, into RELAP 5 i
24 as it is needed.
25 And then as.a test facility has its own i
l T
'. A 1
h
5071 05 09 67 1
AGBeb 1
individual needs, such as MIST or Semiscale, if we have any 2
say over the funding, we would try to funnel some of that.
3 Now we're talking in '86 of small amounts of
()
4 money.
This reduction is slight in '86.
It will be more in 5
'87, and by '87 or '88 it will reach asymptotic low support 6
levels.
7 Complete plant analyzer capabilities in '86.
The 8
major thing we had to do when this decision was made was to 9
reduce parallel efforts that we have with both TRAC and 10 RELAP on certain programa.
The major one was on ROSA-IV, so 11 we just cut out the RELAP 5 analysis with ROSA-IV.
12 (Slide.)
13 That is really the end of my presentation.
I 14 have this last one which answers the question "When?"
O 15 Here is MIST.
Beckner has a presentation later 16 where we are going to say that NRC is interested in a 17 follow-on program with MIST.
But if that does go on and we 18 do get some full-power follow-on testing in '88, we would 19 feel that by '89 we would be finished with the assessment of 20 the codes for B&W.
21 MR. WARD:
Okay.
Thank you, Lou.
22 The next item on the agenda is NRR comments on 23 MIST test groups, and I understand Bob Jones just has a 24 couple of words to say.
25 MR. JONES:
We discussed, I guess last October,
5071 05 10 68 2
AGBeb 1
the NRR comments on the test matrix.
Things have not 2
changed as far as the matrix at all.
At that time we 3
concluded that the test matrix was adequate for addressing
()
4 the TAG issues, and nothing has changed relative to that.
5 There were a few items that we had comments on.
6 Just to update the Committee, one of the items 7
that we concluded back in October was that the test 8
initialization parameters needed to be further verified.
At 9
that time there was -- like the initial system pressure was 10 up in the air, exactly what was going to be used in the 11 facility.
That choice has been made, and several other 12 items on the initial conditions, so they are now frozen, and 13 we agree that they are reasonable.
14 Another comment we had was that the test conduct O
15 needed to be established for the specific tests, including 16 the pre-test analysis and ATOG.
There has been substantial 17 discussions at various PMG meetings on ATOG implementation.
18 They have not been completed.
19 At the last PMG meeting, in May I believe it was, 20 several people, me, Jim Quatermans and I forget who else, 21 wrote down several comments on how to integrate ATOG within 22 the specific test plans.
We have not seen a composite or 23 recommendation yet from the PMG or B&W on what they are 24 specifically choosing, so that it still an outstanding item O
V 25 which should be resolved whthin the next few months,
[
5071 05 11 69 1
AGBeb 1
probably by the next PMG neeting.
2 he other ite= vas candidate tests for pre-tes:
3 analyses were being selected at that ti=e bark in October.
4
~ hat selection process took plare I believe in Cece:ber in a 5
technical neeting.
You have seen that infer =ation presented 6
by Bill and Richard yesterday.
hese are the current test 7
plans.
They see= to adegaately e:::ver the various =ajor 8
testing groups and shocid be sufficient for the planning 9
calculations.
10 So basically we feel that these things are 11 pregressing fairly well and the text natrix is fairly well 12.
in hand at'this tire.
13 MR. WARD: _ Okay.
Thank you, Ecb.
14 Well, that brings us to our next presentatica en 15 MIS' hardware status.
16 Randy Carter.
17 MR. CAN:
Most of these tcpics are really a 18 further expansion of what you saw cn the tour yesterday, sc 19 I think = cst of it can be pretty brief.
20 (Slide.)
21 One thing I did want to talk about a little bit, 22 I showed you the core heaters yesterday and promised you I
-23 would tell you a little bit =cre abcut the prcble: that we 24 had in the -initial installatica of the core heaters.
25 What you saw yesterday was really cur seccad
5071 06 01 70 1
AGBeb 1
attempt at building the core.
We started back in March on 2
the first attempt, and we had a problem where we had a 3
short--
Normally as you go through the heater there should
-( )
4 be a resistor of approximately 2-1/2 ohms between the 5
lead-in wire and the heater rod sheath.
When we installed 6
the first 17 of these heaters, when we checked those we had 7
basically a dead short.
8 After quite a bit of metallurgical examination 9
and whatever, we found that we were creating a short right 10 in the region that we were doing the braze that-I discussed 11 yesterday.
We found out that the major problem that was 12 sort of hidden, the Achilles heel, if you will, was that we 13 had a temperature indication there, or a shift in the 14 calibration of this optic system that we take back to an O
15 infrared sensor, and that had been caused by--
16 I mentioned to you the braze material is put on 17 there.
There is also a flux material that is put on here to 18 help the braze flow.
Initially that flux material 19 contaminated the end.of this fiber optics, changed its 20 transmission characteristics, and caused a shift in 21
. calibration.
When we thought we were brazing.at about 1850 22 degrees, in actuality we were-at about 2200 degrees.
23 And'that caused at this point this~ copper lead-in 24 wire-to melt and create a short from the lead-in wire to the 25 sheath material.
)
N 1
i 5071 06 02 71 i
1 AGBeb 1
That was fairly easily taken care of by 2
controlling the amount of flux that was used.
3 MR. WARD:
That is not a coil there?
That is
- ()
^
4 just a solid wire there.
5 MR. CARTER:
Let's go through the geometry just 6
for a second.
l 7
We are looking at the very lower end of the 8
heater rod as we saw it out there.
This is what appeared to 9
be silt out there, which is a nickel-plated copper solid 10 rod.
The material in between is a ceramic to boron nitride.
11 The sheath material is Inconel 600, and the extension tube 12 is an Inconel 650.
And that is where we create that 13 pressure boundary between the Inconel extension tube and the 14 heater sheath.
O
'5 We then went back into the process of convincing 1C ourselves that we had solved the problem once the 17 calibration had been corrected.
We found out that we now 18 got sporadic failures.
We were seeing like one out of ten 19 failures.
20 We went into quite a bit more metallurgical 21 examination.
This brings us up just about to an April-May 22 timeframe.
Ultimately we found out that the nickel plating 23 had traces of phorphorus in it.
Phosphorus at 10 percent 24 mixed with nickel will cause a eutectic that melts at about O_
25 1616 degrees.
5071 06 03 72 I
1 AGBeb 1
As we examined over lengths of this lead-in wire 2
that has the nickel plating we've seen traces of phosphorus 3
up to as much as 10, 12 percent.
And that has given us the
()
4 problem of sporadic failures.
5 What happens is you braze at 1850.
If you have a 6
condition of a 10 percent phosphorus concentration cause a 7
crack in the ceramic during the braze, you can still short 8
across there and that has been the problem that we have been 9
facing.
10 At the end of April or early May we had a PMG 11 meeting and the technical solution was very obvious.
You 12 get rid of the nickel, but that is difficult on the heaters 13 that are already fabricated.
We did consider replacing the 14 core with only a copper lead-in but based on our work to O
15 date, we showed that we should not have but a roughly 20 16 percent failure of those heaters that we attempted to put 17 in, and that's what we tried to do.
18 (Slide.)
l.
19 We started brazing the heaters in again about the I
20 middle part of May.
We went very cautiously at first, just l
l 21 to try to build our statistics to make sure that we weren't l
22 going to have an overwhelming number of failures.
To date 23 we have attempted 17 brazes.
Thirteen have been acceptable, 24 and those are the ones that you saw out there yesterday O-25
. afternoon.
5071'06 04 73 l
1-AGBeb 1
Four were unacceptable.
The heaters were I
2 shorted.
We took them out and replaced the ones that had 3
shorted out.
()
4 It is building that we are running somewhere 5
around 23 percent, 24 percent failures, and'with this kind 6
of failure mode, we can continue with the spares that we 1
I 7
have on hand or that are coming in from Rama.
So we plan to i
8 continue this mode of fabrication.
i 9-Just a couple of comments about--
I would like i
10 to go back to that other slide for a second.
1 i
11 (Slide.)
12 One of the questions that we certainly asked i
13 ourself also is is there any chance that this could become a j
14 problem later on?
The.only time that-this part of the g
[i 15 heater rod sees any' temperatures over about 300 degrees is 16 actually during the brazing process.. Typically this part of 17 the heater rod has only a-small amount of overheating, and i
18 we've calculated it to be up around 300 degrees versus thel 19 1850 degrees that it sees during brazing.
20 The other thing we have done is we do very i
f 21 careful tests of the resistance before and after the 22 brazing.
In addition, several of the rods after the brazing i
f 23 have now' been taken up to 1,000 degrees and held for 30 or l
24
- 40. minutes just to make sure-that we don't have other O
25 unknown problems.
j
(
I I
I I
._ _.. _ _ _ _._.._.. -... _ _. _ _,, _,., _,. _ _ _ _ _ _ _ _ _, _ ~ _,,
I I
i 4
5071 06.05 74 1
AGBeb 1
DR. SCHROCK:
You have made bundles before for i
I 2
your CHF tests.
Why is this a new development?
3 MR. CARTER:
The bundles that we made for CHF p()
4 were a little different.
Those were electrically-heated I.
5
-resistance versus an internal heater design.
6 DR. SCHROCK:
That was the sheath heating?
7~
MR. CARTER:
Exactly.
Exactly.
It was the skin 8
heating versus the internal heater design.
The whole 9
construction of the CHF bundles are much, much different 10
.than this type.
11 MR. WARD:
What is the maximum temperature you 12 could get to in the heater?.
13 MR. CARTER:
In the heater part itself, which is 14 well above this sketch?
(
15 MR. WARD:
Yes.
16 MR. CARTER:
As I mentioned yesterday, we did 17 test them up to 1500 degrees in air.
The manufacturer 18 suggests that we don't exceed-1800 degrees.
In our 19 calculations if the core is completely voided with steam 20 cooling, we calculate right around 1100 to 1200 degrees as 21 the expected skin temperatures at the kind of powers that we 22 will be testing.
23-MR. WARD:
What about'at higher powers?
I mean 24 if there is a_ follow-on test series,'these heaters are rated 25 for the higher powers I guess.
5071 06 06 75 1
AGBeb 1
MR. WARD:
That's right.
Again this is still 2
basically steam cooling.
I'm not positive of the power we 3
calculated--
I know the power we calculated at.
For 10
()
4 percent power we were at about 1100 degrees.
I think we 5
also did a full-power calculation, and it depends on where 6
you catch it in the K-curve.
And we were assuming some time 7
after trip.
8 So I can't tell you what it would be if we had it 9
in steam at 100 percent power.
We would have to look at 10 that calculation.
11 MR. WARD:
Water at 100 percent power?
12 MR. CARTER:
In water at 100 percent power, you 13 are very, very close to the actual fluid temperature.
The 14 coupling between these heater rods and the fluid is very 15 close.
The only concern would be as you trip the power from 16 100 percent, at what point would you expect to start 17 generating voids in the core?
And that would be really a 18 more detailed calculation.
19 MR. MICHELSON:
Why didn't you chose a lower 20 temperature solder?
21 MR. CARTER:
We looked at that.
That was one of 22 the things we've looked at also.
But typically, the lower 23 temperature braze material--
This is what is called BAU-4, 24 which is basically a gold braze.
Most of the lower k-25 temperatures get into the silvers, the tins, and they do not 1
1
5071 06 07 76 1
AGBeb 1
react very well at all with Inconel.
2 Typically at the initial point of braze you would 3
cause a fish-mouth fissure that will just open up right at 4
the Inconel sheath, so lower temperature brazes give you 5
problems with Inconel.
6 We looked at that very early in the game and then 7
after this problem cropped up, we re-looked at it again, and 8
nothing is immediately available to use.
9 (Slide.)
10 I am going to be switching subjects on you quite 11 frequently here just to update on a few things.
12 The reactor coolant pumps.
Yesterday on the 13 tour I noted that the reactor coolant pump casings were in 14 place.
We, at the point I made this slide up, had had the 15 impeller mockups on.
We have now taken those off and put 16 the gaskets in for the hydro that is going to be coming up 17 for a code HYDRO next week.
18 The impeller mockup will be put on for our hydro.
19 The four pumps right now are sitting at the 20 manufacturer, which is Chempump in Warrington, 21 Pennsylvania.
They are assembled.
22 The first pump has gone through what we call the 23 performance test which includes demonstration of the desired 24 head / flow characteristics, looking at the irrecoverable O
25 losses, both for forward and reverse flow when the rotary is
l 5071 06 08 77 1
AGBeb 1
locked in position, and also looked at the net positive 2
suction head requirements, which is not particularly 3
important for the tests that we're doing but could be
()
4 informative when we get out into the pump bump area.
5 The pump is ready for what we call the acceptance 6
test where now we take it from pretty much just the cold 7
water test, putting them at the pressures and temperatures 8
that we plan to run in MIST, an eight-hour run in a water 9
condition, and then a four-hour run where the impeller 10 cavity has only air in it.
And this is to make sure that we 11 can keep the motor part cool just as we would be doing in a 12 steam environment in MIST.
13 (Slide.)
14 The instrumentation aspects.
O 15 I am going to talk on conductivity probes in a 16 little bit more detail than some of the other stuff since 17 we've done a little bit of work since the last time we've 18 talked with this group.
Tom Larson is also going to talk, 19 after I finish, about the densitomoters and the cool TCs in 20 a little more detail than I will cover.
Those are the two 21 items that Idaho is going to be supplying to the program.
22 We are right now at the point where the 23 instrumentation, specifically the thermocouples and the 24 pressure -- differential pressure instrumentation are being 25 installed right now in the loop.
1
5071 06 09 78 1
AGBeb 1
The densitometers, the ten spool pieces for the 2
cold leg discharge for the cold leg suction in the tube and 3
the hot legs have been installed.
()
4 The electronics for the densitometers are to be 5
installed in the MIST control room during the month of 6
August.
7 And right now we are planning, roughly about 8
February or March of 1986, to install the sources for the 9
gamma densitometers.
That is maybe a little bit later than 10 we talked about before, but the main reason for doing that 11 is so that we can do some of our early debug and 12 characterization where we want to keep the gamma 13 densitometer heat losses to a minimum.
We will put the 14 sources in after that and begin using those.
15 The cooled thermocouples.
There is a total of 16 five of those that we intend to use in MIST, one in each of 17 the four cold legs and one in the downcomer to provide the 18 very low velocity conditions that may be occurring when the 19 flow is approaching stalled conditions.
20 The first of those probes has been ordered and 21 will be used for a benchscale test, and NRC is funding this 22 benchscale test separately from the MIST program.
It is 23 really going after what is the low velocity limits and l
24 accuracy of the cooled TC, and also to address the question 25 of whether a constant cooling temperature to that cooled TC
5071 06 10 79 1
AGBeb 1
is required.
2 The actual installation date has not been set 3
until after we get the results back from the separate
()
4 effects test, but we provide provisions in the facility by 5
changing the mounting technique for those so that we can 6
bring them in and install them at almost any time.
7 (Slide.)
8 Looking for a second at the conductivity probes, 9
I've got several slides on conductivity probes, just to talk 10 about some differences of what we did in MIST -- I'm sorry, 11 in OTIS and what we plan to do in MIST.
12 This particular sketch is of a conductivity 13 probe we used in OTIS, and really to talk about the end part 14 of it is the only thing I want to talk about.
The rest of k-15 it from here back is just really mounting fixtures.
It has 16 been changed somewhat for MIST, but it is not really 17 important to this discussion.
18 The actual working part of the conductivity 19 probe is the central pin separated from the body or sheath 20 by a ceramic insulator.
21 (Slide.)
22 When we install that in the facility, basically l
23 what we have is the conductivity probe body and pin which
(
24 would be exposed to the fluid conditions.
We drive it with
! b) 25 a low frequency AC current and then basically look at that i
I l
i I
l
5071 06 11 80 1
AGBeb 1
current with an AC volt meter that is taken back to a data i
2 acquisition system.
And that is really all just for 3
background.
t 4
(Slide.)
5 When we used these in the OTIS and the GERDA test 6
programs, the major thing that we were looking at was just 7
the raw voltage output from that probe.
What this 8
particular plot shows is what the voltage output versus one 9
of the transients from the OTIS program would look like, and 10 it varies considerably.
If you remember some of the curves 11 that Thad showed yesterday, you saw the same kind of i
12 characteristics.
13 Just for reference we also plotted on here the 14 temperature from one of the hot leg -- temperatures close to 15 the hot conductivity probe.
This shows somewhat of an 16 inverse relationship to what the conductivity probe is 17 saying.
