ML20210B148
| ML20210B148 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 09/12/1986 |
| From: | Atomic Safety and Licensing Board Panel |
| To: | |
| References | |
| CON-#386-807 LRP, NUDOCS 8609170460 | |
| Download: ML20210B148 (140) | |
Text
OIGhAL UN11ED STATES g-NUCLEAR REGULATORY COMMISSION IN THE MATTER OF:
DOCKET NO:
LRP INQUIRY INTO THREE MILE ISLAND UNIT 2 - LEAK RATE DATA FALSIFICATION O
LOCATION:
BETHESDA, MARYLAND PAGES:
1185 - 1323 DATE:
FRIDAY, SEPTEMBER 12, 1986 90
/
l ACE-FEDERAL REPORTERS, INC.
0 OplReporters 444 North CapitolStreet Washington, D.C. 20001 Ilbfl I${$h.L.' [ Nbbbe I
NATIONWIDE CCVERAGE I
1185 CR28040.0 ERT/sjg
-1 UNITED STATES OF AMERICA 2
NUCLEAR REGULATORY COMMISSION 3
BEFORE TEE ATOMIC SAFETY AND LICENSING BOARD 4
- - - - - - - - - - - - - - - - -x 5
In the Matter of:
Docket No. LRP 6
INQUIRY INTO THREE' MILE ISLAND UNIT 2 - LEAK RATE DATA FALSIFICATION 7
_x 9
Nuclear Regulatory Commission Fifth Floor Hearing Room 10 East West Towers 4350 East-West Highway 11 Bethesda, Maryland 12 Friday, September 12, 1986
()
13 The hearing in the above-entitled matter convened"Et 14 8:30 a.m.
16 BEFORE:
17 JUDGE JAMES L. KELLEY, Chairman Atomic Safety and Licensing Board 18 U.S. Nuclear Regulatory Commission Washington, D. C.
19 JUDGE JAMES H. CARPENTER, Member Atomic Safety and Licensing Board 20 U.S. Nuclear Regulatory Commission Washington, D. C.
21 JUDGE GLENN O.
BRIGHT, Member 22 Atomic Safety and Licensing Board U.S. Nuclear Regulatory Commission 23 Washington, D. C.
24
(
'25 ACE. FEDERAL REPORTERS, INC.
202 347 3700 Nationwide Coversee M3364646 l
.. J
1186 1
APPEARANCES:
2 on behalf of GPU Nuclear Corporation:
3 ERNEST L.
BLAKE, JR.,
ESQ.
JOHN NASSIKAS, ESQ.
4 Shaw, Pittman, Potts & Trowbridge 1800 M Street, N.W.
5 Washington, D.
C.
20036 6
On behalf of the Employees:
HARRY 11. VOIGT, ESQ.
7 MICIIAEL McBRIDE, ESQ.
LeBoeuf, Lamb, Leiby & MacRae g
1333 New Hampshire Avenue, N.W.
Suite 1100 9
Washington, D.
C.
20036 10 on behalf of Jack IIerbein:
11 JAMES B.
BURNS, ESQ.
Isham, Lincoln & Beale 12 Three First National Plaza Chicago, Illinois 60602 CilRISTOPHER W.
FLYNN, ESQ.
Isham, Lincoln & Beale 14 1150 Connecticut Avenue, N.W.
W shington, D.
C.
20036 15 On behalf of Gary P.
Miller:
MICIIAEL W.
MAUPIN, ESQ.
17 M.
CIIRISTINA IIENSLEY, ESQ.
Ilunton & Williams 18 707 East Main Street Richmond, Virginia 23221 19 On behalf of Former Metropolitan 20 Edison Employees:
SMTIII B.
GEPIIART, ESQ.
21 Killian & Gephart 217-218 Pine Street 22 Box 886 Ila r ri sbu rg, Pennsylvania 17108 on behalf of the NRC Staff:
24 7y JACK R.
GOLDBERG, ESQ.
( l 25 MARY E.
WAGNER, ESQ.
U.S.
Nuclear Regulatory Commission Washington, D.
C.
20555 ACE-FEDERAL REPORTERS, INC.
202 347 3700 Nationwide Coverage 800 33M646
~
1187 h-1 CONTENTS 2
RECESS:
4 NOON - 1262 5
6 7
8 9
10 11 12 i
V 13 14 15 16 17 18 19 20 21 j
22 L
23 24 ANj 25 l
l l
l ACE-FEDERAL REPORTERS, INC.
202 347 3700 Nationwide Coverage 800 33m
28040.0 (7RT 1188 U
1 PROCEEpINGS 2
' JUDGE KELLEY:
Back on the record.
Good morning.
3 The Board has a procedural question that it wants 4
to put to counsel.
It's kind of a difficult question.
We 5
are not going to ask for any immediate response.
We want to 6
put the~ question to you and let you think about it.
If you 7
want to respond after lunch, at the end of the day, we'll see 8
what your desires are, but it's a question that doesn't go to 9
this morning's business.
In' fact it goes to the next phase 10 of what we'll call individual responsibility-type testimony.
11 But'it seems to us that it's a question that we need to be 12 forehanded about, and look at now rather than later, so'we
-s k.
13 know where we are going and we haven't really discussed it 14 directly before in any detail.
15 The question broadly stated is this:
Should the 16 Board consider the views or comments of Staff wi,tnesses, 17 meaning Mr. Russell, Mr. Capra and Mr. Christopher when he 18 comes, and Mr. Stier,_on the testimony of the party 19 witnesses?
That could be phrased, both-in terms of their 20 views on prefiled or, if they were to be heard from later, 21 their views on all the testimony of the party witnesses.
22 We are not talking -- we don't have in mind direct 23 conclusions about veracity.
I'll put it in those terms.
We 24 are not going to ask any witness:
Do you think some other
(~
25 witness is telling the truth or not?
That's what this Board
\\.s}
ACE-FEDERAL REPORTERS, INC.
AC-347-3700 Nationwide Coserage
%0tL336-6M6
28040.0 1189 N,xBRT
]
1 is supposed to decide.
2 What we are concerned about is possibly getting 3
some helpful input under the circumstances oftthis case, 4
where the history of a lot of these incidents is pretty 5
complicated and you are talking about witnesses who, 6
themselves -- I think in the case of Mr. Stier it was mostly 7
people working for him, but Mr. Capra and Mr. Russell, who 8
have lived with this investigation for a long tims, and we 9
are concerned about, if there are inconsistencies in 10 testimony of party witnesses, if there are gaps in the 11 testimony of a party witness, we are concerned that we know 12 about that, that we pick up.on it, in short.
And we'll do
,_s
(
)
13 the best we can to absorb the record, study the testimony, 14 and ask the appropriate questions.
But we think we might not 15 get the full picture unless we had input from these other 16 sources that I've mentioned, the Staff witnesses, Mr. Stier.
17 It seems to us the question really wouldn't arise 18 in the normal-case where Staff was a party.
Take the Staff 19-input, for example.
If Staff were a party, Staff counsel 20 were participating fully, were cross-examining, then if there 21 were an inconsistency, presumably he or she would bring it 22 out.
But here we have the Staff in this position built for 23 this case only, so it seems.
They are not a party, 24 technically.
The question came up the other day in the 25 context of this panel, namely could the Staff ask a question g s.)
LJ ACE-FEDERAL REPORTERS, INC.
202-347 37(X)
Nationside Coverage.
800-3364M6
28040.0 1190
( ~~)lRT
\\_
1 of one of the other witnesses?
We first said no. The Staff 2
protested a restriction on how they view their role, and the 3
upshot was we' adopted the rule that applies to this panel and 4
the way this panel works, and we left the question open 5
whether at a later stage the Staff could put questions to 6
witnesses other than its own.
So we see it as an open 7
question and let me just state two or three options that 8
simply occur to us.
9 One would be to not get this source of input and 10 just rely on the Board to see inconsistencies, gaps, and 11-phrase the appropriate questions.
4 12-I suppose another approach that is not way at the gQ 13 other end, pretty far away from the one I just suggested, 14 one, would be to recall Staff witnesses when the party 15 witnesses have testified, hear their views in full,'not only 1~6 on prefiled but on whatever they said in the hearing.
I 17 suppose a disadvantage there is it can go on forever, if you 18 feel the party witness is entitled to the last say.
But it's 19 just a possibility, I suppose.
20 Another possibility, sort of a compromise, it 21 seems, would be to solicit comments from Staff witnesses and 22 Mr. Stier.on the prefiled and simply ask them to point out 23 any inconsistencies, gaps, other points that they think we 24 ought to ask a question about, and confine our input from l
/~S 25 that source to that approach.
And there may well be some
(_;
ACE-FEDERAL REPORTERS, INC.
3 0-347-3700 Nationwide Coserage 8 % 3364646
28040.0
<sBRT 1191
(
)
- ~_/
1 other variations.
I simply put those out as something that 2
occurred to us on our first thought about it, and we would be 3
happy to hear any other suggested variations.on point.
But 4
let me just ask counsel, do you understand the basic question 5
that'we are raising?
6 MR. MC BRIDE:
I do, your Honor, and since you are 7
going to have obviously one rule for the proceeding it might 8
be useful for party counsel to have a discussion perhaps with 9
Staff counsel as well before we respond.
10 JUDGE KELLEY:
You might even discuss it among 11 yourselves, sure.
We might bring it up Monday morning --
l' 2 Tuesday.
I beg your pardon.
I think a Tuesday morning t'_J i
13 discussion of this would be adequate for our purposes.
I'm 14 thinking in terms of time.
15 Let's see, if we decide we would like to hear 16 suggested questions on the prefiled -- well, I'm not'sure 17 that these witnesses have even read the prefiled.
If they 18 were going to do that, they would have to have some time to 19 do it so as'to not throw our schedule off too far.
We 20 thought we ought to look at this question now.
But, Tuesday 21 morning -- is Tuesday morning satisfactory for a discussion 22 of this?
I would urge you among yourselves to discuss it.
23 If you can agree on something, fine.
If not, you can lower 24
-- if you can narrow the range of options, that will make it
,w 25 easier for us.
Okay?
Thank you.
t_J ACE-FEDERAL REPORTERS, INC.
202-347-3X0 Nationwide Coscrage 80lk336-6M6
28040.0
("]RT 1192 U
1 JUDGE KELLEY:
We were discussing a range of 2
questions at the end of yesterday.
Mr. Goldberg advises me 3
this morning that'Mr. Russell overnight thought of something 4
he wanted to add.
If you want to.just move to that?
This is 5
a follow-up on yesterday's statement --
6 MR. GOLDBERG:
Yes.
I think Mr. Russell would 7
like to add some testimony in connection with his answers to 8
questions 269 and 271.
9 MR. MC BRIDE:
Judge Kelley, I'm opposed to this.
10 The witness testified yesterday that he answered these 11 questions with advice of counsel.
I did not ask what he 12 discussed with contnsel.
But it seems to.me now.they have ha'd fq V
13 another bite at the apple off the stand.
Typically, on 14 redirect -- if this is what this is -- you are not permitted 15 to consult with your-counsel before your redirect occurs.
It
.16 seems to me that the' Staff shouldn't be able to have this 17 this way.
18 JUDGE KELLEY:
I ask for information.
Did we do 19 follow-up questions on these, yet?
l 20 MR. MC BRIDE:
Yes, your Honor.
I 21 JUDGE KELLEY:
269 and 271?
22 MR. MC BRIDE:
If you are referring to 269 through 23 271 in Russell and Capra; yes, we did.
24 JUDGE KELLEY:
Did we get up to 276?
(~}
25 MR. MC BRIDE:
No, your Honor, we stopped with v
l ACE-FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coserage 8(XF336-(686
20040.0 BRT 1193
, ~s l
4 gj 1
271.
2 JUDGE KELLEY:
And took follow-ups in the middle 3
of a section?
4 MR. MC BRIDE:
It's headed TMI-2 administrative 5
procedures and it's 269 through --
6 JUDGE KELLEY:
I'm looking at a different set.
7 That explains it.
All right.
We understand the dispute.
8 Rather than us conferring on this we'll discuss it later.
9 We'll come back to this if we rule in the Staff's favor ~.
If 10 not, we won't.
11 I believe we pick up with 272.
12 off the record a moment --
9' '/
13 MR. BLAKE:
This should be on the record.
14 JUDGE KELLEY:
We'll be on the record for 15 Mr. Blake's comment.
Go ahead.
16 MR. BLAKE:
Can I ask whether the Board would 17 inquire whether or not it is 'a goof that Mr. Russell -- has 18 he misstated something?
Is he uncomfortable with what it is 19 he said?
Before the Board makes their determination on 20-whether or not this opportunity ought to exist -- I don't 21 think Mr. McBride would take issue with it, if there's a goof 22 in the record, whether or not we ought to cure it while the 23 witness is here.
24 MR. MC BRIDE:
I certainly would not, your Honor.
25 If the witness feels he has to correct something he said or f3
(_)
ACE-FEDERAL REPORTERS, INC.
202-147-3700 Nationwide Cos erage R&336-6646
28040.0
(~gRT 1194 Gl 1
left some misimpression, it would be an entirely dif ferent 2
matter.
3 MR. GOLDBERG:
I think if I have a discussion with 4
Mr. McBride and Mr. Gephart, they may withdraw their 5
objection.
6 JUDGE KELLEY:
Why don't we do that at the break.
7 Thank you, Mr. Blake.
8 Off the record.
9 (Discussion off the record.)
10
' JUDGE KELLEY:
269 reads as follows:
On page 11 11 of your testimony you refer to the accuracy of containment 12 sump inventory monitoring as being " limited." Is it not true
~d 13 that while total leakage collected in the sump may exceed the 14 amount of the RCS.
Leakage the amount of water collected in 15 the sump would not be less than the amount of unidentified 16 RCS leakage?
~
17 MR. KIRKPATRICK:
No.
I don't consider this to be 18 true.
It is trup that on a long-term basis, the amount of 19 unidentified leakage should not be greater than the amount of 20 water -- excuse me -- the amount of water collected in the 21 sump would put a maximum limit on unidentified leakage.
22 However, I have made a comparison between the water collected 23 in the sump and unidentified leak rates, based on daily water 24 additions.
In many instances the unidentified leak rate 25 exceed the sump collected leakage when a comparison is made ACE-FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coverage 23346M6
~x 28040~0 BRT 1195
,q V
1 on a daily basis.
This may be due to evaporation in the 2
containment air or to condensation at some location other 3-than the sump.
4 JUDGE KELLEY:
Thank you.
The question was only 5
put to you.
If other witnesses want to comment --
6 MR. COLE:
I think we take strong exception to 7
that for this plant, being where the sump was specifically 8
' designed -- the sump in this plant is not like TMI-1.
TMI-2 9
sump is divided into two parts.
One is a very large purchase 10 pump, another is a small pump for collecting leakage as a 11 result of Reg Guide 1.45.
Secondly, this is not in 12 containment'and therefore the evaporative loss problem of any
_h (G
13 moisture going out the stack being vented I think is 14 inappropriate, and we found this the most practical, 15 straightforward and one that should have been used for 16 checking the reactor pressure boundary leakage most 17 appropriately.
So we take strong technical exception for 18 that, at this point.
It is not true at all plants.
19 JUDGE KELLEY:
Mr. Stier?
20 MR. STIER:
If I might add, at my request, MPR did 21 an. analysis of sump collection and I think, although we 22 haven't seen Mr. Kirkpatrick's analysis, it is our best 23 judgment that cump collection was quite an effective way of 24 Fdetermining collection of leakage in containment.