18 One thing we always knew early in the program 19 with the conductivity probes was that if you saw the voltage 20 down approximately.05, you knew you had steam; some 21 voltage above that and approaching up to half a volt, you 22 were pretty sure you had water.
But that is about all you 23 really knew.
24 (Slide.)
C) 25 As we started looking at this, making the
I 5071 06 12 81 1
AGBeb 1
transition from what we did in OTIS to go to MIST, if you 2
used the conductivity probes in OTIS, there's a very large 3
amount of manual work that you've got to do, and we wanted
()
4 to try to get to the point of automating that.
5 We went back to the same test I just showed you 6
and looked at the voltage output and we tried to correlate 7
it with the temperature when we knew the probe was in water 8
only.
And what you find is roughly a linear relationship 9
of that voltage as a function of temperature.
And that is 10 not really too surprising if you know what is happening with 11 the water chemistry.
12 (Slide.)
13 We used basically ammonium hydroxide for the pH 14 control on the primary side of the system and also on the 7,
15 secondary side of the system.
On the primary we typically
~
16 run just about with a concentration of four parts per 17 million of sodium hydroxide.
18 The temperature range of interest, which is from 19 roughly 300 degrees Fahrenheit up to about 600 degrees 20 Fahrenheit, the effect of that temperature or the 21 conductivity with this set concentration of ammonium 22 hydroxide is just about linear, and that is what we're 23 seeing with the conductivity probe.
24 MR. WARD:
So that would be an equilibrium O
\\/
25 correction you could make?
I mean if we go back to that
5071 06 13 82 1
AGBeb 1
previous chart--
2 MR. CARTER:
I am going to do what I think you 3
are talking about on the next slide.
()
4 (Slide.)
5 So if the take that conductivity probe output in 6
volts versus time again, and we look at what the probe 7
should be saying to account for the variation in 8
temperature, what you find is at the beginning of the test 9
you would-- I'm sorry, here is what the original 10 conductivity probe was doing.
11 If you look at the effects of the temperature and 12 sort of negate those out, what you see is that really this 13 is nothing more than the effect of temperature variation by 14 the dotted lines.
That is what the probe should have been O
15 doing if it stayed in water all the time.
16 What the conductivity probe really did is what is 17 in the solid line.
And when you see them begin to diverge 18 from each other, particularly in this region, is where you 19 now have a situation of void content in the fluid.
So very 20 quickly you can start using the conductivity to do more than 21 that.
22 (Slide.)
23 We went one step further in some of the analyses 24 that we did and tried to come up with a way of interpreting 25 the void fraction fairly simply, and all we really did so
I i
l 5071 06 14 83 1
AGBeb 1
far was to use a ratio of the voltage that would have been 2
indicated in water versus the voltage if it would have been 3
in steam compared to the voltage that we had indicated by
()
4 the conductivity.
5 We then compared that with some of our local 4
6 delta-Ps that we had available from the OTIS data and see at 7
least -- I've only got one plot, but we have done it for a 8
number of different cases -- some reasonable agreement 9
between the local void fractions we get the differential, 10 and also with the conductivity probe.
11 And this is the technique I think we are going to 12 be using quite a bit more to help supplement some of the 13 data that we're getting from the narrow range delta-Ps, and 14 then also from the densitometers.
O 15 We have on MIST about 13 conductivity probes in 16 each of the hot legs and also on the secondary side of the 17 steam generators.
I think it is going to be a really useful 18 tool for us.
19 MR. WARD:
Maybe I am not quite getting it, but 20 I'm still--
The change of conductivity with temperature is 21 due to the change in the water temperature and chemistry?
22 MR. CARTER:
The water chemistry is constant.
23 What you are seeing is the electrical conductivity of the 24 fluid changing as a function of temperature.
O 25 MR. WARD:
And that is an instantaneous response?
5071 06 15 84 1
AGBeb 1
MR. CARTER:
That's right.
2 MR. WARD:
And not--
3 MR. CARTER:
Right.
It is not a chemical
()
4 change.
It is just the fact that the conductivity, the 5
electrical conductivity is a function of temperature.
6 The beauty about the thing is that really you can 7
get the calibration data from each test.
By ignoring that 8
you are in a single-phase condition you can develop this 9
curve for every test if you have to.
10 We have gone through a period of about two months 11 on the OTIS test program and shown that the conductivity 12 probe from the beginning of the two months to the end of the 13 two months, all the data still tends to fall in the same 14 one-year line.
But it appears to be a pretty promising O
15 technique, I think.
16 MR. WARD:
So the correction would be a function 17 of the water chemistry that you can--
18 MR. CARTER:
Right.
The water chemistry we do 19 control very tightly for two reasons.
One is just purely 20 for operation of the loop.
21 The other is we want the conductivity of the loop 22 to be at a certain level for the conductivity probes to work 23 anyway, so we do check very carefully the water chemistry, 24 and even to the point of making sure that the HPI water 25 chemistry that we have is similar to what is in the loop so l
l l
- - - = -.
5071 06 16 85 1
AGBeb 1
that we don't dilute the chemistry.
2 DR. SCHROCK:
You would expect the calibration of 3
your conductivity probe as a void indicator to depend on
()
4 flow regime.
Are you attempting to distinguish between flow 5
regimes or getting a single calibration that you view as 6
valid for all flow regimes?
7 MR. CARTER:
We did a little work to look at 8
exactly that question.
Let me back up just a little bit.
9 (Slide.)
10 This is probably not the best sketch to look at 11 it, but we did take some of the same OTIS data and try to do 12 cross-correlations and some other type of correlations to 13 see if we can discriminate bubble movement along from one 14 probe to the next.
And it was fairly inconsistent, looking 15 at that.
16 I think part of the reason is that the probe 17 itself--
This center pin is approximately an inch long.
18 The body also, which is not clearly shown in the sketch, 19 actually extends a little bit into there to also make sure 20 he is in the fluid.
21 I think what we've really got is more of an 22' integrater here.
It does not seem to be extremely 23 sensitive to individual bubbles but more on an average which 24 is happening -- what is happening in this region.
So
(']
25 specifically the flow regime, I'm not sure we would be able
5071 06 17 86 1
AGBeb 1
to pick it up.
2 DR. SCHROCK:
Sure.
I think that's right.
But 3
you would expect that flow in profile would be different for
()
4 bubble flowing or transitional flow, and so the difference 5
in the radial distribution of the void would give you a 6
different apparent measurement.
You are measuring the void 7
near the wall in some kind of integral sense.
8 MR. CARTER:
Roughly from the centerline back to 9
the wall.
10 DR. CATTON:
The tip is at the centerline?
11 MR. CARTER:
Yes, the tip of the probe will 12 actually be at the centerline of the pipe.
It runs back 13 about an inch.
The body then will run another approximately 14 another quarter of an inch or so to get into the wall of the O
15 pipe.
16 DR. SCHROCK:
Have you compared results where you 17 extend in only, say, a quarter of the radius, of the 18 complete radius?
I think you will find that the sensitivity 19 is to the near field, I mean like a diameter of the probe, 20 two diameters of the electrode into the channel, but not 21 much more than that.
22 MR. CARTER:
We haven't done what you are 23 saying.
We will have one nice situation in that through the 24 locations where there are conductivity probes there are also 25 viewports right there at the same location, 90 degrees
5071 07 01 87 1
AGBeb 1
around from it.
2 In addition to that there is a very narrow range, 3
on the order of about ten inches span, at the same location
()
4 also.
So at least from the test cases that we will have in 5
MIST we will be able to, at least for those tests, see what 6
the flow regime was visually compared to what the 7
conductivity probe says compared to what we got from the 8
narrow range delta-P.
9 DR. SCHROCK:
Even if my last statement were 10 incorrect, a different radial distribution of void should 11 give a different response for that instrument with the same 12 cross-section average void.
13 MR. CARTER:
I think you are correct to some 14 extent.
O 15 DR. SCHROCK:
Your potential lines are associated 16 in a different way the void is distributed different.
17 MR. CARTER:
Along that tip back to the body is 18 going to be changed.
You're exactly right.
I just don't 19 know exactly how sensitive it is to the length of that pin, 20 and the distribution on that pin.
I really can't answer 21 that.
22 (Slide.)
23 The last item I was going to go over has to do 24 with the guard heaters.
It relates to a problem that we had 25 during the OTIS program.
I mentioned this yesterday on the
5071 07 02 88 1
AGBeb 1
tour.
2 We had added some additional guard heaters to the 3
reactor vessel pressurizer surge line.
During the early
()
4 part of the OTIS program, a number of those guard heaters 5
failed and we had to replace them.
6 For MIST we will be using about 500 guard heaters 7
and we did not want to have a recurrance of that kind of 8
problem, so we did review the type of guard heaters we are 9
going to use.
We have_ selected a different one.
10 And we have had a benchscale test in progress 11 since just about January, and it is mainly to show that for 12 the most difficult locations, which we found is on smaller 13 diameter pipes near valving, near penetrations where you 14 have to wind that guard heater rather tightly, that is where 15 we have the failures.
16 We have it set up so that we cycle between 300 to 17 600 to 300 degrees in roughly a 12-hour cycle.
We have been 18 doing this since January with about 300 cycles with about 19 seven guard heaters, and no failures to date.
So we are 20 feeling reasonably confident that the problem we had on OTIS 21 is not going to reoccur on MIST.
22 That is all I have right now.
Any questions?
23 (No response.)
24 MR. WARD:
Thank you, Randy.
sl 25 Tom Larson.
L.
5071 07 03 89 1
AGBeb 1
MR. LARSON:
I just have a few comments in a 2
little more detail than Randy presented.
I will make this 3
as short as possible.
()
4 The first topic is densitometers.
5 (Slide.)
6 As you are aware, NRC is providing densitometers 7
for the MIST facility.
You saw the spools yesterday when 8
you toured the facility, eight in the cold legs and two in 9
.the hot legs.
10 There was a training session held at INEL on 11 March 20th.
Topics that were covered there were things like 12 the source handling, calibration.
People went through some 13 benchscale tests and actually looked at operation of the 14 electronics setup, the theory of operation, and got their 15 hands on the instruments so that they will be a little bit 16 more familiar with the items in there.
17 (Slide.)
18 Hardware status.
There were two hot leg spools 19 delivered in January.
At that point in time we had some 20 problems with the braze on the densitometers.
We had to 21 backtrack and develop some new procedures for welding the 22 thimbles into the spool pieces.
23 That delayed the delivery on four of the cold 24 legs until the middle of April.
And then we had to order O
25 some new drilling thimbles for fabrication of the other four c______
5071 07 04 90 1
AGBeb 1
cold legs.
They were delivered in the middle of May.
2 The electronics, as Randy already mentioned, will 3
be delivered near the end of July.
They are going to be
(~)
(_/
4 installed in August.
5 (Slide.)
6 This slide indicates that Randy has beat the 7
schedule by a week or so.
The original plan was to get the 8
spools in by June 15th.
As you noticed last night, they are 9
already in.
10 The electronics.
August 15th is the target date 11 for installation.
12 The final installation of sources, detectors, 13 cooling jackets, et cetera, will be in early March, 1986, 14 and as Randy indicated, that is because of concerns about 15 initial calibration of the facility for heat loss, and also 16 during the initial shakedown tests, it is hanay not to have 17 the sources installed so that people can work around the 18 facility without that additional complication.
19 (Slide.)
20 In early March there will also be a final 21 training session for the densitometer which will be a review 22 of some topics that were already covered at INEL.
They will 23 also include the initial run-through of the procedure for 24 calibration, the all-vapor and all-liquid calibration O)
\\
25 points, and how to acquire data to obtain the calibration
.~__.
t L
5071 07 05 91 1
AGBeb 1
constants for the reduction equations that get you from 2
voltage or counts per second to density, s
3 Also at that time we will be discussing t
()
4 uncertainty calculations in more detail than we have to 5
date with the people here at MIST.
We are in the process c
6 right now of developing the software in an automated fashion 7
to do both chordal uncertainties and cross-sectional average 8
uncertainties for each of the densitometers.
9 MR. WARD:
Tom, are these densitometers identical j
10 or similar to those that are used in Semiscale?
l l
11 MR. LARSON:
Neither.
These are two-beam r
12 densitometers that use americium as the source.
The i
13 detectors are all sodium iodide, water-cooled.
And because i
!(
14 of the MIST requirements you saw yesterday, these spool 15 pieces, they are all an integral, welded-in, leak-tight I
t j
16 design.
17 The designs we use on Semiscale are either cesium 18 clamp-ons that have a huge source, 20 or 30 curies of I
f 19 cesium, or a three-beam spool that actually is flanged into 20 the pipe so it is not necessarily leak-tight because of the j
21 mechanical seal.
22 And in addition, on Semiscale we use a system
[
t l
23 where on the low-energy systems we have three different 24 sources and one detector, and it is an uranium detector that 25 you can discriminate energies with.
i a
1 I
I k
+
..., _., _ _.. _,., _ -.... _ _ -,. -. - ~.,. _ _. _ _ _ _., _ _ -. _. _..
2 5071 07 06 92 1
AGBeb 1
On the MIST facility we are using one source and 2
taking multiple beams from that source with multiple 3
detectors, so each source cast on the MIST facility has two
/~T 4
detectors in it.
Each spool you saw yesterday, when in V
5 operation, will have one source, about 250 millicuries of 6
americ ium.
This design has been used in similar form on the 7
ROSA f acility, and so parts of this spool have been used 8
elsewhere, but this exact design has not.
9 MR. WARD:
So essentially it is a design that 10 evolved from EGD experience at INEL and, in particular, what 11 was developed for ROSA?
12 MR. LARSON:
Yes, correct, and also UPTP.
13 (Slide.)
14 The next topic of discussion is the cooled TCs 15 that Randy mentioned previously.
These are also known as 16 thermal dosimeters.
This is the schedule for the devices 17 that we are potentially going to supply for the MIST 18 facilitics.
19 As Randy indicated, we are right now in the
'1 20 process of doing separate ef fects tests on one probe.
As l
21 the schedule indicates, what we are doing is building one 22 probe and modifying our test loop out at Idaho in parallel 23 with that construction.
24 We will do the separate effects tests to actually
()
25 determine what the low end of the flow range on this I
L
5071 07 07 93 2
AGBeb 1
instrument is; in other words, the lowest velocity that you 2
can reliably measure.
And at that point in time, after some 3
preliminary data analysis, a decision will be made as to
(}
4 whether or not to pursue purchase of an additional four 5
probes.
6 And if this looks favorable, then four probes 7
will be f abricated and delivered in this timeframe and 8
potentially installed somewhere out here in September, late 9
September probably.
10 (Slide.)
11 A little more details on what separate ef fects 1
12 testing really means in this context.
13 The probes are designed to handle, naturally, 14 MIST-type conditions, 1800 psi.
I 15 We have done analytical calculations.
Both B&W t
16 people here and we in Idaho have looked at the effects of 17 bulk temperature changes on these probes, on the probe 18 output.
19 There is also a concern about the the cooling 20 water temperature, how that may affect the calibration of 21 the probe.
22 So in our separate effects experiments, our 23 calibration experiments, we are going to look at both of 24 those ef fects and take approximately 20 points, 20 data 25 points at that velocity range for each probe.
This is in 4
L.
~.
~
5071 07 08 94 1
AGBeb 1
ef fect calibration conditions.
You have to calibrate each 2-probe.
3 The next slide is simply a schematic of what the
()
4 probe really looks like.
5 (Slide.)
r i
6 I won' t spent any time on it unless there are 7
'v questions, i
l 8
MR. WARD:
Could you give us a brief description 9,
on it, how it works?
[
l l
10 MR. LARSON:
Let me point out the significant t
11 features of the probe.
i 12 This is the business end of the probe right here, 13 and that ideally is the tip of the probe which would be on 14 the centerline of the pipe.
[
i 15 This would be the pipe boundary right here, 16 contoured to fit the inner ID of the pipe.
)
17 The part from here back is simply the mechanism i
j 18 needed to get the cooling water in and the T/C leads out.
1 19 The basis for operation of the probe is simply i
i 20 one of heat transfer.
The probe is cooled by cooling water i
21 that comes in, runs down a tube on the inside of the probe, 22 makes a 180 degree turn at the end of the probe, and then 1
23 comes back out in the annulus.
That's cold water, cooling i
]
24 water.
()
25 What that ef fectively does if you use a high l
1 f
i k
I i
/
l
, -.- ~ ~
i 5071 07 09 95 1
AGBeb 1
enough velocity on this cooling water is to keep the 2-inside surface of the probe at a reasonably constant i
3 temperature.
4 There are thermocouples, one on each side of the
(
i 5
probe.
This is round, of course.
What you are doing, 6
really, is looking at a top view here.
These tiny tubes you l
7 see on the outside actually contain 20-mill thermocouples.
8 Okay?
They are squashed inside of these roughly 20-mill
}
9 tubes, which is silver-soldered to the outer surface of the 10 probe.
i i
11 So what you in effect have is a temperature 12 measurement with the cooling water at the tip on the inside i
13 of the probe, a wall temperature on the outside, and F
14 somewhere upstream or downstream, you have a bulk fluid j (
15 temperature.
i i
i 16 So if you go through the heat transfer analysis i
)
17 of the probe, what you find is that the bulk velocity can be
)
i 18 related to a constant times a ratio of bulk temperature and l.
j 19 wall temperature, and wall temperature and coolant j
20 temperature.
And on a log-log scale, calibration is very 21 fine.
l l
22 DR. CATTON:
So it would to be hot water?
23 MR. LARSON:
Exactly.
24 DR. CATTON:
How big is it, the tip diameter?
()
25 MR. LARSON:
This is a quarter inch tube.
And j
I m
5071 07 10 96 1
AGBeb 1
like I sa id, the T/Cs are embedded in the 20-mill tube.
2 Back hera is a half diameter inch tube.
This 3
right here is simply a miniature T/C connector block.
t
(}
4 All the leads for the wall temperature 5
measurements and the tip temperature measurements come down 6
the inside of the probe and exit to the T/C connectors here.
7 There are two temperature measurements for the 8
coolant inlet and outlet water.
Hence you have five 9
temperatures on the probe, five temperatures available on 10 the probe, anc then a bulk temperature somewhere.
11 Really to get a measurement out of this thing you 12 need three temperatures, one wall and the bulk and some kind 13 of a coolant temperature.
Whether it is a tip or an average 14 of the coolant inlet / outlet remains to be seen which works 15 the best.
16 DR. CATTON:
Can you get flow direction out of 17 that by having T/Cs on two sides of it?
18 MR. LARSON:
By inference, if you have both wall 19 temperature measurements hooked up, you can look at the 20 shape of the curves.
The downstream one will by nature be 21 more noisy.
We have shown this in calibration tests.
22 So-you can infer direction but it is not an automatic 23 thing.
24 You either-have to visually look at it and make a I
25 decision, or I suppose a statistical analysis could be done l
l i
(
_-_.,,c
t 5071 07 11-97 1
AGBeb 1
to audit it.
You have to have both wall temperatures hooked
~
2 up to do it.
3' MR. WARD:
That is why it is legitimate to say
[
i i
( )'
4 that it 'is measuring velocity rather than speed I guess.
f 5
DR. CATTON:
It is measuring speed.
[
6 MR. WARD:
Well, you're getting an indication of 7
direction.
8 MR. LARSON:
If you have both hooked up you have j
9 a record of quantity and velocity.
If I have one hooked up j
10 I guess I've got speed.
i i
11 MR. WARD:
Where was this--
Is this used in l
12 Semiscale or LOFT?
i i
' 13 MR. LARSON:
We are using two of them in i
14 Semiscale.right now, yes.
There is not a lot of operating l(
^
15 experience with them,. if that was your question.
16 MR. WARD:
Yes, that's what I was just-wondering 17 about.
18 MR. LARSON:
It is not enough to show--
They are 19 hand-built.
20 We did build one some years ago with the thought j
21 that it had potential application in TMI, so we built one 22 that would actually fit in the ther=al well, but it was j
23 never used, never installed.
In fact, that is where the i
idea for the probe came up, is with that application.
24
.( )
25 That concludes the additional details that I had i
6 t
F
~_.,4
,,:,.-,.,,e
__,w,w,m,_-.,.r_m..w
.,_,,.wc,
,.7.,y_
_y, m,
i 5071 08 01 98 1
AGBeb 1
to offer.
2 MR. WARD:
Our next topic is on the long-range 3
plans and who is going to cover this?
4 MR. BECKNER:
Jones is going to start, and I will
{}
5 follow up.
6 (Slide.)
7 MR. JONES:
What I am going to discuss is the 8
follow-on testing plan or program which NRR requested on 9
full power tests for MIST.
10 (Slide.)
11 Basically back in October of '84 we issued a User 12 Need letter to Research requesting the follow-on program.
13 In that what we did was we stepped back and looked at the 14 IST program to decide whether or not it appeared that there
()
15 was adequate data being gathered for complete code 16 assessment for the B&W plants.
These include non-LOCA 17 transients, what are the risk dominant sequences, and small 18 break LOCAs.
19 Out of that review we concluded that the current 20 IST program was adequate for small break LOCAs, which was 21 its basic objective when it was constructed.
22 There was -always, when the original tag was
~
l 23 developed, a concept of possibly using this facility for a 24 follow-on program, hence we put in the capability for 100
()
25 percent power, for example, on the rods.
u m-#
4 e
-s-
w'--
g-L---tetw e
t-
5071 08 02 99 1
AGBeb 1
On a cost-benefit analysis performed back in '82 2
when we were looking specifically at small break LOCAs for 3
this cooperative program, it did not appear beneficially to
(}
4 totally upgrade the facility to a full power situation.
5 So what we have done is we have basically j
6 requested in this User Need letter the typical NRR position i
7 which is show us that you have adequate data for code 8
assessment for B&W plants.
Show us your codes are good 9
enough, or get us the data which we believe was best 10 obtained in a follow-on program in MIST.
11 And that's essentially what it came to, is that f
12 we are recommending more or less this follow-up program.
13 That is our view at this point.
f 14 Based on what we would like to do -- The next 15 slide will show the type of transients we would like to see 16 the facility be capable of doing -- we came to the 17 conclusion that a full power simulation appeared to be 18 necessary.
We asked Research to look at that and determine 4
19 whether or not that was required, f
20 We didn't sit down and do a detailed review of 21 various ways of operating this facility to try to simulate 22 full power ef fects of the various phenomena.
We just have 23 such an extensive list, as you will see in a minute, that it 24 appears to us that full power simulation was the best way to
, ()
25 go, but we weren't prohibiting further looking.
4 I
..n v,,--
n.,-..
., ~
,---,-,--+-n----
i 5071-08 03 100 1
AGBeb 1
We asked Research to look at that item and tell 2
us whether or not it was required.
3 DR. CATTON:
With the reducing budgets and that 4
cost of eight million dollars, it seems to me the scaling
(}
5 study had -better be done first to show that you can't do it i
6 before.you do it.
7 7
MR. JONES:
Agreed.
And NRR is being nice guys 8
and pushing it over to RES.
They are the researchers.
l l
9 DR. CATTON:
They are the ones that are supposed 10 to do it.
]
11 MR. JONES:
That's right.
That is our view, any 12
- way, f
13 As you see, the estimated cost was eight million j
14 dollars.
The User Need letter said approximately seven
(
15 million.
About half of that is for the upgrade to' full 16 power.
The other half is operating costs for getting the 17 data, the reports, the analysis, et cetera.
t 18 The eight million dollars is a rough estimate
'19 that we obtained from B&W about -- some time in the last six 20 months.
And based on j ust the currer.t 'IST program and some l
21 time to put in an upgrade to the facility, it is scheduled f
22 to be roughly September '87 to October '88 before testing.
R23 Actual hardware stuff would start or precede the September 1-24
'87 date.
25 (Slide.)
L 1.
l
~
5071 08 04 101 1
AGBeb 1
What we've constructed in the User Need letter as 2
shown here is kind of a wish list.
These are what we viewed 3
to be the various transients we would like to see the
(}
4 follow-on program be capable of running.
5 We did not attempt to prioritize this list, nor 6
work with Research at this point to come up with a test 7
matrix.
We just said we think about 25 tests would be 8
adequate, and here is the list of tests or areas we think 9
the Code ought to be capable of handling, and that Research 10 and us would work together to define the specific test 11 matrix if it was deemed necessary to go on with this 12 follow-on program, and what we would look at.
There were 13 the issues, more or less.
14 As you can see, we have steam line breaks.
The f) s/
15 current test series as one case, the so-called AEOD case, 16 which is a steam line break with a tube rupture, with a 17 combined primary / secondary LOCA, but it is limited in its l
18 size due to the current hardware configuration of MIST.
19 We have reviewed that with AEOD.
We recently 20 wrote a letter to the Commissioners which said that we feel 21 that the current test in MIST is adequate for resolving the 22 AEOD concern on the combined primary and secondary LOCA.
23 And we have AEOD's review and concurrence on that letter.
24 This is the combined primary / secondary LOCA that
/~T
',_)
25 was talked about in that SECI -- I can't remember the number s
d 5071 08 05 102 1
AGBeb 1
now -- 82.296.
It addresses that one.
2 Now I recognize you have another AEOD issue which 3
I will point out where I think this testing program should
(}
4 be able to handle it.
We've got basically steam line 5
breaks, feed line breaks, loss of feedwater scenarios, steam 6
generator overfill.
7 Now again the steam generator overfill is the 8
series that we could look at, this other AEOD case, which is 9
filling up the steam generator, and possibly use that even 10 as a condition within one of the steam line break series, 11 and that would resolve, as I understand, the other AEOD 12 issue which is filling up the steam generator and having a 13 steam line break as a result of that filling.
14 We also have in there, based on some of the 15 Subcommittee's recommendations, looking at a couple of 16 alternate ECCS designs such as the High Point sprays in the I
17 hot leg or the use of the High Point vents for feed and 18 bleed.
19 In steam generator tube ruptures, we may wish to l
20 extend what is in the current MIST program to include some 21 failure scenarios such as a stuck-open relief valve, but we 22 haven' t given a lot of thought to that series at this 23 point.
It just appears that maybe we can do some additional l
24 failure scenarios that could be incorporated within the test 25 matrix.
l l
l l
l
- ~ - _ _ -.,,
w =/ l n *~ cu g
. V.L n
=
~ e-w v.
s
.t
.7a'a.h
.T
.(7..
_.. s _.'.c'*d t{.
2%g
?. 2.,2
.s
- 3. a. zy-T. 4.. s.
- b..e a..m g ~.3...
w * *W o.
O.
ya.. 4 3p. e +.".)
,- s
.s.
w:
.T.*,fii_t.
p e.e o.
a$
.e,y,.e s::r,
y._. 3 *..
p.
.es..s ? M..y g.3 b 3*
- j.
- . '.-' I
- 3 *p3..* * "y e
]
f
- e..b. e.
.C *;* $
- g ** 3 a s.a i. A.
w.E3*,
E. 3 - [. E. *1*,
^8 4*
.y g
s gf.'%
.M *
- p. *.* T M. y.
- p.
8.* I 1
- d. w..e.b T p jk m. J ped, -
T C"
w7 a. '. - Cpa. g.
U d..
w
- r*
t j.
. ~.
.'i. tsrc.
3,*y.2 i..
3.::.e.,.w.i.y,-
7 w
- 3+
ys.
_3..
,,a....M m e m.*I n gg.ns.m*4 ys.
3 Z..~ 1.T y gs..e s.
m.
y.
y g.
- 2
- h 6. S i.g w e.g ea s w.2
.e3.4*
S e
- *Je *..* 3 A T7 7. e 2*
= =.
w w.
.u
- s.
.w;5
- T,e****e Si. **#. #.'..'.
Tve".
's
V ' ' '.
- i. g y,a.
- ^ "
y
.a T*.*.
y.
'l vid.*. s pe. *_ um e..'
..-d.' *..' o *. s *. * *_'* i. _= _' _=.- i. .' *.r. s..- * *- " = *. v=.
4
- 3. 3.%.;.;.4. y
._A a.. g g) y2 g~.
T.a.
"' 5 *. '. *..
t?
- - --3.. E,.a
+%
h 3 + e
.3 'a ?
.~.2,..
w y.
4.
13
'As.
3,153 k v a.
3.a r,..t s.
~.*
? _3...
- m. o.cm'! d vf
-e
'l 4 S ir.*.* * *..' ^.^.c.
1.#S *.e.d
- ". e.* e.
- 20.5e
- .**.1.7.. '. ',", e m.. '.'. # "V.*
~
s.
Tlevew, 15 pC5510.7 ooc.d cs e scr:t e 13 S*enarlOS YOlCO V 1 c.P o r ma. em 4 wa. e. *. a..e.1 Am.2
- e w3.fe
.s~e 3.
2 M
-J..
%sP Abd"ab h. M e
-my I7
- d A 21.C,*
I" 3 *m* a.
.E.*. a. l *J 4.e.
2
.a.g,.q 3 d
4u 5**1 4
- 2....-. *. =.
..C.,*.
W*
m e
S.n.~.I1 *. # ^*
'e. **#*t I$
DAg*..
?. E.
g g f,* e *
- M..' 'v'- 3
"*d.* **'e.". *. Q V ^
- i as y.
. +. a
.~.-
19 it v ld be very varthwhile to do Ibe sn11 t eat L'>^A a.d
- 3. 3.t 3,.
..u3,e e.e.s.
.e.+, m,.,. e. a r.,.., s
.'e
=s*=' m e'
.~
r----
+
' 'L 5 *.' ' *
- T
.#.'Om
?. %. M. '.~ e*.a *'
'.P.'.
.s y
7
- Q ]. Q *.=* 3. *..-1. g. E. n g, *. a. e. a
.1..
]7 13
.p..* gy
.s w=. g se.
w.p w
$$@. N. $.*. P. G.3'"* 3.*..* i g C I 3**
- SYN 3 2*E**
2 **. 4 WYC*
M*
. -.. y e..s.
b.
f.
r----**
- - * - - *'- * - --' --~ - - - - * *
.._.ee { -. A..a.
- A3+
- 8, K
N 7.4
- 3c+
o #.
. b e..g. +... h. 3 + y?a.
.e2e+s e=
g,. _..
s
.=.
..s r_
- y
-% g e.
psg.*.Cw.
a=%.". 3
,,.4. ? ?.
- 3.
.e. m.y.3,4.,. __ 4 &.
- t ;
- . e. y *gm 1
4.
/
7 i
1
5071 08 07 104 1
AGBeb 1
you wish, close the loop to ensure that the test data is 2
indeed accurate.
It is kind of confirmatory, and we think 3
it would be very worthwhile to get such a test if we get the
("')
4 100 percent power simulation.
v 5
DR. CATTON:
Which of that list actually really 6
requires the full power?
All of them?
It doesn't look like 7
it.
8 MR. JONES:
Not the plant cooldown simulations, 9
certainly.
10 The tube rupture scenarios could be useful to 11 simulate a single tube and the plant runback, for example, 12 and that you want to be at full power.
13 The steam generator overfill I would think would 14 be very useful under the full power simulation.
(~)
k/
15 DR. CATTON:
I am hard pressed to see 16 phenomenological differences as a result of full power in 17 running down this list.
18 MR. JONES:
I think the big actor is just getting 19 the right stored energy in the loop, and its possibly impact 20 later on, and the reactor coolant pumps running.
If you run 21 this now with the reactor coolant pumps running, you are 22 going to get a very low delta-P point across the core.
23 Now part of that is the limitation of the control 24 room of the pump.
They can only control it down to 20 O\\_)
25 percent speed.
It may be possible to simulate it by l
l l
5071 08 08 105 1
AGBeb 1
holding the pump speed down.
But my understanding is the 2
best the manufacturer can guarantee is 20 percent speed 3
control.
Below that they cannot guarantee that the 4
control room will work very well.
5 And then finally, there are a couple of 6
risk-dominating sequences:
a small break LOCA without HPI, 7
and we can look at various scenarios such as aggressively 8
blowing down the steam generator, station blackout, and 9-ATWS.
I'm not sure how you are going to do ATWS on this 10 thing.
Maybe with a power program similar to what Semiscale 11 is might be a way, and it is probably very challenging for 12 the facility's capabilities as it exists, and it may not be 13 possible anyway.
14 But as I said, this was basically a wish list 15 which said these are all the type of things we think the 16 codes ought to be capable of handling.
We are basically 17 asking Research to confirm either the codes can handle these 18 cases adequately, that there is an adequate data base for 19 code assessment.
If not, then certainly those should be the 20 candidates for a follow-on testing program, i
21 That is the whole purpose of this list.
We 22 picked a bunch of people's brains and we got a big list-to 23-start working from.
24 MR. MICHELSON:
On the estimated schedule you
()
25 indicated September '87 to October
'88.
Is that the time I
i l
l
.=
5071 08H09 106 1
AGBeb 1
period in which all of these tests would be performed, or--
l 2
MR. JONES:
Basically, yes.