I don't 25 want to use the phrase " unidentified leakage" because as we ACE-FEDERAL REPORTERS, INC.
202-347-3700 Natiortwide 8'.oserage RW33&M
28040.0 g3BRT 1196 U
1 discussed the other day, there's a definitional problem.
But-2 leakage in containment can be measured with a good deal of 3
precision by collection in this particular sump.
I can't 4
answer for sumps generally.
5 JUDGE KELLEY:
Mr. Kirkpatrick, further comment?
6 MR. KIRKPATRICK:
My basis -- my answer is not 7
based on theoretical possibility.
It's based on the data 8
that was actually available to the operators that we used in 9
plotting _the sump leakage and-there was a very large 10 variation on a daily b' asis during a time period when you can 11 show that the actual unidentified leakage did not have very 12 much variation.
,-U 13 JUDGE KELLEY:
Mr. Russell?
14 MR. RUSSELL:
I think it's also significant to 15 note, we have this in the record already, that the shift in 16 daily procedure did not include a conversion of water 17 collection in the sump to gallons per minute.
The water.was 18 collected there.
If one did the appropriate calculation to 19 determine _ gallon per minute. leak rate, as you can do after 20 the fact, yes, it then becomes an indicator that can be 21 used.
But other sources of leakage from the steam system or 22 from other fluid systems in containment would go into the 23 sump.
24 It would appear in this case to be an upper limit g-25 but it would be very difficult to correlate.
I think k)J ACE-FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coverage 8@336e46
20040.0
, sBRT 1197
\\
't./
1 Mr. Kirkpatrick's. testimony is that it did vary significantly 2
and, so, we don't see that it is that significant to this 3
proceeding because it was not,.in fact, used and the 4
technical specification was not completely implemented.
It 5
did not convert to gallons per minute.
6 JUDGE-KELLEY:
Dr. Harrison?
We'll just go back 7
and forth until we get this honed down.
8 MR. HARRISON:
Let me just make a few comments on 9
this particular subject.
10 JUDGE KELLEY:
Sure.
11 MR. HARRISON:
I don't know, I have not seen the-12 analysis that Mr. Kirkpatrick refeIs to.
However, we made a r'-
13 very extensive examination of a sump, based primarily on the 14 alarm printout, which gives you an indication of when the 15 pump started and when the pump stopped.
And we 1-ooked at all 16 the information that we could find on the alarm printout.
We 17 looked at a very large number of times, almost over the-18 entire operating period.
I think my perception looking at 19 those is that the sump was not an inaccurate, or gave any --
20 it was a very good measure of what was happening in the 21 plant.
In some cases there were instances of actual problems 22 in the plant with leakage of some kind.
The sump would 23 respond very promptly, would show those changes, and I was 24 very, I guess pleased is perhaps the right word, as to the 25' accuracy and the responsiveness of the sump.
We explained Os.
ACE-FEDERAL REPORTERS, INC.
Ac.347-3700 Nationwide Coserne 800-33MM6
28040.0 r~'3RT 1198
(~)
I this in, I think, considerable detail in our report, showing 2
the curves of how the information was obtained.
It shows how 3
it responded.
And I believe I would like to say we just 4
feel, not knowing what the analysis is that Mr. Kirkpatrick 5
based his assumption on, our analysis would not reach that 6
conclusi.on that it was inaccurate or unresponsive.
7 JUDGE KELLEY:
Can I just raise a question here.
8 I think the last time we bad a disagreement on this order, 9
there was an adjournment which led to more understanding.
10 You say you haven't seen Mr. Kirkpatrick's 11 analysis.
Is this the kind of -- it seems like a fairly 12 narrow issue that would lend itself to further discussion.
['sV) 13 We are happy to keep going now if you think we are refining 14 the question, but would it be better to take it to the 15 informal context and bring it back?
16 MR. STIER:
In all honesty, Judge, I doubt it.
17 But we can try if you would like.
-This is not a question of 18 calculation of a particular set of data.
This is a question, 19 a broader question of the reliability of a major component of 20 the plant.
21 JUDGE KELLEY:
Why don't you just go ahead, then.
22 MR. COLE:
I would like to clarify our position.
23 We are not trying to say the operators did or didn't use the 24 sump.
All right?
I find the same problem.
But I'm saying
(-]
25 the data that was available, that was computerized readings
~%)
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-33M686
20040.0 BRT 1199-
,r S V
1 that we got out so there was no issue of inaccurate logging 2
or things of that nature, we have discussed-the numerous 3
instrument errors and oscillations and mathematical errors' 4
for the inventory balance.
Those types of things were just 5
not involved.
There was obviously a small instrument error 6
involved in the sump but nothing to the magnitude that was 7
involved in the inventory balance test.
8 MR. KIRKPATRICK:
I would like to add that I do 9
agree that the sump is a very good indicator when something 10 changes.
We saw peaks in the sump data at times when we knew 11 there were peaks in the unidentified leakage.
We also saw 12 peaks in the sump data at times when we saw evidence that 13 there was also steam leakage.
However, just taking the 14 simple pump-outs, which is available to the operator, we
-15 plotted those and there was simply too m'2ch variation on a 16 day-to-day basis.
I didn't try to go back and do a 17 complicated error analysis or correction of that data.
But 18 just pumping -- plotting the sung pump-outs, and the volume 19 you would expect with each pump-out, there was too much 20 variation to be able to say that you could tell, today, that 21 the leakage would not exceed such a number.
22 MR. COLE:
Just so you can see what our data shows 23 for the thing he just talked about, you can see it on. figure 24
-- in~Stier report IV-A, chapter IV, figure IR-2.
It plots 25 it for a period of October through March and it appears that O-ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationside Coserage 80tk336-6646
28040.0 7 xBRT 1200 b
1 the sump would indicate there were a total of 10-times that 2
the leakage was above 1 gpm, as defined, unidentified leakage 3
was above I gpm as identified'by Reg Guide 1.45.
4 MR. STIER:
There are some other figures in that 5
section, also, that deal with sump collection.
The figures, 6
IV-1 and IV-3 and IV-4, all plot out the rate of sunp 7
collection for different periods of time.
The section that 8
Mr. Cole cited before describes how MPR did their analysis, 9
how they arrived at their conclusions.
I think it is an 10 important issue.
Because a lot of our work was based on 11 looking at that sump data and making judgments about the 12 extent to which leakage was actually occurring in the plant, 7-k-)
13 using that as a basis for evaluating leak rate test results.
14 We had methods of checking that.
But we felt that this was a 15 very reliable way of getting a baseline for unidentified 16 leakage.
17 JUDGE KELLEY:
Thank you.
18 MR. RUSSELL:
I have reviewed the information in 19 the Stier report and I would agree with the statements that 20 have been made about its use.
I would like to point out, 21 however, that while there were, I think it was 10 periods of 22 time when the rate of collection of leakage in the sump 23 exceeded 1, there were. fewer than 10 times where that was 24 potentially from the reactor coolant system.
They could not 25 confirm those.
That is, in effect, the type of inconsistency O.
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cos erage SM33MM6
i 28040.0
)
/^33RT 1201 k-)
I that Mr. Kirkpatrick was talking about, because other sources 2
of water collectedgin the sump.
3 Staff agrees that the sump is a.very accurate 4
means of collecting all leakage into the containment.
But 5
the question, as I understood it, was related to reactor 6
coolant system unidentified leakage.
If you have other 7
leakages that can go in the sump, it would affect that 8
measurement.
But it is an effective way of showing 9
conformance to the technical specification 6.1 that we have 10 talked about previously, that is the limiting condition for 11 operation for sump collection leakage.
It just was not 12 implemented that way and not used that way by operators.