You would start--
3 I'm not sure of the exact time of hardware implementation f
(J'T 4
that you can put i t o n--
Well, certainly you can't put it 5
in until MIST is done, and it is probably not recommended to 6
touch the MIST facility until such time as we are confident 7
that all the data we have is good data, and that there are 8
no required reruns.
9 I believe MIST--
We have a whole period on the 10 facility for three months, I believe is the current status, 11 which puts you into roughly March of '87, maybe early '87, 12 somewhere in that timeframe.
13 MR. MICHELSON:
Do hardware changes take long to 14 make to do these tests?
15 MR. JONES:
This would be an initial facility 16 upgrade, such as changing the secondary side condensers or i
17 blowing it to atmosphere, or a bigger tank, the power supply 18 changeover, et cetera.
19 MR. MICHELSON:
It might take more than six 20 months then.
21 MR. JONES:
We are hoping and planning to go 22 earlier than that.
That was our assessment, that we should 23 be able to do some work up front from the design and 24 procurement standpoint so that what you are left with is
()
25 primarily installing hardware, and you might even be able l
l
~
l i
r-5071 08 10 107 1 'AGBeb 1
to put together some of the hardware on the side and be 2
ready to just hook it up to the loop.
And six months to 3
nine months appears reasonable from our perspective to do
(}
4 that.
5 MR. MICHELSON:
Does the '87 budget have this in 6
it now?
7 MR. WARD:
No, I don' t think it does.
8 MR. JONES:
There is a line item identifying the 9
research budget for a follow-on program of some sort.
10 MR. WARD:
It was identified--
3 11 MR. SHOTKIN:
Bill is going to have some--
But i
12 there is a line item.
Right now we are looking at several 13 options now, and it would depend on whether we get industry i
14 support for the follow-on which option we take.
r 15 MR. JONES:
And then as far as the timeframe, the d
16 one year is based on the MIST program currently is running 17 30 -- well, 40 tests I guess it is, alternately, and the 40 18 tests plus the reporting, et cetera, is roughly a one-year 19 timeframe, for a year.
i 20 MR. MICHELSON:- Have you looked at some of these l
21~
questions in terms of if it is going to provide a. timely 22 answer?
For instance, some of the AEOD' questions which have i
23 been outstanding for a long time already, and this means a 24
_1988 answering period.
()
25 MR. JONES:
I think right now the way we view it o
{ '
l i
l
,. - - ~
-,m.
,.. _. - -.......,.. ~, _ _ _
, - - +,. -, -
,,--_,..m
5071 08 11 108 1
AGBeb 1
is we don't have any significant or outstanding safety 2
concern which needs us to rush and get this data.
I think 3
our primary problem right now is we view this facility as f')
4 it, when it finishes its program, it's done, it's gone, and J
5 there will be no more data for B&W machines.
6 We think it is appropriate then to try, to the 7
extent possible, to fill in the data base in the various 8
areas, non-LOCA and f ront end and the risk-dominating 9
sequences in order to assure that should future problems 10 arise, we know the codes are capable of handling the 11 situation, or are reasonably confident they can handle it.
12 Otherwise, we are left with a series of small 13 break LOCA data and that's all we have, or primarily all we 14 have from MIST.
15 So it is just to close the loop and kind of bring 16 it up near the status of where we view current Westinghouse 17 and CE issues.
18 19 20 21 22 23 24 25 4
5071 09 01 109 1
AGBwrb 1
MR. WARD:
And BWR issues?
2 MR. JONES:
I'm just not that familiar with BWRs 3
to answer that.
But I believe that is correct; I think that
{~}
4 we have concluded that what we have now is sufficient.
5 That's my understanding.
I 6
DR. SHROCK:
These high point vents you' re 7
talking about, feed and bleed, and so forth, it seems to me 8
these are a new potential for small break.
I wonder, have i
9 you considered whether the. nature of the small break, the 10 location of such vents, would be substantially the same as 11 on hot legs, or is there a difference in the size potential, 12 or what-have?
13 MR. JONES:
Well, right now the high point vents 14 on most of the units are fairly small,-they are very small.
15 I don' t remember the orifice sizes, but they do range 16 from --
17 MR. WARD:
They can' t be too small if you' re 18 thinking about using them for feed and bleed.
19 MR. JONES:
This item is put on the list 20 primarily due to Mr. Ebersole's questions and his 21-recommendations.
My gut reaction is, at least the current l
22 size systems are not adequate.
And, in fact, that was the 23 staff position during the TMI restart hearings, during the l
24 Appeal Board hearings, that they're just too small for this
()
25 function.
And it probably is, at this point, a low priority l
l t
I l
l
~
5071 09 02 110 2
AGBwrb 1
test, in my opinion.
2 I think the high point vents, as was mentioned 3
ye sterday, for assisting recovery from small break LOCAs and 4
filling the loops up, I think that has a lot of promise.
(}
5 Even though they are very small, they seem to be very 6
effective in doing that.
And the system just is awful 7
stubborn, so to speak, when it is time to try to fill it up 8
and recover natural circulation.
It's just because of that 9
trapped vapor that you just have to wait a long time.
10 So that has already been identified, as I 11 mentioned yesterday, as an open item on ATOG, the use of the 12 vents for such items.
And that's where I personally see the 13 primary use of the vents.
I just have a feeling it will not 14 show up to be very useful in feed and bleed.
15 MR. WARD:
I don't see why you need experiments 16 to come down with an-opinion on that.
I mean, I guess it 17 seems fairly obvious to me, without any calculations, that 18 they can' t be useful for feed and bleed.
And certainly with 19 calculations you could show it one way or the other.
20 MR. JONES:
I agree with you.
I really do.
I 21 agree with you.
22 This was a wish list, this was put together last 23 October to just try to identify all the various things we 24 heard, the comments from the Committee, talking to the risk
()
25 people as to what they would like to see.
And I'll admit,
5071 09 03 111 2
AGBwrb 1
those aren't well-defined as risk-dominating sequences 2
except small break LOCA without HPI.
I'm not even sure what 3
they're looking for from the station blackout situation and
{}
4 how we would construct the tests that would give them the 5
information that they might want.
6 The various other stuff, the ATOG stuff was based 7
on a little bit of a review with the ATOG, and where it 8
appears there's data, where it appears there isn't data or r
9 calculational backup or the calculations could really use 10 some backup.
The others are just basically Chapter 15 type 11 transients where there are no data at this point.
12 MR. WARD:
Okay.
Thank you.
13 We'll take a short break now.
14 (Recess.)
15 MR. WARD:
Okay, can we proceed?
16 (Slide.)
17 MR. BECKNER:
I'm supposed to talk about long l
18 range plans as to what we do with the facilities from here, 19 and also address some questions that I think Ivan was 20 alluding to.
21 I had a little trouble putting this presentation 22 together, really for two reasons:
First of all, we are in 23 the early planning stages of this effort, and the second 24 reason is that this is very closely tied to a larger overall r
()
25 ef fort that we are undergoing, looking at what is the future 5
i l
f l
5071 09 04 112 2
AGBwrb 1
of all our test facilities; and, again, that is also in a 2
preliminary stage.
3 I think if you will remember from our budget
(}
4 discussions, there are some line items in there which are, 5
again, varying with time.
But there are line items for the 6
MIST follow-on, and also line items for some kind of 7
continued integral testing in general.
8 So, really, what I'm going to present today is a 9
snapshot in time of where we are.
10 (Slide.)
11 First of all, we are interested in continuing a 12 full power program.
And I have it worded as calling it a 13 f ull power program right now.
And I think I would also 14 concur with some of the comments that Ivan made as to "Do 15 you really need full power?"
I'm not sure.
16 There is a need for some follow-on testing that 17 Jones has identified, and the question of "Do you need full 18 power to address them?" I think is an open item.
And I 19 would agree with Ivan on that.
I think if we could get his 20 commitment not to ask us any questions in the future on full 21 power, then we could maybe do without it.
But basically I 22 agree with him that that is something that we should look at 23 as to can we do these tests without full power or not??"
24 MR. WARD:
One of the items on Bob's list was f'T Q
25 ATWS.
i
l 5071 09 05 113 2
AGBwrb 1
MR. BECKNER:
You need more than full power for 2
that.
3 MR. WARD:
Let's see; remind me what sort of ATWS 4
test was done on Semiscale.
5 MR. BECKNER:
I'm not sure if I can.
4 6
Have we run ATWS on Semiscale?
i 7
MR. SHOTKIN:
It was done on LOFT.
You need the 8
feedback to really do an ATWS te st.
9 There was one done on LOFT, but it was a very 10 minor test.
You can't pressurize very high.
11 MR. WARD:
So if there were to be an ATWS test in j
12 MIST that would be really going beyond what was done for the 13 other reactor types?
14 MR. BECKNER:
We did not--
15 MR. SHOTKIN:
There was probably something called 16 an ATWS test in Semiscale, but I wouldn't really call it an 17 ATWS test.
18 MR. BECKNER:
Remember, we did that for FIST.
19 But the way we approached that was, first of all, the first t
l 20 couple of second.were pre-programmed power based on a t
21 calculation, and from there on we went to a series of l
22 separate effects tests, looking at phenomena we thought were l
23 important during ATWS.
i 24 MR. JONES:
As I remember, there was a test.
It
()
25 was done-- I think it was part of the power loss test series l
l l
5071 09 06 114 2
AGBwrb 1
in Semiscale, where they used a calculational module to 2
drive core power to simulate ATWS feedback.
I don' t 3
remember the details of the test at this time; except that
/'N 4
there was that special power controller, and they tried to V
5 pick up the nuclear feedback in the test.
6 MR. WARD:
Say it again for the record.
7 MR. SHOTKIN:
It was just the small 8
pressurization.
9 DR. CATTON:
What are the big questions, 10 thermohydraulic questions, associated with ATWS?
I don't 11 think there are any, are there, for BWRs?
12 MR. BECKNER:
I think for BWRs that your 13 pressurization rate is to some degree a nutronics 14 question." And high power natural circulation is a 15 phenomenon.
s 16 DR. CATTON:
But if you can do low power natural 17 circulation.*I don't know why you can't do high power natural 18 circulation.
19 MR. BECKNER:
It's certainly a subset.
20 DR. CATTON:
I mean, it's a feedback question, 21 but you can' t answer that with an electrically powered 22 system.
I don't really see the need for it.
23 MR. BECKNER:
I think, as Bob Jones said, that 24 was a wish list, and I think our response is that we feel
,m
(_)
25 some testing is necessary.
I think the reason for that is
,. =
i 5071 09 07 115 2
AGBwrb 1
that what NRR has told us is that these are the types of 2
things they feel need to be calculated, and they are putting 3
it on our shoulders now.
And I think what we are saying is
{ {}
4 that we need to do some of these types of testing to 5
quantify the uncertainty, basically, where we would be 6
accessing the codes with these transients, and we do not 7
need that uncertainty.
8 Certainly, we haven' t looked at this in great 9
detail yet, getting into the types of things you're talking 10 about.
There are some things where I think we do need 11
. testing to quantify the uncertainty.
12 So what do we do with this?
i 13 Well, first of all, we are considering a possible 14 follow-on to the existing program, which would be an upgrade I ()
15 here.
16 I've also got a third bullet up there that I 17 think has to be mentioned: this is not directly related to 18 the MIST program.
But if you' will recall, in looking over 19 our long-term plans, one of the options that we have been 1
t 20 looking at would be to consolidate all of the facilities at j
21 Idaho, or something like this, to minimize -- in effect to l
22 consolidate staff in that manner.
You maintain one expert i
i 23 testing staf f as opposed to three or four, depending on how i
24 many integral facilities you have.
()
25 Now, again, that is only an option: we don' t know J
I l
r o
I
.,,.,,,,,-.--,,-.y,,,
,-.----__y.--
,~~~,m,mc-,~....--y+--e v
-w..,e_-,---.-,-
5071 09 08.
116 1
AGBwrb 1
if it is practical yet.
But it should be pointed out that i
2 that's one thing that is being considered in a separate 3
study.
4 Let me say a little bit more about that.
5 We do have a commitment to get a lot of 6
information together, I believe by the end of this fiscal 7
year, on cost and practicality of a number of different I
8 options for continuing testing: this is not just in MIST, 9-this is in all our facilities; such things as the cost of i
10 different options, the practicality of going to smaller 11 low-pressure facilities, as Ivan was discussing.
Idaho is i-
. orking on that for us now, and is supposed to be giving us 12 w
13 some input by the end of this fiscal year.
14 The fourth bullet up there I guess is more a 15
" motherhood" statement.
We're going to take a look at the 16 dif ferent options and hopefully come to something that is 17 technically acceptable and cost effective.
18 So there are a number of options out there, i
19 Likewise, on the last bullet, I think that this l
20 will be true for any option we go with, as far as we need 21 some type of industry support.
And let me explain what I 22 mean by that.
i
.3 Tnere's two types of industry support: one is 2
l-24 providing money, which obviously helps; the second one as
()
25 far as industry support is that when we go before not only 4
b i
i s
5071 09 09 117 l
3 AGBwrb 1
our own management but when we go before different 2
committees and Congress, it always helps to have some 3
support in terms of general agreement with the technical
) (}
4 community that these types of tests are needed.
5 We are contemplating a tag-type ef fort where we 6
can get a consensus between the technical community, 7
including industry, the government, and other individuals 8
such as university people and consultants, that this is the 9
type of thing that needs to be done.
10 so we will be seeking industry support, and we 11 will be going forward and trying to talk to some people this 12 summer on this thing.
13 So, again, that's j ust a snapshot.
It's rather 14 vague.
But I'm going to tell you about some plans and
(
15 on-going efforts.
We haven' t come up with the answers yet, 16 but we are pursuing those answers.
17 MR. SHOTKIN: -Bill, can I add something to that?
18 As far as getting industry support, I think there 19 are many people who really wouldn' t believe industry support 20 unless there was some financial, or in-kind contribution.
21 So the type of support we will be seeking is some 4
9 22 co st-s haring.
23 MR. BECKNER:
Okay.
Let me just touch briefly 24 again on this alternate testing strategy, or what-ever.
()
i 25 (Slide.)
. ~.. -
.3n,13 3
U. S i. 4 As%
1 AG3vrb 1
This is the type of thin; ve' re d::in; i 2
- de've been ic6r.ina at inis as j us t one s
1 a ?.. a...a *..'. e *
- n a i..".. 2. i n,. =.
'..+.+.,..=?.
- e s *. - =._= s-i ' '.
- v,
=*. -
3
. mee
~ s.
2.n.4. w... s. o.n.,. e. ae. te,..a,_s 2.x,--.s
- s_
4 s
A 5
lo ting at right nov in ter2s of everall effort, and they 6
vill be nating us of studies ender this pro;ran to previde 1
4-.,
~.-
. e7e.,,i
< >. ; i. ;...
7
- cer.
m x;.g.we.3 -
3.,.33 4 s '-- 3
.aa..
." n.7-. <.e b:v
.a..
r.
- ne timin; via., v:rt Out.
=2: it is a case 2,.
vtere we v.,,
y
. ha *.*.; n, *
- s*;*.i
~='
.#.*.- 5. 7.* *.v., ~~ ' ~.~. r-es s.. a...' -n e c
,a m.,
x'l
- ali i
- e.'.'*.-.,~~.~.~.ey-*.*.*.
- a. s *. s #. on *.
.n, v. ~,.'.
e-g 12 alternate scalis; concept.
- 1. 3
- o *..'.i s i. s = a ;, e e,-,c.-*. -. ' *.v.,
.r.. p.. '..
W*
4
,.a..,,-
sz w..
-.e.as c' c..,e.e..=.'
- m. e..=..'.' * > s ~. =.-r, s.-
s-.
i ca r., - 4 3. a...,., e a 1., a. e % s... e e..
.i -3 s a, 5.,..e.u 3.
a <<
.3 5:
.43 T 4.~A.a., 3. g.4
. w.,
. h. 4.a 5,7. *. 9.,
33 wo.
g. 4.-.
e.
-f 17 has started b:th in terns which Jean-Pierre and Ton 1,ars:n
-.e 3,;,,..,-.g..usg - o.. ma,
a.
a
.a,-
..ge.--.,-
.g 2 ; a, 19 arallel.
3 t emicusly they are loatin* at vaat ea:t cther e
r 1
2G is doing to icct at this scalin; eencept in general.
v.
x,
. ;s e<<-. umas s.a..s..
d.
4.
.<.2,.m..;....
22 vten we start getting data and coap.ario; results fran
,3
, a_ 4 3 4.. ; e s.
24 T:st is a very quiet everview of weere ve are p
i
\\,
25 going so f ar as the future.