73dJ 13 But it is true that because of periods of time 14 where they were getting large leakage to the sump, that they 15 thought they had leaks.
And other evidence will come out 16 that shows that they were literally looking for leaks because 17 of that phenomena.
That becomes significant later.
18 In other words, it didn't calculate it to a 1 gpm 19 limit and say:
We are above that and therefore we have to 20 enter some kind of an action statement but it did give them 21 an indication that they are having some kind of real leakage 22 problem.
23 When you combine that with the difficulty they 24 were having with leak rate tests a t the time it becomes 25 significan', but it.'s not a quantifiable number as the ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage MB3%6646
-20040.0
., 7RT-1202
%-)
1 operators used at the time.
2 MR. STIER:
If I might add just one additional 3
piece of information to that.
Mr. Russell is right.
The 4
sump collects leakage or drainage from sources other than the 5
For example, there was -- one of the 6
spikes on the sump collection figure that we referred to 7
before, one of those spikes was due to the fact.that they 8
were hosing down a piec" of equipment in containment and-that 9
collected'in the sump.
So you saw that on the printout.
10 However, it does give you an outer limit, an outer 11 boundary, a conservative measure of what is happening in
-3 12 containment.
As long as you can go back and identify sources (V
1 13 of collection in the sump other than reactor coolant system, 14 you can get a pretty clear understanding of what is happening 15 in containment.
And we did that.
When we had a spike we 16 would go out -- we would look in plant records to see if 17 something else was happening at the' time that would have led 18 to collection in the sump and we were able to eliminate some 19 of these periods of apparent high-leakage as being something 20 other than leakage from the reactor coolant system.
21 MR. COLE:
The Reg Guide 1.45 indicates that, you 22 know, you are supposed to be monitoring the sump.
And the 23 fact -- and it also sets up that if there is leakage from 24 other than the reactor coolant pressure boundary that goes in
("N 25 the sump, if you can identify it and you know how much it is, ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationside Coserage 800-33MM6
20040.0 BRT 1203 f3N) 1-then you could subtract it.
.That seems to us a very 2
practical and pragmatic way of taking a number that you get 3
that's.on the upper bound and gives you some provision to 4
correct any wrong -- say leakage from a building cooler of 5
some type.
It seems to us that when you get an alarm in the 6
building, per the reg guide, it seems a very prudent reaction 7
to assume it is the unidentified leakage as defined by the 8
reg guide until you can prove otherwise.
9 JUDGE KELLEY:
Anyone else?
Let me just state my 10 overall reaction to the discussion.
It seems to begin as a 11 discussion in which there was sharp disagreement on a 12 significant point.
As the discussion progressed, both ends
\\#
13 of the table agreed on various points.
Would it be possible 14 to distill what comes out of this discussion?
At what major 15 points -- perhaps we could glean-this from the testimony --
16 but what are the major points of disagreement on this, if you 17 can say?
18 MR. STIER:
Disagreement?
19 MR. RUSSELL:
I don't think there is any.
20 JUDGE KELLEY:
Or agreement.
21 MR. RUSSELL:
I don't believe there's a 22 disagreement.
Mr. Kirkpatrick, to the best of my knowledge, 23 and he can testify for himself, looking at it, did not go 24 back to identify other sources of water which leak into the 25 sump.
Therefore, a correlation to reactor coolant pressure O
ACE-FEDERAL REPORTERS, INC.
202 347J70)
Nationwide Cmerage 80rk33MM6
28040.0
(~iRT 1204 L)
I boundary unidentified leakage would appear to be not very 2
good.
3 That was done by MPR.
The Staff did review that.
4 We used their analysis because they had done it and the 5
difference appears to be one of emphasis.
I think MPR would 6
say that the s~ ump should have been used more.
You'should 7
have calculated the 1 gallon per minute limit and you should 8
have had an alarm on it.
All those are "should have" in the 9-past.
We put emphasis on the inventory balance and not on 10 the sump and the tech specs.
.It's Monday morning 11 quarterbacking.
I don't think there's any technical dispute 12 between the two parties or two sets of witnesses, since we
-~
13 are not parties.
14 MR. COLE:
I guess the only thing we were 15 concerned on, the testimony would imply, the prefiled 16 testimony of the NRC, that they make very flat statements 17 about the accuracy of the sump method.
That is the one thing 18 that we think really distorts the record on that subject.
19 JUDGE KELLEY:
Lacking this explanation that we 20 now have?
21 MR. COLE:
Absent this explanation, if you take 22 pages 10 and 11, the context, we think that's grossly 23 misleading.
24 JUDGE KELLEY:
But with the discussion we've had 25 this morning?
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 80tk33MM6
l 28040.0 BRT 1205 I
i U-1 MR. COLE:
I think that clarifies it.
2 JUDGE KELLEY:
Anything else?
Fine.
3 270:
Would you agree that if the amount of water 4
collected in.the sump is less than 1440 gallons per day, the-5 reactor operators may properly assume that unidentified RCS 6
leakage is less than I gpm?
7 MR. KIRKPATRICK:
Here again, I think if you did a 8
lot of analysis, you could probably come to this conclusion.
9 Based on the information that the operators readily had, I 10 think they.could really not make this conclusion.
11 JUDGE KELLEY:
Okay.
Any comment?
Mr. Stier?
I 12 MR. STIER:
Well, again, in terms of pressure
('/
13 boundary leakage -- we haven't done a calculation'of the 14 number -- but there is some number for sump collection that 15 would indicate, if you figured out the rate, that you did not 16 have leakage in excess of 1 gpm from the reactor coolant 17 pressure boundary.
Because you would have an outer limit 18 established by what goes into the sump.
It could be from 19 other sources, but you would know you weren't greater than 1 20 gpm.
However, you've got to recall the definitional problem 21 that we have.
22 JUDGE KELLEY:
Right.
23 MR. STIER:
Subject to that, I would agree.
24 JUDGE KELLEY:
The prior question went over that 25 as well as some others.
Okay.
ACE-FEDERAL REPORTERS, INC.
202-347-3 310 Nationwide Cos erage 800-336 # 46
20040.0 1206
(~'}RT v
1 On page 30 of your testimony you discuss the 2
effect of switching the makeup --
3 MR. MC BRIDE:
Excuse me, Judge Kelley,-there may 4
be a slight lack of clarification in the record.
At the very 5
end of Mr. Stier's answer, he said subject to that 6
clarification he would agree.
I wonder if you would ask him 7
what he is agreeing with so the record isn't left unclear?
8 JUDGE KELLEY:
Okay.
9 MR. STIER:
I would agree that, as I indicated, I 10 don't know whether the number is correct, the number that was 11 included in the question.
12 MR. MC BRIDE:
The number in the question is 1440, 7-V 13 which is 60 minutes times 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.
14 MR. STIER:
That is, I believe, calculated at room 15 temperature.
As we know, under the TMI-2 procedure, leakage 16 from the reactor coolant system, unidentified leakage, was 17 calculated at reactor coolant system temperature; therefore 18 that number would not be correct.
But it could be corrected 19 by the difference in density of water at the two 20 temperatures.
21 The point is, I agree that there is a number at
-22 which the rate of collection, you could be sure the rate of 23 collection in the sump would indicate that leakage from the 24 reactor coolant system in containment would not exceed 1 25 gallon a minute.
ACE-FEDERAL REPORTERS, INC.
202-347-37(x)
Nationside Coserage 8m336-646
28040.0 BRT 1207 (m
s_/
1 JUDGE KELLEY:
271.
2 MR. RUSSELL:
I think the definitional issue that 3
Mr. Stier referred to ought to reflect that unidentified 4
leakage to containment as is defined in Reg Guide 1.45 is 5
what we are talking about, and.not unidentified leakage as 6
defined in the tech specs, which includes intersystem leakage 7
to other closed systems.