And I <;;ess I'll open up fer I
4 i
i 1
i i
~
y
-e g
-.m
.-m
~-
5071 09 11 119 i
.1 AGBwrb 1
questions.
l l
2 MR. WARD:
I guess that last item would sort of l
l 3
imply to me that we might want to fund, for example, TRAC
(}
4 calculations beyond what might be considered as absolutely l
5 necessary for the IST program in order to get a better i
i 6
understanding of this particular issue.
7 MR. BECKNER:
That's a good point.
We are 8
planning to do calculations, but that's a point that I think 9
should be considered.
This is a very important concept i.
10 which we want to evaluate.
l 11 That concludes what I think we have had scheduled I
j 12 on the agenda.
13 MR. WARD:
Very good.
We might have a couple of 14 comments.
15 Could I ask:
Randy Carter, I noticed--
This I
i 16 information, I'm sure it's available, but do you have in
(
l 17 some handy form, the current costs for MIST in some sort of 18 a breakdown for the MIST facility, in some sort of 19 breakdown?
I
(
20 MR. CARTER:
Within the contract, the costs for l
21 the MIST and the OTIS programs are in, I believe it is I
22 fifteen separate cost control groups.- And that lets you i
23 have visibility of the costs for what was done on OTIS and I
24 also for what has been done on MIST, including construction, 25 analysis, test specification, testing and reporting.
{
f l
) -
- _ - _ _. _ _ _ _ _ _. _. -.... - -.-..~,-. _,.
5071 10 01 120 1
AGBwrb 1
MR. WARD:
Well, maybe I could just get this from 2
the staff afterwards.
Maybe I have it, as a matter of fact.
3 MR. CARTER:
It is within about the first five 4
pages of the contract.
5 MR. WARD:
Okay; it is.
6 Do you know roughly what the breakdown is on the 7
costs for the control room equipment, not the instruments 8
themselves but for the equipment for controlling and 9
analyzing data?
10 MR. JONES:
I believe Paul is on the distribution 11 of the various PMG documents that are issued.
One of those, 12 of course, is the monthly progress report, which has the 13 various cost elements associated; how much has been spent on 14 each or projected to be spent.
So I'm sure Paul can dig it
(^%
(-)
15 out for you from that.
And it is tied to the contract.
16 MR. BECKNER:
One of the cost control groups is 17 building modification and control room enclosure.
18 MR. CARTER:
That includes the actual building 19 itself.
20 MR. BECKNER:
And then there's a dash--
It may 21 not be clear: there's a "--control group," but that also 22 includes some other stuff.
i 23 MR. CARTER:
I guess what I would recommend, if 7
24 there are some specific areas you would like to know the n(_)
25 cost on, that those be identified, and let them be given to l
w
5071 10 02 121 1
AGBwrb 1
you.
2 MR. WARD:
Okay.
Tony, maybe you can ask Paul to 3
provide--
It's probably all available, but it's just a 4
4 matter of breaking it down, and then if we have any
{}
5 questions we can come back.
6 Okay; tha t's fine.
Thank you.
7 We would like to take a few minutes now, the 8
Committee members and the consultants, to make any comments 9
they would like to on what they've heard in the last day and I
10 a half.
11 Virgil, can I start with you?
12 DR. SHROCK:
I have really not thought this 13 through enough to respond in a comprehensive way, but I'll 14 do the best I can off the top of my head.
(
15 I think the things that we've heard in this 16 meeting have been quite encouraging in the sense that both 17 the calculational efforts and the experimental program for i
18 MIST seem to be moving along very well.
i 19 I think the areas in which additional emphasis 20 needs to be given have been brought out in the discussion 21 here.
One of those I think is the vent valve behavior, and i
22 certainly the experiement at the University of Maryland is l
23 the one that has the greatest potential to give further i
24 insight into whether some surprises may come from the vent (f
25 valve induced characteristics of the system.
l e
F-m
- e
-m mi-g y-vn-7.-
+m y
w
=e---y z--*i
- + - - > -
+1M i
w -
-y e
- --mw
5071 10H03 122 1
AGBwrb 1
So I think the use of those data is something 2
that perhaps deserves more emphasis.
And it sounds to me, 3
from this meeting, they will be getting it according to the 4
present plan.
}
5 This potential for scaling understanding I think 6
has certainly been emphasized here, and I agree fully with 7
the need to use these experiments to the fullest extent 8
possible better to understand scaling and these complex 9
two-phase flow systems.
But I do think that we have to keep 10 very clearly in mind the fact that the ultimate scaling is 11 the important one to the full plant.
It is largely the 12 code.
And'it does seem to me that in some respects there 13 may not be enough emphasis in lo,oking at the codes in some 14 detail, and in the sense that they are accomplishing this 15 scaling step.
16 The comparisons that were shown us this morning 17 which were characterized as being, I guess, satisfactory; no 18 major concerns arising from them; in my mind seems to show 19 more simply that the codes are tuned to the small scale 20 experiments when you make the calculations on the full scale 21 system with the code.
What you see is basically the same 22 phenomenon.
I don' t think that's surprising, but, at the 23 same time, I don' t think it's very conclusive in any sense 24 as a demonstration that the scaling that the code has O(_j 25 accomplished is necessarily doing a first class job.
l l
l t
-- ~
_ = _ - -
5071 10 04 123 1
AGBwrb 1
When we get some data from the larger scale 2
facilities in the other countries, that will help in some 3
respects.
It's not going to help much, I guess, in terms of 4
the B&W system peculiarities.
But, in a general sense, it 5
will help to build confidence in the scaling capabilities of 6
large codes.
7 I guess the character of TRAC as a first 8
principle calculation is something that has been emphasized 9
somewha t over the years, and I think an awful lot of people 10 out there believe that everything that TRAC does is indeed a 11 first principle calculation.
And so, for that reason, I 12 think that this type of problem that we spent a good deal of 13 time on this morning concerning the need to do a more 14 detailed jcb of analyzing what is happening in the steam O)
\\_
15 generator, to my mind is a bit of a disappointment.
- But, 16 more than that, I think it may lead to -- if a surprise 17 should arise, it may lead to a lot of dissatisfaction with 4
18 the code program in general.
Because I think there are an 19 awf ul lot of people that really believe that it is first 20 principle, first principle, and that's that.
21 So I would give more emphasis to.that aspect.
l 22 In terms of the steady state initialization in 23 the calculations, this is a problem that has existed for a 24 long time in the codes.
And I think that it is something
(~%
(_)
25 that is still disturbing.
l w
._,a-
,s
5071 10 05 124 1
AG3vrb 1
I think the codes should be able to generate 2
highly accurate steady state data.
If they cannot, then I 3
think that casts doubt on their ability. to deal with the 4
After all, t he transient should be nore
(}
5 dif ficult to calculate than is the steady state.
6 So these dif ferences in the steady state I think 7
should get a little nore attention, and those differences 8
ought to be exact.
9 There are =any individual co=2ents I would agree 10 with, such as Ivan's point that the code ought to be 11 incorparating the inforn_ation that slug ficw vill not occur 12 in the large pipe.
If that is an accepted conclusion, it 13 sounds to ne as though it is rather lucid.
14 Scae of the comparison's that 2 tad snoved whien, 15 again, indicated that the phenonenon looked pretty sinilar.
16
- guess in ny own sind, or like nany that I've seen in the 17 past, where it is a sort of subjective conclusion, and 13 different people see it differently.
19 An example of what looked very dif ferent to ne j
20 was the broken loop voiding which seems to cover a nurh 4
21 longer time span in MIST than it does in the full-scale i
22 pl ant, and has rather dif ferent characteristics.
There are
)
1 1
23 sose features that: are sinilar, but those that are 24 dissLnilar seen to ne to be greater.
Whether that is
)
25 impo r tant in the context of any safety issue is, of ecurse, l
4 4
5071 10 06 125 l
AGBwrb l'
another question.
But that certainly is of some 2
significance in terms of the codes' ability to follow the 3
details of the system behavior.
And I think that's what we 4
need to be looking for, is the overall ability.
( }~
5 I guess, in terms of the follow-on in MIST, it 6
certainly seemed to me desirable to have full power data.
7 The cost seems rather high, and I gather this will be a 8
factor in what Bill Beckner is looking at as an assessment 9
of the options.
Certainly $8 million would go a long way 10 toward the general facility.
So that, I would think, would 11 be given a lot.of consideration before deciding which way to 12 go on that.
13 That pretty well covers the key things that I 14 would say.,
(
15 MR. WARD:
Okay.
Thank you, Virgil.
16 Ivan?
17 DR. CATTON:
Basically, to reiterate some of the 18 things that Virgil. said, but I'm going to go backwards.
19 MR. WARD:
Okay, sir, go ahead.
20 DR. CATTON:
The last thing we heard, one of the 21 last discussions was on f ull power testing.
I really don' t 22 bel i eve it will be necessary.
I think by scaling of the 23 proper parameters and running in two different pressure 24 regimes, I think you can get the same in f ormation.
[()
25 The steam generator modeling I think needs I
m
-e
,.w...,,-
.w x
x-.
c
5071 10 07 126 1
AGBwrb 1
attention.
It may even require some separate effects 2
testing.
3 Here is a case where tuning of the code to small I (}
4 scale facilities may lead to troubles when you try to 5
extrapolate to a full size plant.
Some of this was brought 6
out in the plant comparison study where Thad showed the importance of the multi-dimensional downcomer.
7 8
-Multi-dimensional downcomers don' t exist in these small 9
facilities.
10 I think the multi-dimensional downcomer is i
11 important to more than just the B&W reactors, it's also 12 important to the CE reactors.
13 The multi-dimensional characteristics in the vent 14 valve behavior to me clearly emphasizes the need for using l(
15 both the University of Maryland and DSRI-2 in any code 16 assessment for it to be meaningful.
17 Our tour last night and the presentation this 18 morning by Randy I think showed that the hardware under the 19 MIST program seems to be very well in hand.
20 The multi-dimensional downcomer found by Thad 21 seems to hinge quite strongly on the possible inadequacy of 22 RELAP-5 for these kinds of calculations.
23 I think we need to take a better look at that 24 with the people in industry using RELAP-5.
()
25 That, with the comments made by Thad about the I
5071 10 08 127 1
AGBwrb 1
costliness of that kind of calculation, I think Research 2
ought to take a look to see whether or-not they can handle a 3
one-dimensional core' and a multi-dimensonal downcomer.
4 Because I think, really, the only aspect in any of these 5
studies where the multi-dimensional part is clearly needed 6
is in the downcomer.
7 The scaling studies that are underway I think 8
should yield some very helpful information.
In my view, 9
this is the.first time that I know of such a study being 10 made where each proponent is going to do a scaling study and then hopefully there will be comparisons, and but of that we
~
11 12 ought to get some good ideas as to how this kind of testing 13 ought to be done.
14 MR. WARD:
Thanks, Ivan, i
15 Carl, did you want to make some comments?
16 MR. MICHELSON:
A couple of small comments.
17 on the overall picture, I would first of all like 18 to thank B&W and the staff for a very fine presentation.
It 19 was extremely informative and interesting.
Unfortunately, I 4
20 had missed some of the previous things and, as a 21 consequence, had to do a little catching up.
So it was 22 particularly helpful.
23 The-tour of the facility indicated to me that 24 they are doing a very fine job of building a piece of f()
25 equipment which hopefully now will give us some good i
l t
I-l
=..
-- -. = - -
5071 11 01 128 1
AGBwrb 1
~ results.
2 I have some minor concerns.
As was mentioned 3
earlier, the vent valve arrangement.
Certainly there was a 4
dif ficult decision as to how to do it.
It still seems to be
- , {
5 an interesting compromise that was arrived at.
We are just 6
going to have to wait and see now whether it really is all
[
7 that important or not.
A nd, in part, that will perhaps be 8
observable from work that the University of Maryland is t
i-9 doing.
t 10 One minor comment on the facility.
I would 11 sincerely hope that someone' would do something to quiet down b
12 the center of the research activity in the control room.
l 1-13 You can hardly hear yourself think and converse with j
i j
14 anyb ody.
There really doesn' t seem to be a wise saving to
)
15 allow that high a noise level in the principal area where 16 decision-making will have to occur during a test.
It is i
17 remarkable that one could tolerate that level of noise.
18 Until one begins to see the results from the L
19 early-on testing, it is very difficult to tell whether the 1
20 higher power testing program is needed, or the extent to l
21 which it is needed.
I just don't know.
But there are
{
22 certainly a number of questions that will be. addressed in l-
- 23 such a higher power program that clearly need to be i
' addressed.
i
-24 l
(f 25 Unless there are some quite clear-cut answers i
i l
[
J I
f l
1
-,,.-e-e e.r,,,-4.-.-
em-m--
ew
.-%.n,,
9.c-_..m_,..
.-yy,,m-..-we,.,c-
%.~, w.
y-.-mm,,.-,w.
-ww.,.m,,3.--y-y-.1w---c-.m,,-,-
5071 11 02 129 1
AGBwrb 1
coming from the first part of the program, I kind of have a 2
suspicion that there will be a few surprises in the early 3
part of the program that will alter the plans for later.
(~
4 Anyway, it's wise to start thinking about the plan, I 5
believe.
6 That's about all I have at this time.
7 MR. WARD:
Okay.
Thanks Carl.
8 I've just got a couple of comments that I would 9
like to make.
10 First, I think what we've seen in the last day 11 and half indicates that the program has come a long way in 12 becoming more integrated.
I think I am quite comfortable 13 with the integration and the extent of the cross-checking i
14 among the various facilities, and the analytical methods
(
15 that have begun and planned.
16 I wasn't particularly comfortable with that 17 aspect of the program a year or two ago.
So I am very, very 18 impressed with how well that has come along.
i 19 The physical program, I mean here at the MIST 20 f acility, seems to be going ahead well and in a professional 21 manner.
I certainly don't foresee any problems there.
22 The need for something like a TRAC calculation of 23 some of the University of Maryland experiments seems to me 24 something to consider.
It strikes me that, you know, a
()
25 major weakness of the MIST facility, which is inevitable, is l
L.
1
5071 11 03 130 1
AGBwrb 1
the inability to model some of the asymmetric effects that 2
would exist in a plant, especially those effects in the 3
downcomer and the vent valve arrangement, and so forth.
And 4
it seems to me that that's the key thing that the University
{}
5 of Maryland facility, which has other deficiencies, that's 6
the key thing that that facility attempts to model 7
experimentally.
And it just seems to me very important that 8
those dif ferences be as well understood as possible, so that 9
we have some confidence in the understanding of how that 10 plant is going to behave in asymmetric transients.
11 I don't know how that can be done, other than 12 with a pretty good set of TRAC calculations.
Maybe there i
13 are some other ways: I'm certainly open to it.
But I think 14 it is important.
()
15 The steam generator auxiliary feed spreading 16 distribution question, it seems to me, is extremely
- 17 important.
I think it is an important issue.
Whether it is 4
18 important in the transients that a plant would undergo, I 19 don' t think we know ye t.
It looks like it would be -- take 20 a fairly major ef fort to explore experimentally, or to, you 21 know, go to first principles with TRAC or something like 22 that in order to provide an assessment of it.
23 I certainly think that some sort of sensitivity 24 analysis to give you an understanding of how important
()
25 differences in steam general behavior might be are pretty l
r
.~
_ ~ _ _ _. _. _
5071 11 04 131 1
AGBwrb 1
essential at this juncture.
2 On the code assessment program, I guess I would 3
just like to encourage you to remember that the country has 4
got a whole lot of B&W reactors, and let's not get lost in 5
that whole effort.
6 You know, most of us say that we really have 7
three kinds of reactors, and you've got a code development 8
and assessment program which is for two kinds.
And I think 9
you need to give particular attention to the third kind; and 10 maybe you are.
But it seems to me that that has been a 11 problem in the past and will continue to be, unless you get 12 some focus of attention.
13 I would like to get some feedback from you, I 14 guess, Lou, within a few weeks on what -- af ter you have had
(
15 time to thina about some of the comments you heard here, in 16 particular abaut the analysis of, you now, the TRAC 17 calculation of the University of Maryland facility 18 transient,and about how you plan to get a handle on the i
19 steam generator atx feed distribution question.
If we could 1
l 20 get some feedback on that af ter you've had time to think i
21 about 'it, we would appreciate it very much.
22 I don' t have anything--
23 DR. CATTON:
Would it be possible for us to get 24 the letter that NRR wrote in support of the full power
()
25 testing?
l
--4
+
_.L
-r A-4
- 4 5071 11 05 132
,R.
WARD:
Sure; I'm sure we have that.