8 JUDGE KELLEY:
Right.
9 MR. RUSSELL:
The other comment is that, in 10 calculating such a number, you would have to identify other 11 potential collection points in the compartment which may not 12 drain to the sump in addition to the correction for density, 13 bef. ore you would be willing to establish such a limit.
14 JUDGE KELLEY:
These are metered points that you 15 could just read a number off?
16 MR. RUSSELL:
Not necessarily.
There may be other 17 points in the compartment where it would not drain to the 18 sump.
19 JUDGE KELLEY:
Like just a pool of water just 20 sitting there?
21 MR. RUSSELL:
Water sitting on a floor, for 22 example.
They should be small quantities but depending on 23 how adequate the drainage is from all the collection points 24 in the compartment, it may not go to the sump.
25 JUDGE KELLEY:
I guess I was assuming the way we fxU ACE-FEDERAL REPORTER 4, INC.
202-347-37(R)
Nationwide Coverage E '16-6M6
28040.0 emBRT 1208 1
talked about drainage and sumps that everything in a nuclear 2
reactor containment is sort of tilted down; is that not 3
' rue?
4 MR. COLE:
In this particular case there was a S
real effort made to drain to that sump and trying to clean it i
6 up.
We used this sump because everything drains to it so 7
far.
We haven't found anything that would normally catch to 8
it.
So I don't know if he is speaking generically or if he's 9
speaking specifically with the design of this plant.
10 JUDGE KELLEY:
You are suggesting that this 11 particular sump really did catch everything?
12 MR. COLE:
There could be some-isolated little pO 13 piece, but as far as we can tell, it looks like it catches 14 everything, basically.
15 JUDGE KELLEY:
Anything else?
Okay.
16 271:
on page 30 of your testimony you discuss the 17 effect of switching the makeup level instrument during a leak 18 rate test.
Would you agree that there is no documentary 19 evidence that this ever occurred?
In other words, has the 20 NRC Staff's examination of the filed leak rate tests and the 21 makeup strip charts for TMI-2 identified a single test where 22 the strip chart indicates that the operator switched channels 23 while the test was in progress?
l 24 MR. KIRKPATRICK:
I have observed one instance on 25 a makeup strip chart of a level transmitter switching during r3 u
ACE-FEDERAL REPORTERS, INC.
202-347-3'00 Nationanic Coserage MG334646
.28040.0-1209 (es)BRT v
I a test.
I believe this was inadvertent.
That's the only 2
instance I'm aware of.
J'DGE KELLEY:
Can you tie that down?
Is it in 3
U 4
the record, this particular case you are talking about?
S MR. RUSSELL:
NRR report identifies it.
We can 6
identify the specific test.
We have many questions on makeup 7
level transmitter switching.
We could go into that now --
8 JUDGE KELLEY:
No, not particularly.
I just 9
thought since there was a single test referred to I wondered 10 which one it was.
11 MR. RUSSELL:
We'll be' covering that later on.
12 JUDGE KELLEY:
Yes?
13 MR. STIER:
I believe that that's our test number 14 108 that he is referring to.
But we will be talking about 15 that later on.
16 JUDGE KELLEY:
Okay.
17 272:
On page 31 of your testimony, you refer to 18 the effect of incorrectly reporting the RCDT voltage 19 reading.
Would you agree that there is no record from which 20 to document that this ever occurred?
21 MR, KIRKPATRICK:
Yes.
I would agree.
I don't 22 know of any such record.
23 JUDGE KELLEY:
Comment?
24 273:
Would you also agree that if the voltage 25 reading may have been improperly recorded by an operator on O
ACE-FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coverage 800 33MM6
20040.0
'rw BR T 1210
-]
1 one or more' occasions, there is no evidence (except perhaps 2
for Hartman) that such misrecording was willful, rather than 3
an honest mistake?
4 MR. KIRKPATRICK:
Yes.
I am not aware of any such 5
evidence.
6 JUDGE KELLEY:
274:
On page 32 of your testimony 7
you discuss whether water additions were required to be 8
recorded.
Would you agree that the addition of 100 gallons 9
of demineralized water to the makeup would have no mer.urable 10 effect on the boron concentration in the RCS?
11 MR. KIRKPATRICK:
No.
I would not agree.
12 Dilution by 100 gallons of demineralized water at this stage V
13-in core life would cause the removal of about 1 ppm of boron, 14 which would be represented by
.1 percent of the amount of 15 boron contained in the reactor.
This should have a 16 significant effect on reactivity.
17 JUDGE KELLEY:
You said ppm?
18 MR. KIRKPATRICK:
About a ppm; a part per t
19 million.
Excuse ~me.
20 JUDGE KELLEY:
I remember.
275.
21 JUDGE CARPENTER:
Did you say it would represent a 22 change of 0.1 percent?
That's the magnitude of the change?
23 MR. KIRKPATRICK:
No.
Of
.1 percent.
24 Excuse me, that's what I said.
25 JUDGE KELLEY:
If demineralized water were added (O
l ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 84xk33MM6
I l
28040.0.
l 1211 I
ks}BRT
,-~
r l'
to the makeup with no intent on the part cf the operator to n
2 alter the boron concentration of the reactor coolant system, 3
would it be reasonable for the operator to conclude that he 4
was not diluting the boron concentration within the meaning 5
of item 3.3.11 of AP 101, and thererere he was not required 6
to record the water addition?
7 MR. KIRKPATRICK:
No.
I don't believe this would 8
be a reasonable conclusion.
Regardless of his intent, any 9
normal addition of demineralized water, particul'arly at thiv 10 stage in core life, would alter the boron concentration i
11 significantly and, consequently, the reactivity.
The 12 addition should, therefore, be recorded.
13 JUDGE KELLEY:
Any comment?
14 MR. STIER:
No, your Honor.
15 JUDGE KELLEY:
Okay.
Let's go ahead with the i
16 other four questions that remain for Mr. Kirkpatrick.
17 Under the caption "TRI-2 interviewe, number 276:
1 18-Were you one of those who conducted interviews of control 9
19 room operators at.'TMI-2 in 1980?
20 MR. KIRKPATRICK:
Yes, I was.
21 JUDGE KELLEY:
277:
During those interviews, were 22 members of the TMI-2 D shif t shown documr:nts or told that r
23 seven out of eight leak rate tests where the NRC Staff found 24 hydrogen added during the test were performed on their 25 shift?
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage
- xt336-Are
I I
28040.0 T'7RT 1212 V.
1 MR. KIRKPATRICK:
During these interviews I 2
interviewed only one of the D shift operators.
That was j
3 Mr. Coleman.
My notes ir.dicate that this operator was not 4-stown any documents relating to hydrogen additions.
5 MR. MC BRIDE:
Judge Kelley, I don't want to risk 6
the Board's wrath, but I think the witness just answered half 7
the question.
It says were they shown documents or told 8
that.
I wonder if you could just pin that down.
9 JUDGE KELLEY:
You responded -- I wanted to ask 10 I you somet:iing slightly different.
I'll ask that too.
- But, 11 was there a team of NRC people, of which you were one?
12 Mk. KIRKPATRICK:
That's right.
13 JUDGE KELLEY:
Do you know one way or the other 14 a out what other members of the team may have said in this 15 regard to the D shift people?
MR. KIRKPATRICK:
Yes.
At that time I identified 16-j lfromthecontrolroomoperator's log several instances, I 17 I
l ')
believe it was three, in which hydrogen was recorded during a 19 time period for which there was a surveillance test.
One of 20 the members of the team went through the strip charts and, 21 based on the times on those strip charts, he found those 1
fperiodsofthose surveillance tests for which hydrogen was 22 23 recorded in the control room leg and made photographs.
Some 24 of the operators were shown these strip charts for 25 surveillance tests that were done during their shift.
My
}
ACE-FEDERAL REPORTERS, INC.