Just ask 1
AGBwrb 1
M 2
Tony to send it to you.
{
3 MR. CAPPUCCI:
I'll send it.
4 DR. CATTON:
Excuse me; the letter dealt with the f")/
A-5
-AOD question in SECY 82-296.
6 MR. MICHELSON:
We just got a copy of thatl.
7 Which letter do you want?
8 DR. CATTON:
The one where they rationalized that 9'
MIST is suf ficient.
10 MR. MICHELSON:
I don' t know if this is the one 11 you' re referring to.
12 MR. JONES:
If that is the recent PMG distributed 13 letter, that's the right one.
That is the recent letter we 14 wrote back.
)
15 MR. WARD:
You just got it.
i j
16 MR. MICHELSON:
On another question which was 17 raised just a wee bit earlier, can we get some kind of a 18 general feel now?
I just don' t recolled in the budget for 19
'87 how much of this was being covered.
I
~
20 MR. WARD:
Let's see; the budget for '87 showed 21 either 4.5 or 3.5--
22 MR. MICHELSON:
The million-dollar discrepany was 23 because of some other argument.
Could you repeat?
24 MR. SHOTKIN:
The budget process for '87 is just
()
25 starting.
We' re not sure what our '86 budget is f rom i
l
,w-.
=
i 5071 11 06 133
'l AGBwrb 1
Congress.
The '87 budget item that you're talking about has 2
just been internally reviewed by NRC.
?
We could pay for roughly 50 percent of the
' _( }
4 u pg rade s.
So we are going to seek f rom industry roughly a 5
50-percent cost share.
6 MR. MICHELSON:
But what you asked for in that 7
budget is commensurate with reasonable needs for the long 8
range program; is that right?
9 MR. WARD:
I think, from what he just said, it's 10 assuming there is a long range program and that industry 11 will pay for half of the upgrade.
4 12 MR. MICHELSON:
Isn't industry paying for about a 13 third of this now; or is that right?
14 MR. SHOTKIN:
We ' re paying, I think, 42 percent.
t 15 But with in-kind contributions it gets complicated.
i 16 MR. MICHELSON:
So you' re asking really for. a 17 higher level of industry support when you come up with your j
18 budgedt number; is that correct? --or about the same level 19 as now?
20 MR. SHOTKIN:
In order for this to go through 21 Congress--
As you know, they are being very critical of the
.i 22 research utilization of our research, and industry has to 3
23 support program in order for it to go.
That's our feeling.
24 MR. MICHELSON:
Well, when you tell me that
)
25 budget was based on about half of the support coming from i
s c
1
5071 12 01 134 1
AGBwrb 1
industry, I'm wondering: is that about comparable to the 2-support you are now getting for MIST 7 Half?
3 MR. SHOTKIN:
The nu=ber that you saw was just a 4
back-of-the-envelope estimate, as I understand.
That 5
8-millon cost, now some of that could come from in-kind 6
contributions.
7 I don' t know how much detail you want to go into, 8
but it could be waiving fees or other things that might not 9
be exact costs.
10 MR. WARD:
I think Carl has a simple question, 11 though.
Given the restrictions on NRC research budgets for 12 the on-going program, are you really looking for more 13 support from industry than you did for the MIST program so 14 far, or around the same?
15 MR. SHOIKIN:
About the same.
16 MR. MICHELSON:
That's what I wanted to get an 17 appreciation for.
18 MR. WARD:
I guess that's it.
19 I'd like to express my appreciation to Bob Turner 20 and the people here at the Alliance Research Center for 21 their hosting.
The f acilities are excellent, and it has 22 made the meeting go very well.
We appreciate it very much.
23 Thank you again.
24 (Whereupon, a t 12:23 p.m.,
the Subec=mittee
()
25 was adjourned.)
l t
i r
-.w
CERTIFICATE OF OFFICIAL REPORTER This is to certify that the attached proceedings before the UNITED STATES NUCLEAR REGULATORY COMMISSION in
%e matter of:
NAME OF PROCEEDING:
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON EMERGENCY CORE COOLING SYSTEMS DOCKET NO.:
PLACE:
ALLIANCE, OHIO
\\
DATE:
THURSDAY, JUNE 13, 1985 7
were held as herein appears, and that this is the original transcript thereof for the' file of the United States Nuclear Regulatory Commission.
(sigt)
(TYPED)
ANNE G.
BLOOM Official Reporter ACE-FEDERAL REPORTERS, INC.
.O Reporter's Affiliation V
i r
I L
T.
f 6
a n --
Com aa ri son Ana ysis 3
~~ R AC 3 ' / v 0 D '
by Thad Knight Jime Lime Saf ety Code Development Los Alamos National Laboratory l
l ACRS ECCS Subcommittee Meeting Alliance, Ohio June 12-13, 1985 l
l O
O O
I t
BACKGROUND i
l
- PLANT ANALYSIS FUNDED BY NRC/RES UNDER FIN A7217. TRAC CALCULATIONAL ASSISTANCE NRC PROGRAM MANAGER H. S. TOVMASSIAN l
PRINCIPLE INVESTIGATOR J. F. LIME i
- DRAFT REPORT OF ANALYSIS ISSUED APRIL 1985
- TRAC PLANT MODEL REFLECTS DATA RECElYED FROM B&W f
- PLANT CALCULATION IS COMPARED TO MIST PRETEST-7 l
CALCULATION FROM 0-2500s. PLANT TRANSIENT BEYOND 2500s IS BEING CORRECTED FOR REACTIVITY-FEEDBACK INPUT ERROR.
53
~:RAC PLANT v0JEL I v 3 ROVE V E N ~~S LOWER SECONDARY-SIDE WATER INVENTORY REACTIVITY FEEDBACK MODELED l
B&W REACTIVITY INSERTION RATE B&W REACTOR COOLANT PUMP CURVES MODELED RCP POWER INCLUDED IMPROVED VENT VALVE MODEL (CODE UPDATE)
UPPER-HEAD RECIRCULATION FLOW SLIGHTLY LOWER MFW FLOW COASTDOWN S/G AFW AXI AL SPREADING MODELED S/G OUTER-SHELL HEAT STRUCTURE MODELED 9
e e
e
81 f
10 esi!
i
.E r e
s"1E i
@9 m
e i Sgi e-Sis O
- , / :1
~(
Y j; i'y3x
- @ g /, s $r t: t. i.f
.~4 2<
ja a
a
- J eD a.
[.i'i A
- Y' d
5
- k.,
.g
" I % *,.~.:,-;:,-
c
- \\
O 8
9 5
2 T~I H
a f
T*
a!
y Z
s s
Igg g
l e
ii a
C-d l
I i
i K nn N
M..., f_M.....
, <.v.%.,7 iii i
O W
m 5 a
o h
h 'y M I
.gl'ILCO T ca i
e g5.
s s
s
-.1 j
I
=s eg D
o w
a.
g y
h h
h3 e
e
=
=
s
.? e
- 2 ;g D
2 A, v I
- J Ni TJ2 E
e Q
@E
- 1-w - G's$,
CeD - y.: sag;
-,,,G, h. g's:.. -i g
r t
1 7 Ds 5
~~
(fi W
l
. g... j
..-a7-e.
'$. IhN I E
e tI../
g I
15)E l
6 49' I
E I.'\\
V h,- I g : 1 @!i
^
U3
)
- }
lIU h !
I azi!
si TD *
's
=II
V ST 3 L A sF-VODE._
REAC-OR VESSEL N OD l b G
( r=2, z= ' 2, 0=6)
CONTROL-ROD GUIDF IURF S (1 PER SECTOR)
ELEVATION (m)
/
.............I.1. 3 7 6.,,, _ _,
U 10.033
_j LEVEL to e 2
S vtwr utvts is VENT 3
d
/__
8.762
-f VALVES T
to T
t__o 8.356 CFT
'I E
9 J UT
-cri swJ C
~
l D....,7.8 4 8,,,,,,
8 HOT LEG 1___
O 7.137
+
1 8
" ~ ~ ' ' ~ ~ " " 'C O L D 7
LgyLL g j
- -- --._ C ".... 6.10 4,.... LEG.
3 3
Q in
,6
........ 5. 2 2.1, _., _ _ _,
-5 ACTIVE
........ 4. 3.3 8.........
4 coto Ltc at coto Ltc s2 REGION
........ 3 4.5.5.........
3 i
s
........ 7.5,7 7 _ _ _,,,,, _
Hot itc a 5
por Ltc e 4
2
')
\\
.......... 0.8 4 5........
Coto Ltc 42 coto LtG B1 1
LEVFL 8 O
l 1.9177
' 2.1669 h) r(m)
Q Q
Q
< JU AL-CH AN NIL ST Ah/
GEN ER ATO R STE AM-EX I T ANNULUS 2
1, M
3 143]
- 07. AFW 4
141 Qi4o 2 J$~ &
1007. AFW 102 in3
+EE 2
13 13 7
1 6
146--* STEAM 12 12 14 0 14 2 190-fy 11 11 11
-J3@ BEE-15 190 [
1 JIOO*-- O 10 10 10 1
E X g 9
a g
8 8
4 DOWNCOMER 8
8 7
7 7
7 5
6 6
6 6
6 AD JUNQT__ COMP _ONENTS NO.
DE_SCR.l P_T_I_ON IJ.PI 10 5 10 % PRIMARY PlPE 130 90 % Doll.ER TEE 18 Ur Er-132 10 % HOLLER TEE 134 MrW SPLIi TEE
~
~
~
106 104 13 4 130.
ASPIRATOR SPLIT TEE 907.
107.
907.
107.
14 0 90 % SUPERHEAT TEE' 142 10 % SUPERHEAT TEC PRlWARY SECONDARY 14 4 5IEAM EXIT TEE 190 DOWNCOMER TEE b
f PRIMARY-TO-SECONDARY HEAT STRUCTURE MODELING NON-WETTED HS.
OF 10% CHANNEL
% TUBES l
WETTED fp
+y 13
- -~
JAFW 13
.h[k.;
12
- h.- -
7-(100%)
6%
12
+ --
11 E$
II
-'t. ~
8%
i.
10 i!
10
- d 10 %
+
~g 9
y 9
- t 8
',)
8
- [
7
?
7 6
[j$
6 i{.
p
- = _=.
w,'-
v_
= _
=. - -
- c. <.
PRI HS SEC PRI HS SEC 90% CHANNEL 10% CHANNEL O
O O
O O
o f* TRIPS AND SAFEGUARDS ASSUMPTIONS TRIPS SETPOINT QEL Y I. REACTOR TRIP ON LOW PRESSURE 13.203 MPa 0.2 s i
(1900 psig)
- 2. TUROINE TRIP AND STOP VALVE CLOSURE ABOVE O.5 s
- 3. RCP TRIP (LOSS Of OffSITE POWER)
AROVE 2s
- 4. MAIN FEEDWATER TRIP ABOVE 2s+12-s COASTDOWN
SAEEG.UARDS I. HPI FLOW = MIST-SPECiflED FLOW SCALED 817
- 2. AfW FLOW = MIST-SPECIFIED FLOW SCALCD 817 BOTH HPI AND AfW REFLECT FULL-CAPACITY SYSTEMS.
I 4
t
,~,
TRANSIENT SEQUENCE OF EVENTS TIMES (s)
EVENT PLANT MIST *
- 1. BREAK OCCURRED 0
N/A I
- 2. REACTOR TRIP SIGNAL 71 N/A
- 3. AFW DELIVERY STARTED 111 229 j
- 4. TOP OF HOT LEG A SATURATED 186 266
- 5. HPl DEllVEtY STARTED 203 229 l
- 6. 3.5% POWER REACHED 229 229
- 7. TOP OF HOT LEG A VOIDED 500 350 l
- 8. TOP OF HOT LEG B VOIDED 1100 1150 1
- 9. BCM STARTED IN LOOP A 1800 2500 l
10.~ CALCULATED TO DATE 2500 3730
I
- MIST EVENT TIMES SHIFTED 17 7: SO THAT START OF CORE DECAY CONCIDED WITH PLANT 3.5% POWER OCCURRENCE.
1 l
le e
e e
O O
O MIST / PLANT SBLOCA COMPARISON PRESSURIZER PRESSURE 16 PLANT
-2200
MIST (SHIFTED 177 S) 34 -
-2000
-1800 12 -
?
-1600 g
3 v
g w
g E
D 10 --
W N,
-1400 3w m
w m
(r W
o_
m
-1200 8-
-1000 6-
-800 4-
-600 i
i 0
250 500 750 1000 1250 1500 1750 2000 2250 2500 TIME (s) 2 je
MIST / PLANT SBLOCA COMPARISON INTACT LOOP SECONDARY PRESSURE e
12 i
i PLANT
-1600
MIST (SHIFTED 177 S) 10 -
-1400
-1200 8 --
[
2-
-1000 w
l g
6-
\\.
E3
-800 (n
U) g W
E m
Q-o_
4-
- -600
-400 2-
- 200 0-0 0
250 500 750 1000 1250 1500 1750 2000 2250 2500 3
Tit (s)
O O
O MIST / PLANT SBLOCA COMPARISON BROKEN LOOP SECONDARY PRESSURE e
12 PLANT
-1600
MIST (SHIFTED 177 S) 10 -
-1400
.-1200 8-9 s
a Q_
g d
-1000 3
W l'
x y
y T
a 6-3 W
m
.-800 m
m W
w g
g O_
o_
_ -600
-400 2.-
-200 0
0 0
250 500 750 1000 1250 1500 1750 2000 2250 2500 TIME (s) 4
MIST / PLANT SBLOCA COMPARISON INTACT-LOOP SG-PRIMARY LEVEL e
18 PLANT
--- ------ MIST (SHIFTED 177 5) 15 -
--50
%ee v
W y
{
_J 10 -
',7 d
5 Q
-30
=
=
3 8-3 6-
--20
~
4 ---
-10 2
0 250 500 750 1000 1250 1500 1750 2000 2250
2500 TIME (s) e is d
O O
e
O O
O MIST / PLANT SBLOCA COMPARISON l
BROKEN-LOOP SG-PRIMARY LEVEL 18 PLANT
-.::::::::::. MIST (SHIFTED 177 S) 16 -
q
-50
~~....
14 -
12 -
~
-40 g
_1 3
W W
i J
10 -
b W
g (r
W
-30 k
Q 3
8 --
3 6 --
-- -20 4 --
-10 2-0 250 500 750 1000 1250 1500 1750 2000 2250 2500 TIME (s) i
MIST / PLANT SBLOCA COMPARISON BROKEN-LOOP SG-SECONDARY LEVEL 18 PLANT
MIST (SHIFTED 177 5) 16 -
-50 14 -
~
- ~40 2
12 -
v W>
8
'_j
_J 10 --
___________________________y_ ce__:cccc-:_:ce:c g
-30 w
g e
g 8-l l
l l
6-
/
- -20 l
l l
4-l l
-10 l
~
l g
2 0
250 600 750 1000 1250 1500 1750 2000 2250 2500 TIME (s) r 9
e
00 5
O 2
0
,5 2
)
2 S
77 1
0
,0 D
0 E
2 G
T N
F I
I N D H
I 0
S V T (S
,5 OO N
7 1
T I
A S R D L
I A N P M P E 0
g M B
,0 5
O 1
C U
)s A G
(
0 C E OOL
,5 E
2 M
1 L
B T TI S O T H 0
N P
,0 0
1 A O LP O L
/
0 T
S T
,5 7
IM C A
TN 0
I
,0 5
.,e tl l ll:.l.
- l 0
!!l' 5-
,5 2
)-
O 0
e 2
8 6
4 2
0 2
1 0
0 0
0 0
1 z O G O < e k O' O Q Y o
g l
- l I
ll
MIST / PLANT SBLOCA COMPARISON BROKEN-LOOP HOT-LEG U-BEND VOIDING 5
1.2 PLANT
~
MIST (SHIFTED 177 S)
- '.y v l3
~ ],
i 0.8 -
l zO i-l O
0.6 -
l l
T u-l e
l c) o_
o.4 -
I 5
i
.t.
0.2 -
- A l :l l
l l
- l i
^ ^-
0-
- 0.2 0
250 500 750 1000 1250 1500 1750 2000 2250 2500 TIME (s) u dn O.
O O
O O
O MIST / PLANT SBLOCA COMPARISON i
e COLD-LEG-Al MASS FLOW 1500 PLANT
-3000
MIST (SHIFTED 177 S
~
1250 ~
AND SCALED 817)
-2500 1000-
-2000 750-9 d6,
-1500 g
S 3
h
-1000
[r 500-9 W
m W
<C 2
250-j 500 f
{...
v.
IV M /M\\
[
-250-
--500
~
~~
-500-0 250 500 750 1000 1250 1500 1750 2000 2250 2500 g
TIME (s) so
(5/C;) /.iO li SSWi O
C C
o O
o o
o O
P O
O C
O o
C a
o o
o o
a o
0 O
I I
C eq N
N C
0 N
N C
_a N
N LD Nm NN C
nO
_C w
=
C OC N
ru L.
=<
=
4 O
O t; v@ U
-Q 4
to (O
"wC
<.qZ C-S J E r<
O.
13<
l r
"2 O
~
g
-w O 'O 4
O s
I k
a
<3
,6 o
x O.
N.
D m
I
$O
- t w
O J
C a 7 f
O a
~
< C J a C-O DU
( :*,
^
W N
2 n
_x y
6b m
-B N
e i
~
O O
O
^
C O
C C
o e
o o
o 0
0 C
C N
O N
C N
N 7
l t
(s/53,) fio13 ssyn z
l Q%
O O
O MIST / PLANT SBLOCA COMPARISON INTACT COLD-LEG B MASS FLOW l
1500 t
PLANT
-3000
MIST (SHIFTED 177 S 1250 ~
AND SCALED 817)
~
-2500 1000-
-2000 750 -
'UP
,ds',
-1500 g
3 3
500--
9
-1000
[E W
m
.W l
m 4
2 250 --
--500 2
~,
~
~. -
~~.......-.............
~
~
-250 -
--500
--1000
-500 0
250 500 750 1000 1250 1500 1750 2000 2250 2500 TIME (s) 32 02)
(s/qj) /AOB SSVW 8
o o
o o
o o
o o
o o
o o
o o
o o
o e
o a
o m
n m
o o
I i
o o
c
~
o
_a m
R SR o
_o m
O o
WO g
N F-y t_ a l
5<
v-3 go o
OO Hvm
~S W,'
kwa cr -
aOz
<L m
- a. s <
C w
o os l
2 <
-8 o
o o6 oO 9
O
_a l
u w
- 2
&o I
F m
-t
.~
o W O 9
_o Z o
_.)
L.J
'f NX
-ao O
g-x 8
2e o
-oa o
-a
.. ~
~
,.o o
o o
o o
o o
o o
o Q
o D
o c
c a
m U
o n
a
~
m a
i I
(s/6 ) /AO U SSVW 3
n m
J
O O
O MIST / PLANT SBLOCA COMPARISON BREAK MASS FLOW e
12 0 PLANT MIST (SHIFTED 177 S
-240
~
AND SCALED 817) 10 0 -
- )
l,,,
- 210 l ,'
l i
I
- 5 I
gie It g,
g l
' l'h
_ - 18 0 80-
~
I Q
'l 2
h.
i, i
~ '\\
%n
.- 15 0 o
S k!
3O 60 --
l W
m
- 12 0 W
l M
I W
<c
<C 2
l 2
40-l
_ - 90 l
l
-60 V
l 20-l
-30 l
l 0
0 0
250 500 750 1000 1250 1500 1750 2000 2250 2500 TIME (s) 2i
MIST / PLANT SBLOCA COMPARISON VENT VALVE MASS FLOW 1500 PLANT
--3000
MIST (SHIFTED 177 S 1250 --
AND SCALED 817)
-2500 1000-
-2000 750--
?
6
-1500
_o o
3 3
i
' j
O O
500-
-1000 u_
l; g;
!!r mp i
-500 2
2 250-i 4/
3
)k'lA*I
- -O 0_
_l k J/ r L s-L 4 n,L,k i
L lt c
a
-250
_ --500
--1000
-500-0 250 500 750 1000 1250 1500 1750 2000 2250 2500 TIM (s) 9
oS g w < 2 9 d J_. ? o m o C L
t 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 5
0 1
0 5
0 5
0 5
0 0
3 2
2 1
1 0
~
5 f
2 i
O l
(
0 f
l
,5 h
2 S
2 7) h 77 0
1 1
,0 8
o D
0 ED k
2 t
(
FL l
N H A f
I C
0 I
S T (SS
,5 O
7 N
I A TD r
1 S N R W LI i
A PMA P O 0
h L
l
,0 M F O
/
5 l
1 C S f
S g
)s A A
(
(
0 C M
,5 E
O q
2 M
OBL T 1
I E
T L
S
(
N 0
T I
s 1
,0 N E 4
0 A R 1
(
L O P C
/
0 l
T s
l
,5 S
7 I
M
,'i N
0
,0 5
/
/
0
,5
,[
l !lll'!
2 i
i' 0
O 0
0 0
0 0
0 0
0 0
0 5
0 5
0 5
5 0
5 2
0 7
5 2
2 5
1 1
1 Q[
Ww<2 nLdJK LTO a
5 g
!llI
CONCLUSIONS AND RECOMMENDATIONS l
1
- MIST AND PLANT TRANSIENTS SHOW SIMILAR BEHAVIOR
- PRESSURES AND LOOP VOIDING COMPARE REASONABLY WELL
- COLD LEG TEMPERATURES DIFFER BECAUSE OF MASS-FLOW DIFFERENCES AND PLANT COLD-LEG FLOW CIRCULATION l
- LOOP ASYMMETRY OCCURS IN BOTH TRANSENTS
- CORRECTED PLANT CALCULATION SHOULD SHOW PRIMARY-SYSTEM REFILL AS IN MIST PRETEST-7
- CONTINUE TO SUPPORT FIRST PROGRAM WITH FULL-SCALE PLANT CALCULATIONS WITH BOTH 3-D AND 1-D VESSEL MODELING.
l
~
l ASSESSMENT OF CODES:
110W AND WHEN?
MIST ACRS MEETING ALLIANCE, OHIO JUNE 13, 1985 1.0VIS M. SHOTKIN ll.S. NilCLEAR REGULATORY COMMISSION 9
O O
O O
O IllLBBALilYDRAblIC TRAi4S.ltl4TS (CONTIllDED1 hAJOR TOOLS TEST FACILITY ASSESSMEliT AllD i
(VEliDOR GE0 METRY)
C0tiPUTER CODE QUANTIFICATION'0F ACCURACY PLAllT ANALYZER SEhlSCALE (W, CE)
TRAC-PWR i
i
> liPA-PWR l
ROSA-IV (H, CE)
RELAPS l
l
)
MIST / UNIV. OF MD. (L8W) n)
COBRA / TRAC i
l 1
l i
& DOE LABS i
biilVERSITY RESEARCli 1
i x NPA-BWR COMP 0iiENT TESTS TRAC-BWR i
i ATF-tiWR (GE)
O)
RAM 0liA/illFA 1
I I-IST (GE) i I
- INTERNATIONAL CODE ASSESSfilliT PROGRt+1 (USillG DATA FR0tl INTERNATIONAL TEST FACILITIES)
5 5
5.
1 i
3 s
i
=
=
=
1 g
E E
i
=
=
=
1
=
=
=
0 S
2 2
I l'
5 W
=
=
=
l O
e
.E E
=
M M
=
E f li w
w w
.e o.
e.
u
.B B
=
2
=
3 5
+
+
I g
=
=
e e
=.
-g
=
=
Ma.
W W
=
e D
Q=
3 E
g
&~
t
...Tk s
=
r=
=
mO eC 4
-3 33
=
3
$=
o 5
es E
8 o
5
-5
-5 e.=
v4 N
>N
>N Ee E2
-o
=
g
!= $2 k:
=
3
.. W
_g _3.g.. y- -g
~
E E
3 5
o O-h N
=8 5
N m
g E
=
5 5
s 3
=
a e
5
~
g d
d
=
=
r r
J t
=
E 0
=
1 o
2 9
E a
O
-8 x
e a-
- - o=
3
+
a
=
a E
2 e
e e
.=.
==
e O
m w
w
=
=
E e
E a
2 5
y
'T w
Ji Cg.
g
=
g I.
E 0
W e.
2 E
e 4>
+
at
==*
=
2 e
g e
e e
5 M
M m
=
g Q
og
==
o~
a e-e "f %
8 EO
==
e
.r
=
n-
==
s e.
$4 o
~<
>q sa
>.,e
>4
- M w9 3 at e tas eC ee q>
g d5 4w
=
- w.
w ad w taf
~
t 1
=*
i m-e 8
=-
w h
M t.
2
=
o u
a a
d.
o w
O
'O O
FUTURE PLAlls FOR TRAC-PUR ANI) RELAPS TRAC-PWR CODE OF Cll0lCE FOR FUTURE PLANS NATURE ASSESSED 3-D CAPABILITY IlVMERICS COMPANION TO TRAC-BWR.
Cll0SEN TWICE BEFORE (1978,1990 1
CAH ADDRESS ALL PERCLIVED T/II SAFETY ISSUES CONTINUE PLANS FOR " DESK-TOP" PLANT ANALYZER
. FUND USE AT INEL (T/H TECliNICAL INTEGRATION CENTER) l
FUTURE PLAllS FOR TRAC-PWR AliD RELAPS (CONTIliUED1 i
RELAPS C0liTIllut IMPROVEMENT AND MAlllTENAllCE SUPPORT I
DECREASED DIRECT llRC SUPPORT INCREASED USER SUPPORT DOMESTIC USER GROUP l
ICAP TEST FACILITY NEEDS (MIST, SEMISCALE)
COMPLETE PLANI ANALY2ER CAPABILITIES IN FY 1986 ELIMlllATE PLANNED ROSA-IV ANALYSES IN FY 1986 AilD FY 1987 l
O O
O
h O
v v
b)
INtemt Hv0RautlC l' AN58 tuts: MAJOR itst FACitilfr5 p
FTIMS FT 1986 FY IM7 FT IM8 FTIM9 FY 1990 Ty 1991 SEMISCALE F/5 IEST SERIE5 58t0CA W/0 HPI CLO5EOUT ANALY5t$
LIQelD NotDUP 1[515 5CALING COUNIERPARI NOINSALL MI51/0ils (84W)
MI51 CONSTRWCiloll MIST TESTS:
MI5I IISI5:
Futt-POWER TE511MG CLOSECUT ANALTSE5 FNAERWillAll0N. MODE liiLift6ANOelfED.
IRAN 51110N. 5BLOCA FULL-POWER M005 FIST (Gt)
PNA5E 11 ANALT5ES AND CLOSIGUT MALY$tS
- AtPORIING A054-19 (JAPAN)
IPIT 5.G. H.I.
LIQUIS Hol0UP AlitRNATE ECC. PLANI N.T. IN 5.G.
CLOSEOUT ANALT$ts L5ff IISTING TNI. 5810CA.
REC 0vtRY itCNNIquf 5 MIURAL CIRCULAll0N IMIIUiMtED$fu ANilCIPAIED TluutSIENT PRIMART LOOP COOLING MECMANI$M MB-2 (f)
TESTS:
I I ill75Gia CCTF-II (JAPAN)
SE5T-tSilmit TESTS l
SCIF-II(JAPAN FORCED FitD if 515 SCIF-III (JAPAN)
SMEEDOWN IESIS COLD LES INJECil0N CopSINED INJECil0N CLO510UT ANALYS15 itsis 11515 UPIF (GERMANY)
CONSTRUCileN COMPLEltD SEPARAlt EFFECTS 11515 INTEGRAL EFFECis T[515 CLOSEOUT ANALTits 5NAttDOWN 11511NG i
ADWANCED TE5I SCALIIIG STUDT DESIGN SIART CostSTRUCTION AIF-PWR TE5IING Alf-PWR It5flNG AIF-PWR ItsilleG-AIF PWIt TESTING FAtlilif (ATF)
(W/CI)
STARI CON 518UCil0ll ATR-BWR AIF-BWR IESTING AIF-BWR TE5illeG AIF-BWR TE5fileG R
w l
Yi g
c s
B&W Owners Group 9
'N/'N/
MIST STATUS ACRS MEETING NRC acne 13. 1983 B&W PRESENTED BY:
H. R. CARTER A/'N X
X l
t 1
l EPRI l
l i
Babcock & WIIcox a uccemen company l
8 HEATER ROD AND BRAZE JOINT CROSS SECTION INCONEL TUEE g
7 l
l i
BORON NITRIDE 7
4 INDUCTION COIL
.002"+m N
Uf !
'U Short Circuiting Due tc:
2n!i 1
i ri lZ;i l
e Temperature Indication lU[j l Ol-j Error O
l['
e Nickel Plating on Leed
,0 has Reduced Meltino 1
I 1
.b o nt BRAZE MATERIAL I--
I i
l HEATER ROD l
SHEATH i
l SILICONE MOISTURE l
SEAL g
i i
/-
%j
'. $$ 0 l]ATU_n NICKEL PLATING
l l
S~~ A~~L S O r CC R E
- EA~~E R RC:
- \\ S~~ A
_ _ A~~::C \\
e 17 BRAZES COMPLETED e
13 BRAZES ACCEPTABLE e
4 BRAZES UNACCEPTABLE l SHORTED HEATERS REMOVED:
1 O
O O
i STATUS OF REACTOR COOLANT PUMPS e PUMP CASINGS AND IMPELLER MOCKUPS INSTALLED e PUMPS (4)
FABRICATED AND READY FOR TESTING e PUMP i HAS COMPLETED PERFORMANCE TESTS
- HEAD / FLOW
- IRRECOVERABLE LOSSES
- NPSH e PUMP i READY FOR ACCEPTANCE TESTS HOUR RUN AT PRESSURE / TEMPERATURE HOUR RUN WITH AIR ONLY O
O O
W z
e' HW w
U w
F-c Z
J M
07 o
a 2
a o
a m
H cD tn Z
W L7 CD C
w w
c cn u
C Z
Z aF t.o c
w z<
J 22 O
C!
M
<W a
- o F-I z
C I
C Z
M C
wU O
W3 W
C z
CC C
1.:
-O Z
DH w
Z C
IP-M (f)l Cn CD J
M w
C D?
cn z J
C.
w >-
H c
F-- >-
C 5!
WH F-W C
W H in W
<l C
tt)
H C>2 U
Q Lj !
z J
MO Z
cn c
or-a=
F-O O,c g'
F-w
>- - d-w J
w fT2 W
C Cf3 < W >-
CD c
c z
w tn z -
E-w Z
a to w
c_
o F-w=
<J C
in to <
C c.
ww o
3 O
J P-W Q
W Z
Cm
=
O
&a zo m
. G.
tn C.3 w
zw au F-.
tn Cn in u
o a
C w<C
~
c
_2 O
F-W W _3 m
Z H
F-y z
C
<C H
d<
W O
tn C
C.
CC<
z Q_ H 1-tn C
w W
u) < C
_J 3 5-a w
O T
F-En ICw
< CO W
C U
C Ef P-O C O.
F-b -
OIU I
W C Cn zOZ n Cn C
w D
zlt U Lt3 _.J O
c.
o ww waw zw Di OE_!
}-
En W
Cn Q
c:
F-C < &--
M*
E3 IW<
H W
g g
g g
g g
F-E t.:
}-
CD
_J t
W LJ CD Z
O (f)l IHZ W
O z!WQH Q
U m!
S 9
4 t --I tn!
si O
xi
A%.8 x
n e.
Va u
oo gna; 0
'h t
c 0
o N
5 v
V N
to k
O=
S t
l s
c j/
1 1
9
O Data Acquisition System l
h i
(0-1 Volt)
Pipe Wall n
[vwe.c%r (RMS Output) iA
.t.
14p 400 E Conductivity "b
Probe
+
Frequency l
/
Oscillator i (30 HZ) l 7
Conductivity probe electrical circuit g
650 0.5 - Test 220756 0.45 625 330 S A u ",.,
d
~
4 s
Hot leg U-Bend 600
+-
t onductivity Probe W-310 P (HLCP14) 575 6 6
+
0.35 i
-Thermocouple (HLTC09) e
.o h
'\\
550 290 0.3 s
o.
a.
.b
\\
E-E y
s,
' y >
525 F b
s e
m o
0.25
'~s 270 s
N v
N u
f o
0.2 500
~
N Conductivity Probe Output (HLCP14) 250 0 15
~
475 y
g
Thermocouple Output (HLTC09) 1 1
I I
I I
I I
I I
450 0.2 0.4 0.6 0.8 1
1.2 1.4 1.6 1.8 2
2.2 Time, hr Hot leg conductivity probe HLCP14 and hot leg thennocouple HLTC09 signal history of Test 220756 O
O O
l 1
Number of data points = 780' ECP14 9.3. --- W=1.644.9923*TDWTRAftME y
- 9. G _
O L
T S
O.4 l
~~-, ' ~.