Mdl? Et10 Nationwide Coserage
- n3%fM6
s 20040.0 5,_q BRT 1213
,- )
1 answer only pertained to Mr. Coleman as he was the only 2
operator on D shift that I interviewed and he was not shown i
3 one of these -- this piece of evidence.
There was a 4
discussion with Mr. Coleman regarding hydrogen additions, 5
however.
6 JUDGE KELLEY:
Okay.
Number 278:
Can you 7
identify today the seven tests to which you referred?
Or did 8
you refer to seven, specifically?
I'm not sure.
9 MR. KIRKPATRICK:
No, as I already indicated, I 10 believe at that time we only knew about, I think it was 11 three.
And'I couldn't identify them today.
12 JUDGE KELLEY:
Okay.
And, number 279:
Is it true O
~/
13 that, based on later analysis, the NRC Staff no longer 14 believes that hydrogen was added during those seven tests?
15 MR. KIRKPATRICK:
I wasn't party to those later 16 analyses and so I don't really know.
I don't believe they 17 could pertain to the records that I was aware of at that 18 early date because I don't believe we had that many at~that 19 time.
20 JUDGE KELLEY:
Mr. Russell, do you have any 21 comment as the later analyst of these tests?
Whether seven 22 or otherwise, I'm not sure.
23 MR. RUSSELL:
I don't recall the exact number but 24 I did review the records that were in the Region 1 office in 25 April of 1984 at the direction of the commission.
I believe ACE-FEDERAL REPORTERS, INC.
202 347 3700 Nationside Coserage 8(433MM6
28040.0
(~7RT 1214 N~)
I there were a number, I'd have to refer to my notes, on the 2
order of about 15.
3 Some of those the Staff later concluded that, 4
because of errors in the way the strip chart timing was done 5
'or for'other reasons, that the additions did not occur during 6
the. test.
There were some that we concluded were.
And we 7
can identify that later but I don't believe it's that 8
significant.
As later testimony will show, we had to revise 9
our own analysis many times, as a result of feedback from 10 interviews from operators reviewing work that had been done 11 by Dr. Chung and reviewing work that had been done even 12 earlier in April 1980.
g-
\\_/
13 There were, in my opinion, because of the time 14 that they had done it, significant errors in some of the 15 early work that was done in April of 1980 with respect to 16 identifying hydrogen additions from strip chart times to log 4
17 times, and because of timing errors and other factors.
We 18 just did not make an effort to try and carry along in time 19 the April 1980 identifications to the version 1 of Chung's 20 report which was a part of the grand jury effort, version 2 21 of the Chung report to the multiple versions, even, of the 22 NRR report.
We finally locked in, in July of 1985, what we 23 think is the best technical analysis of the various leak rate 24 tests.
25 JUDGE KELLEY:
Thank you.
We should go to ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-3W%m
28040.0 r-3RT 1215 V
1 follow-ups on the questions this morning, primarily to 2
Mr. Kirkpatrick, but of course there were other 3
participants.
Do we have follow-ups on those questions?
4 MR. MC BRIDE:
Judge Kelley, I don't.have any, 5
with the proviso that it may become important later in the 6
proceeding which of these tests the NRC investigators 7
initially believed had hydrogen added during them, they no 8
longer believe had hydrogen added during the test.
Therefore 9
I would just like to reserve -- I know Mr. Russell said he 10 didn't think it was that significant.
I suspect what he 11 means is it doesn't really matter if it's 10 or 1 or 12 something, but it may turn out to matter a great deal to one
~
13 or more individuals if one of the two tests that we are 14 talking about here, or 12 tests or whatever they be, is the 15 only one that is so-called questionable for that operator.
16 So I would like to reserve on the possibility of 17 asking Mr. Russell to tell us at a later time which of those 18 tests they no longer believe had hydrogen added during them.
19 JUDGE KELLEY:
You can reserve on the 20 possibility.
The Board -- since Mr. Russell is here in 21 Bethesda -- I suppose if there's a good reason to call him 22 back, we'll do it.
I don't want to leave it that the Board 23 is committed to calling back Mr. Russell, necessarily.
24 MR. MC BRIDE:
Then I would like to ask the Board
(^g 25 inquire of the Staff which of the tests they believed in 1980
\\_)
ACE-FEDERAL REPORTERS, INC.
202 347 3700 Nationwide Coserage MG336-6646
28040.0 s
r"7RT 1216 U
1 had hydrogen added during them, and identified those to D s
2 shift operators, they no longer believe had hydrogen added
.+
3 during those tests.
4 JUDGE KELLEY:
From your prior answer,
~
5 Mr. Russell, are you in a position to do that this morning?
6 MR. RUSSELL:
We can do that.
But for any tests 7
that we subsequently concluded did not involve 'a hydrogen 8
addition we did not zely on that. test and we, in' fact, 9
ignored it.
So it's not part'of the basis for NRR's 10
' recommendations for any operators.
We can, however, discuss 11 those earlier tests.
The ones we discarded, more or less.
12 I think the issue is really with respect to
.f 3
~
not myself.
Because I'm expecting 13 recalling Mr. Kirkpatrick, 14 to be here.for a while yet.
15 JUDGE KELLEY:
Right.
16 MR. MC BRIDE:
Judge K'elley, can I ex' plain to you 17 what the oroblem is?
It may not be significant for 18 Mr. Russell's conclusions or Mr. Capra's conclusions at the 19 present time, because of the nature of their reports, which 20 are much later in time than these April 1980 interviews we 21 are talking about, and, therefore, they relied on their own 22 conclusions.
23 But what happened here, and this was the very 24 opening gun of these various investigations was that the
(~}
25 operators -- and the notes of interviews and reports of those
\\J ACE-FEDERAL REPORTERS, INC.
202 347 3Nt)
Nationwide Coverage 800-336- % 56
28040.0 rs3RT 1217 U
1 interviews would show this; this is not my speculation but it 2
is in the record -- operators were brought into the trailer 3
and were told:
Look, it appears to us X out of Y tests on 4
your shift, leak rate tests, hydrogen was added and those X S
times were on your shift.
What do you think about that?
And 6
various operators were made to answer those questions.
They 7
didn't know which tests they were, when they occurred, who 8
did them or what ha e you, but those were the facts they were 9
presented with and they gave some testimony.
10 A lot of issues arose out of the addition of 11 hydrogen during leak rates because of the predicate to those 12 questions and those issues have carri'ed forward for six and a
,s 13 half years through some of these operators' testimony.
14 People in subsequent interviews keep going back to the 1S question of:
Well, why did you tell the NRC in April of 1980 16 that either hydrogen was or was not 'added during leak rate 17 tests on your shift, to whether or not Mr. Russell and 18 Mr. Capra now say, well, we are-not relying on those 19.
particular tests or those issues; those issues have a life of 20 their own in the testimony or certain of our clients.
That's 21 why it's significant to us.
22 JUDGE KELLEY:
That may very well be that when 23 your clients are on the stand, that that issue would come up, 24 as you have described it.
I guess I'm not clear that em 25 Mr. Kirkpatrick -- he says he talked to Mr. Coleman.
Maybe ss ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coserage 804336-6M6
28040.0 7'7RT 1218 e
\\_J 1
he'd come back and talk about Mr. Coleman, if that is 2
appropriate.
3 We are hung up over whether you have an absolute 4
right to' recall Mr. Kirkpatrick.
I'm not sure it's a real 5
issue but it's something that we are just reluctant to say 6
right now this morning:
Yes, you can do that.
7 MR. MC BRIDE:
Here is the problem, Judge Kelley.
8 When you get D shift on there, on the witness stand -- and 9
the question of their answers about the interviews in April 10 1980 may occur to the Board because you have directed them to 11 be aware of their prior statements -- you are going to see, 12 in some reports, that these operators were told.on seven out g-
-(/
13 of the eight tests that we've identified where hydrogen was 14 added during leak rate tests, it happened on your shift, or 15 such similar question.
And you are going to ask the 16 operators, perhaps:
Was hydrogen added during those tests?
17 And the operator is going to say:
I didn't know which tests i
18 they were talking about when they asked me that question.
19 JUDGE KELLEY:
And if we need Mr. Kirkpatrick to l
20 find out or shed more light, we'll call him back.
i 21 MR. MC BRIDE:
All right.
22 MR. BLAKE:
Judge Kelley, is that the end of l
23 Mr. Kirkpatrick?
24 JUDGE KELLEY:
It is going to be.
We asked about 25 follow-ups.
(~}
m-ACE-FEDERAL REPORTERS, INC.
m s>4>.
~
. c _,.
--r--
---,w----m-,-
--.i-.-w+4
~.
28040.0 f-'3RT 1219
.U i
1 MR. BLAKE:
I didn't know if it was just 2
follow-ups to this group.or it's follow-ups and 3
Mr. Kirkpatrick is going to be dismissed.
4 JUDGE KELLEY:
I believe we may have a question or 5
two from Mrs. Aamodt, one question for Mr. Kirkpatrick.
6 MR. BLAKE:
I would just ask Mr. Kirkpatrick not 7
be dismissed before there's a break and I have a short 8
opportunity to review some potential questions.
9 JUDGE KELLEY:
I was going to note that we were 10 asking for follow-ups initially on the questions immediately 11 p.eceding.
We did establish at the outset that there would 12 be two chances, so to speak, sort of a cleanup chance at the 7-U 13 end.
That may take a little longer for people to pull 14 together.
If you could be around after the break for a bit, 15 Mr. Kirkpatrick, for a possible question or two.
16 MR. KIRKPATRICK:
Okay.
17 JUDGE KELLEY:
I have a question of my own, 18 general follow-up question that I would like to put before 19 you go.
But apart from that, that is we had this one 20 question from Mrs. Aanodt, which I will now ask.
Then there 21 would be this general follow-up opportunity and then we would 22 envision letting Mr. Kirkpatrick go, pursuant to the 23 understanding he may be recalled.
We just don't know.
24 Excuse me a moment.
25 JUDGE BRIGHT:
You stated that if you added 100 O
ACE-FEDERAL REPORTERS, INC.
202-M7 3700 Nationwide Courage 80lk334 % 46
28040.0
/ '?RT 1220 U
1 gallons of nonborated water because of certain' reactivity 2
perturbations -- could you give me some idea of what this 3
might do, insofar as the power level is concerned?
What 4
would the perturbation be like that would be effected?
5 MR. KIRKPATRICK:
Let's see, I made an estimate 6
here of reactivity of about.01 percent, delta K over K.
7 What I would expect, normally the reactor is on automatic 8
control and what you would see is that the rods would go in 9
slightly to compensate for this boron addition -- excuse me, 10 for this boron removal from the core.
So that you shouldn't 11 see much in the way of power.
You might see a slight change 12 in the temperature, however.
,O-13 JUDGE BRIGHT:
I was just noting from yesterday a 14 small change in temperature made a big change in the 15 gallonage.
16 MR. KIRKPATRICK:
That's true.
I think you would 17 prol' ably see an effect on the leak rate test due to the 18 temperature change.
19 JUDGE BRIGHT:
Would you consider this to be one 20 of the better arguments tor not adding makeup water during i
21 leak rate tests?
22 MR. KIRKPATRICK:
That would certainly be an 23 argument.
In light of the 1rrors they had, though, the fact 24 that the water addition was not properly accounted for in the 25 calculation, I think the effect of the water addition itself, ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationaide Cmerage HUO-3%6M6
28040.0 yRT 1221 1
the miscalculation of the density, had a much larger effect 2
than the temper'ature.
The only time temperature should have 3
introduced an error was when they were'above 582 degrees 4
Fahrenheit.
5 JUDGE BRIGHT:
Mr. Russell?
6 MR. RUSSELL:
I would like to supplenent that.
7 The Staff has reviewed the logs and we find many times where 8
water additions were required for reactivity control reasons,
-9 such as Xenon transients following shutdown and startup, and 10 the amounts-of water were added and there were cases where 1 1~
there were dilutions going on and there were cases where 12 there were borations going on.
It is required that they fgb 13 maintain the controlling group within a certain defined limit 14 and there is a procedure.
15 We will be getting into this, I think, in fairly 16 significant detail because it was important for us to show 17 that there either was or was not an operational necessity for 10 some of the additions that may have occurred during leak rate 19 tests.
I think the answer that Mr. Kirkpatrick gave in 20 general is correct.
I did also review the' calculations.
It 21 is a simple ratio of vclume reduction for 100 gallons, use a i
22 mass ratio, mass of the reactor coolant system to the mass 23 added.
That's the amount of dilution you'll get in ppm.
The 24 typical number for ppm is about 1000 to 1040.
1 ppm change r^
25 for 100 gallons addition is about the right ball park.
m}
ACE-FEDERAL REPORTERS, INC.
202-347 37(X)
Natiotiwide Cos crage 8(xk3%ue
28040.0 r~SRT 1222 Y]
1
. JUDGE KELLEY:
I have a question for 2
Mr. Kirkpatrick from Mrs. Aamodt.
It reads as follows:
3 Whose responsibility was it for taking the reactor to hot 4
standby within six hours af ter a leak rate test in excess of S
technical specifications where the source of leakage could 6
not'be identified or reduced?
7 MR. MC BRIDE:
Judge Kelley, may I ask for a 8
slight modification of that question?
.She's omitted the 9
first step of the action statement, which is you have four 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> to run another test, look for leaks, attempt to 11 identify leaks -- then you have six hours to'go to hot 12 standby.
O 13 JUDGE KELLEY:
And we do have extensive record 14 testimony on that point.
So,. fine.
With that understanding 15 built in, could you respond?
16 MR. KIRKPATRICK:
Yes.
The general answer is that 17 every licensed operator who is on shift is responsible for la carrying out the conditions of the technical specifications.
19 JUDGE KELLEY:
Thank you.
20 I would like to put a sort of general follow-up 21 question which frankly I wish I had put yesterday afternoon 22 before Mr. Rockwell and Mr. Moore left.
But, Mr. Kirkpatrick 23 now is going to leave us, who has been speaking to test 24 validity extensively.
Perhaps one could infer the answer to 25 the question from all the tes timo'ay, but in my mind it is O
ACE-FEDERAL REPORTERS, INC.
M-347-17m Nationaide Cmerage
- 43 36-t646
28040.0 7' ORT 1223 V
1 sort of a bottom line and I'm not entirely sure of where the 2
witnesses are on it.
3 Namely:
Was the test at TMI, with all of its 4
problems, all of its inaccuracies, was it a reliable test if 5
administered properly, under steady state conditions and so 6
on?
Or was it so shot through with errors, built in 7
problems, that it was essentially arbitrary?
It was like 8
playing a slot machine and you didn't know whether you were 9
going to get a bell or a lemon or what.
And even when you 10 got a number under 1, of
.8, that looks good, the numbers you 11 might get the next go-around would be so erratic that the 12 seemingly valid number is called into question?
I just f3U 13 wonder whether an operator there in the control room just 14 pushing the button should have stuck to the test and should 15 have believed it or whether he might have been justified in 16 saying:
This thing just doesn't work and it is irrational; I 17 don't have to pay any attention to it?
18 Maybe we can take a break on the basis of the i
19 question and come back to your answer.
How is that?
Take 10 l
20 minutes?
21 MR. KIRKPATRICK:
That's okay.
22 MR. GOLDBERG:
Judge Kelley, I think it would be I
23 helpful if we had 15 minutes on this particular break because l
24 I want to discuss with Mr. McBride and Mr. Gephart the 25 proposed testimony of Mr. Russell, which may remove their ACE-FEDERAL REPORTERS, INC.