9.2 _
e i
I I
I I
430 See 520 548 SES 539 Get TDETRATtME F l
Hot leg conductivity probe HLCP14 output as a function of temperature and linear curve fit for test 220756 1
1
9 Temperature, *F 0
100 200 300 400 500 600 700 40 l
i I
I i
i 1
O O
O o
30 52ii 2
i O
3 t
20
.a E
O
~
v l
I Uo
~
10 O 4.05 ppm NH 0H 4
0 I
i f
f I
I f
e g
0 50 100 150 200 250 300 350 400 Temperature. *C l
Amonium hydroxide electrical conductivity versus temperature for concentration of 4.05 ppm 1
1
0.5 _
S.45 _
. ~,
0.4 _
8.35 _
y 0
l L
T
- 3. 3 _
S 8.25 _
__..-[*'
s.
- 9. 2 _
Conductivity probe output
- - - - - Calculated wetted probe behavior 0.15 _
b I
I I
I I
I I
I I
I I
c.2 9.4 0.6 9.3 1
1.2 1.4 1.E 1.5 2
2.2 IS.CP14 TIIE BRS Hot leg conductivity probe itLCP14 signal history and the predicted transient response for fully wetted probe.
O O
O
1 e.s _
V
./.
1 O
I
- i' 3
y ',\\l
/
\\
S.E _
?
F
.li.
R n
C T
g,4 _
I J
0 M
s 9.2 _
1 ij I.
Y b
e 1
q Io 1
I I
I I
I I
I I
I l
8.2 9.4 S.S 9.5 1
1.2 1.4 1.E 1.3 2
2.2 BEIP14 TIIE BRS
-- - Void Fract ion f rom DP Calculated void fraction profile from the hot leg conductivity probe llLCP14 output signal.
O O
O
GUARD HEATER STATUS o HEATERS (500)
ORDERED AND RECEIVED i
i e BENCHSCALE TEST IN PROGRESS SINCE JANUARY TO VERIFY RELIABLE OPERATION
- TEST ON SMALL DIAMETER PIPE
- CYCLE 300F TO 600F TO 300F IN TWELVE HOURS
- APPROXIMATELY 300 CYCLES COMPLETED WITH NO FAILURES O
O O
^
+
9 Y
e e
- e 4
- Og u
as e..d
"" 8
~
3
'Me
,. ~ ~. h
- 7 W7
~
4 W1 K
Ged W
l
>=
I l
t E
tea O
l l
e
~
E d
5 9
O h
a "9
+
~
9 e
O e
9 9
O e
,,/-
e MIST DENSITOETER TRAINING C0WLETED AT INEL MARCll 20, 1985 TOPICS e SOURCE IIANDLING AND INSTALLATION s DENSIT0ETER llARDWARE ASSEMBLY l
e DESCRIPTION OF ELECTRONICS e SET-UP, Tile 0RY, AND ADJUSTENT TECHNIQUES I
e CALIBRATION PROCEDURE AND TECilNIQUES e llANDS ON SET-UP AND CALIBRATION 4
e e
e e
O 0
O O
O
9 MIST DENSIT0ETER HARDWARE STATUS O
2 HOT LEG SP0OLS DELIVERED 1/18/85 4 COLD LEG SP0OLS DELIVERED 4/11/85 4 COLD LE6 SP0OLS DELIVERED 5/17/85 6 SETS OF ELECTRONICS TO BE DElfVERED 7/26/85 4
0 9
e e
1
l i
l MIST DENSITOETERS INSTALLATION SCHEDULE 2 HOT LEG SP0OL PIECES JUNE 15, 1985 8 COLD LEG SP0OL PIECES JUNE 15, 1985 6 SETS OF ELECTRONICS AUGUST 15, 1985 SOURCES AND FINAL TRAINING MARCH 1, 1986 i
t 4
l l
l
1*
MIST DENSIT0ETER FINAL TRAINING TO BE C0WLETED AT MIST MARCH 1, 1986 l
TOPICS REVIEW SOURCE HANDLING AND INSTALLATION e
REVIEW SYSTEM CONECTION AND GEERAL OPERATION e
e DEMONSTRATION MD REVIEW 0F ACTUAL SYSTEM SET-UP MD DATA ACQUISITION e
CALCULATE CALIBRATION CONSTMTS MAKE MD EVALUATE SINGLE POINT CALIBRATION CECKS o
UNCERTAINTY ANALYSIS MD ' LESSONS LEARED*
e 0
i O
O O
4
6_
m a_.
W Deeh w.
w l
4
- 7 s.
. ?.
9
--m-um 6
2 l
.r r,,,-
-. s.
i e
I
- f-,
4 e,
_, I,
,,j ' t 4 e
r.. _,
...~
gg O G3 4
v; e,m i
..r g
-o-"e s
g6 -g
- m a e s
W
=
r e.
e
~
g.
g,
-,-, t..
i
~
r....'-
.. 3
, e I
=
e
. [p-wm.
ed4 e
p' 41 ~ -
6 I
- *8**
f e
A
'e g
e.
=
j
-.J
~
g e
9 g
~
o J*
e r-I f.
t M
e-e
>=
M U
>=
U M
M E
1 w
9 g
O.
l O'
e
=
t
- t j
e
4 I
i REVISED MIST INCONEL 600 CTC FABRICATION / CALIBRATION SCHEDULE 9/10 4/23 5/21 5/28 6/3 6/24 7/5
'8/2 9/3 Prelim Fab four Calib Four Qual.
Order Inconel Fab /Assem Sep Eff Data inconel Inconel Pack / Ship Hateria Weld Proc First Pro Test Analy Probes robes -
Probes v
v O
v v
v v
v v
l L
V g
9/17 I
l l
Instr.
Sep Effects /Ca11b Data Analysis Test Loop Mod Setup s
g 4
O p
9, O
e e
/
b
- j r.
- ,, : : :: =. -
~ ~& - --- : - -- :.
7
.s.
.J
~ - -
s.
..~,-.i,..
6
- ^
,,. s,....~'. $'&j g,. & H.
z
..,,....i..,.P,*
% h,.. p ~..,,..
.p-..
e q,..
,a
-r
..i
.. I t. g.
s.
e
.f e
.s g
.r D
'.g,-.
-)a.',..: [*
j-3
. s p'..',, g v...,' a
..n. ;J a e -
- f r - T.
...W. w,.
, (;.-.? ?g* a
.i
..s.
?..
w e
+e
- tt.:. %.. ':.
v..
',;;. C.h.
Q.f. K.? :,,6.,j., ?. ' ".%}. ' %+......; g'..)* '. \\
,_[
., : A. ;.' ~.,,' 4..*
. co
.. -9
,j 'r ' %
..y..
L._ k. Q..
s','~.
t a
...r.
,3,.w..
c
. r,,s.,,-. r <..,a
,. m,.
,. s e,
.. a e
.u.
s.....
.s... A.
s
. ;, s..
e. <. -
.. +.
? $ wi ?
,.**F Q,,
'[*=
f;
..,,'A'.,.
e.
(8[g,,.ht.'.7 p f. !4 ~
, g-
, ~ *.
t
.e
. - - a 2*,.ze,... < s ! g...(.s.r.;.v....g.
y
..a s-w
~.,
.s.
. ',...: v g.
,J,... @ $
- W..$ iW..h,
. w-.
.t
' Z. m-I
,4.*..' ?.'.4.
y.. c f L. i.; g
-8).S, *<
.e
- t.
./
.p
.6.,
4..,.....8.,4
.)
,e
-. m ;,.1
.,,;-. ~p.
. ;, f. '..,
.- e ~
s.s ~,..,.,s....g.
3
~
s,.....3....,~g-...
<.vq.....s..,...s..,,.,.-
s,.
- w... -.
.w-
..., r y,. \\,+... * ; r.,,..L., %. n..
. + s.,
.e r.t *W.... W.h.:.' s.%...y ';C / ' ; + " @
..'.l C
..,.. y...
s.
.e,...,. e.. *
.., ' w.
% n -M.,s m,..o..jf.i'. N 'h,y* h W ;.,c :... ' ', '
@..,..4.'
y:*-
4 3..r.
...s..
.s
- .~ ;:,*,.- W *c.. m.
- 2. 3 1,.
..,. : L,
'- 4...
- .r...;..... - ; n....y ses n. p,...
... m ?->
?
r
, %..~.
e
. m.
..r a rw e
,.7:. ~-:., p:, t-M. M,.
w ' :,y[ a v+
' w,
..?-
..s u..
z:
,,. ?.
- .,';.:.).,,W..ie'.'..^..
.A,i v.m.
o c
..t.-
,4.'...*
.n. ('-
, m. 9.. -
p.
.. -,-,*.,, -.j -. j... e.
.r.
t '. 9 l' *'. ; s, s
- l. -.'; W. ;(
ij i
- s.
.,..,,..a 7..'.,..',.
..;. ' /.>5,- t ~-
I,.
b - m g :- u..
s.
... 'O s,0 ?;.
- b Af..
- r;i.
...,u g
-e c
L.D.s..".2 -.J ; f. 2 i9.
O N CN -
- lll.m-.
,r
.S.r-
.~
v1' 3-
,*. C
-cN.y-p ;
.g.. U g
- W CN seg c 4*
s-,A H'
W W
5
-t- ' y. ;;'s ;4 ;W.-
.,,.,f.., *y, p,.
e
' '.. '.', A,13 y,..
M v
g e
~-
,Y j
c m,
l'
.J
, 4,, p 2).
s,,,.
e.,
Q.
t
... ?; 7.
N, g.
2.;
m
.J
..;, f J!. f.4 & *, -~,., h..,..."g: 7.,, M n, N
- 4.*
M*
. M Lo me*
.,.w
-q.-
u.,..
y
..I W
.. t. & -.. e.
". p.
y p
- +
,( 1,.;,,,, W;,. $...,.4.,t
-i' Cz:.
E 7.
.e
- ll:3
.-=
,. A)..' LA s.
- .'.%*g W '. :'.
- ,.;,, j..; '.
>=-*
C
... t. y W..', ':
- L. ;.,.'
t'.
'c ;
ca:
ag.
A. ca 4
.4
,.A
,, - ".'W l,.N.f'4 W ".",,.G. :j fff. c t. '.f. >:ll3 =. t gg; y
.. s.
8 Iv
,. 4..' wI c ::n, w g., e W.;,4...s ' ". +.'s, M' ~ ca:,~ v N C t
d,
V '. %,. ' * 'y':
. '. d.
W'
?'* W
. W w*
.. :.. h
. nL t..q;'%)f w 9 J^
p, Ly cc p
- d i:.,!.r ~. u.L 3..,a% N ::s ;s -
Q
,. ca p-
..*- c.;.:.i- $J M. J[. q,* q g. M.c';. & -*
w
=
- 8. - -
g, M-p p
e s p...g u ;:,.a 1.. c:
W.:% v-4ycz: ? e ' - =3
- C 5
- n.% >-,6 A. 4,..i;.v.'. A
.S;-:.
4 2
J t
.s -
.. :::n
.n..,,
.A.a...,......
,, >- ;. t.
.s.,. ?.
J.
t,g
- s..m.4 r j,4 _M,,..c F... t~ A s,;.u.n 4,
u., -
.at
. o 1
c ca
.-=
.n.
... A. },f de
..:.M-
- .*4A M U.d W 7QC
- .;r. m. 2,..
v.'* g:. W n E an s =
- n s
- f.
Y '.m. <.v.&g5.h.pw;~,f'3.M j'i 833e'-~.,.' e'.-
i-<- c:? w.
- - 9pf W
W e m,..
c.f W.
. ft.' p,,....a.,yi H" r.g su ~
e e
.a 5 k' ~t 8.h d*g "h iI,Fhf'K-3,@* ;s
..N I-
- fy
. h k.. P..-s e a.i.* =..
r * * *
',6.
a
- p- *;.*. I.
.,... L.i... -...
a tW r
2 th s,-
).
6,.%.,... f,.,s. e. s..s t -
s, s
,c
.. t
\\
- ..$ p A-
..u
.n..=
7 i
- +. 1 1 ;. t 'h n.y s 3.. "* ftC; i.
.6,,
., q. a,
.. +... -..,
s.
- p..
..,.-.,;.*'.s 3...,. c,
.. ? '.
,.'.4," f ' a...
.. "4 4
?. r.
.e:..,
r.,....e.,.
.s..
.:J.
2.:
.,y
- . f
.1 3; " g f. i ) -. -.1.
.a*...-
. c. g.,., '
..o.
g
.t.
-A.. h.
.,...-a s..
.c.
jf
?
r
..> s
~
t b
g O
t,
g,.,
g
-=.e
O i
e.l 2
N
.IwI e
x s
o 4
O
- w:
e en k
a f
1 W
e Y
k O
h!
3 o
e l
v i
/
l i
D-=,
v.
G w
e-i 8
3 L
h M,1 L..qt C
i
.tch f
e
'/,)
ia4 y
l {.)
g I
.-?
=
.i O-T S
I M
G R
N I
O T
E F
EW B
E 5 O
S S
E 8
S T
T 9 I
E T
R A
T H
1 O.
O N
I
/
D M 3 S
0 E
M I
1 E
E J
T O
S N
C 2 C
T D
E B
N 1
S U
A
. /
E S E I
C L
R R
N L
R P
S U R
C J
A 4
E R
C N
E W
S O
R C
P A
L V
F O
s l
l i
NRR DATA NEEDS
- USER NEED LETTER (10/31/84)
- CURRENT IST PROGRAM ADEQUATE FOR SBLOCA
- DATA NEEDED FOR CODE ASSESSMENT
- NON-LOCA TRANSIENTS
- RISK DOMINANT SEQUENCES
- F0LLOW-0H PROGRAM IN MIST RECOMMENDED
- FULL POWER SIMULATION BEllEVED NECESSARY l
- ESTIMATED COST - $8 MILLION
- ESTIMATED SCHEDULE'- SEPT.,1987 - OCT.,1988 O
O O
l TABLE 1 l
TRANSIENTS TO BE ANALYZED IN FOLLOWON MIST PROGRAM
- Steam Line Breaks **
O
-Spectrum of Sizes y.
-Spectrum of-System Failures
- Feed Line Breaks **
-Spectrum of Sizes o::.
- Loss of Feedwater**'
- Steam Generator Overfill **
- Alternate ECCS Designs
' High Point sprays
-Use of High Point Ventt for
(
Feed and Bleed
- SG Tube Ruptures 1
-Failure Simulations
- Plant Cooldown Simulations
-Single Loop Natural Circulation Cooldown
-Natural Circulation Cooldown with a Solid Pressurizer
- Normalization to MIST Tests'*
-SBLOCA
-SGTR
- Risk Dominating Sequences
-SBLOCA without HPI
-Station Blackout
-ATWS O
- Allows assessment of effect of using 10% power as initial conditions in current MIST program.
e
filST L0flG-TERf1 PLANS PRESE!iTED TO ACRS ECCS SUBC0fV11TTEE MEETING JUNE 12-13, 1985 ALLIAi1CE, 01110 llY:
WILLIAfl D. BECKNER REACTOR SYSTEMS RESEARCil BRAflCil DIVISION OF ACCIDENT EVALUATION OFFICE OF NUCLEAR REGULATORY RESEARCll 9
O O
EUTURE OF MIST FACILITY AFTER PHASE-III TESTING NRC INTERESTED IN -PROGRAM 0F FULL-POWER TESTING.
CONSIDERING POSSIBLE FOLLOW-0N TO EXISTING PROGRAM.
FULL-POWER UPGRADE WOULD ALSO BE USEFUL AT INEL FOR POSSIBLE FUTURE BWR TESTING.
WILL CONSIDER TECHNICALLY ACCEPTABLE OPTIONS GIVING MOST COST-EFFECTIVE TESTING CAPABILITY.
INDUSTRY SUPPORT WILL BE SOUGHT OVER THIS SUMMER.
4 e
e O
O
ASSESSMENT OF EXPERIMENTAL STRATEGIES ALTERNATE SCALING CONCEPTS OFFER THE POTENTIAL FOR MAKING USE OF SMALLER, LESS EXPENSIVE FACILITIES FOR INTEGRAL TESTING, MIST, UNIVERSITY OF NARYLAND, AND SRI FACILITIES OFFER A UNIQUE OPPORTUNITY TO EVALUATE THESE DIFFERENT SCALING METHODS, PRELIMINARY SCALING COMPARIS0N IN PROGRESS, BUT FINAL EVALUATION OF NEW SCALING CONCEPTS WILL DEPEND ON POST-TEST EVALUATION OF RESULTS, e
O O
O O