200 3:7 3700 m:;=.wi.?: Ces crage mn33MM6
28040.0 r'
1224
(.)9RT 1
objection.
Mr. Blake indicated he wants to' prepare his final 2
follow-up questions for Mr. Kirkpatrick, now that he is going 3
to be going soon.
4 JUDGE KELLEY:
Fine.
15 minutes.
5 (Recess.)
6 JUDGE KELLEY:
Okay.
We can pick up again.
7 Before we pick up on the pending question, counsel, I 8
understand, discussed the matter that arose right at the 9
beginning pertaining to the possibility of Mr. Russell saying 10 something further -- Mr. Goldberg?
11 MR. GOLDBERG:
Yes.
I requested that Mr. Russell 12 be allowed to provide some supplemental testimony in LJ 13 connection with his answers to questions 269 and 271.
I 14 apologize to Mr. McBride for not having the opportunity to 15 speak to him about that before I requested it of the board, 16 but after discussing it with Mr. McBride, I believe that he 17 is willing to withdraw his objection to that supplemental 18 testimony.
Of course he can speak for himself.
But, if so, 19 then I would ask that the Board allow Mr. Russell to amend 20 his testimony by supplementing it with respect to his answers 21 to 269 and 271, 22 MR. MC BRIDE:
Mr. Goldberg's understanding is 23 correct.
We do withdraw our objection.
24 JUDGE KELLEY:
Fine.
Mr. Russell?
25 MR. RUSSELL:
Let me start by rereading question ACE-FEDERAL REPORTERS, INC.
202-347-37(O Nationwide Cmcrage
$n336 6646
28040.0
/7RT 1225 Q) 1 269:
Were the administrative procedures applicable to TMI-2 2
included in the legal requirements imposed on the TMI-2 3
operators?
4 I answered that in a narrow sense of what was the 5
chain of responsibility to the individual operators.
I 6
believe it is also very significant and relevant that I 7
believe the operators did not understand that.
And I base 8
that upon my interviews of the operators, where I 9
specifically queried them on their understanding of 10 administrative _ procedures, their legal obligations, et 11 cetera.
12 This was a part of the proceeding that was.not gb 13 included in the Staff's report because it related to current 14 knowledge, current understanding.
But in the process of 15 getting information on current knowledge, current 16 understanding, it also became obvious that at the time of the 17 accident they didn't have that.
18 If I could put that in context, many operators 19 that I have talked to at many facilities, even today, 20 perceive that an NRC license is a technical qualification.
21 They have achieved some level of expertise that permits them 22 to operate the facility.
And in many cases they have not had 23 an appreciation of their responsibilities to the NRC, 24 independent of the facility licensee, to assure that_the 25 rules and regulations are followed.
I think that's very q
ACE-FEDERAL REPORTERS, INC.
202 347-3700 Nanonwide Coverage 800-3h6M6
q 28040.0'
/'UR T 1226 i
)
1 significant.
Because there was testimony by operators that 2
decisions about shutting down the reactor or meeting the 3
technical specification requirements would be made by 4
management, or their. shift supervisor.
Or that was someone 5
else's responsibility.
They would make recommendations but 6'
they did not understand their affirmative obligation to 7
assure that they were carried out.
So the response I gave
~
8 was a narrow one and did not address their understanding at 9
that time.
10 I felt that was rather significant and the record 11 ought to reflect that.
12 JUDGE KELLEY:
Let me be clear to the extent, if
\\_/
13 any, that these discussions of yours that you just described 14 are in the record, say in the form of copies of interviews 15 and so on.
16 MR. RUSSELL:
I believe that there may be some.
17 But for the most part it was in enclosure 2 to each 18 individual report that I forwarded.
The approach that I used 19 was that interviewing-individuals with respect to their 20 involvement in leak rate testing in September 1980 through 21 March--- September 1978 through March of 1979 was conducted 22 with Mr. Capra.
I intentionally used a different individual 23 to assist me in the. interviews of the operators on their 24 understandings of the terms and conditions of their license 25 and reaching conclusions as to whether they had a sufficient
(~}/
l u-ACE-FEDERAL REPORTERS, INC.
2tc 347-3700 Natiottwide Cos erage Wh346646
28040.0 RT 1227
(~]/
(-
1 understanding of the facility procedures that they would be 2
actually able to carry those.out.
So there was a two-part 3
approach that I used.
It was in this second portion that I 4
came to the understanding and conclusion that they really did 5
not have a firm understanding of the legal obligation that 6
they had under their license.
7 It is also true that that's a -- I believe, not a 8
real significant point because they did understand that the 9
company expected them to follow their administrative F
10 procedures.
11 JUDGE KELLEY:
I'm still not entirely clear on 12 where, other than in your testimony now, we would find this
,f-()'
13 documented.
I think, for example, of the numerous exhibits 14 appended to the OI report, come of which go back to 1980.
15 You are not referring to that material, I take it?
16 MR. RUSSELL:
I don't believe, unless we erred in 17 redacting the reports, I believe my testimony, now, is 18 probably the only place'that it would be in the record.
19 JUDGE KELLEY:
Thank you.
20 MR. GOLDBERG:
Judge Kelley, Mr. Russell conducted 21 a separate interview of then-licensed operators in connection 22 with an aspect of hic investigation which is not within the 23 scope of this proceeding, namely current performance.
24 JUDGE KELLEY:
I understand.
L
(~T 25 MR. GOLDBERG:
Those transcripts are not a part of i
(/
l i
ACE-FEDERAL REPORTERS, INC.
I 202-347-3XX)
Nationwide coserage
- 10-336-(M6
28040.0
('pRT 1228
\\_,1 1
this record and all the information in his report which was 2
related to current performance has been removed at the 3
request of the Board.
4 JUDGE KELLEY:
All right.
5 MR. RUSSELL:
I might also comment that in the 6
enclosures -- and the approach we used was essentially that 7
of an NRC examination of the operators; that is, we kept 8
track of the questions we asked them and the responses to the 9
questions and whether they were correct or not.
The 10 information on their understanding of their obligations in 11 1978-79, while they came out during those discussions with 12 the operators, there would not be a transcribed interview.
13 We did not use a court reporter.
In fact we allowed them.to 14 use the procedures and other information that they would 15 normally have in doing their job in the control room.
We 16 wanted to make it as close to actual, on-the-job application 17 as possible.
So it was more an NRC oral examination during 18 which this information came out.
19 JUDGE KELLEY:
- Okay, 20 MR. RUSSELL:
The second question I wanted to 21 respond to was question 271.
That question was:
Do you 22 agree with the suggestion that NRC should or might take 23 action against an operator found to have violated only an
'4 administrative procedure?
And I believe I responded to that i
f7 25 with text that indicated "maybe."
\\/
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationaide Cmcrage
- G 336446 r
20040.0 f^7RT 1229 O
1 I-wanted to further put that in context, because 2
there are some important aspects of deciding whether to take 3
action or not that were not discussed.
I essentially 4
indicated that, if it were willful and the individual had S
knowledge, that enforcement action may oe appropriate.
6 There are two other, I believe, important factors 7
that should be considered.
One is that action should only be 8
taken when an individual violates a legal obligation directly 9
imposed upon that person.
This will become significant 10 later.
I have a question from, I believe, Mrs. Aamodt, as to 11 why didn't we look at operators or individuals other than 12 those who were licensed.
The other aspect is that it has to f-13 be, today, a safety concern.
That is, that continued 14 participation of the individual could jeopardize continued 15 safe operation of the plant.
16 Those aspects are significant in deciding what 17 actions we would take and those aspects were not included in 18 my prior testimony.
And the reason I didn't include them was 19 because I felt they were beyond what I had understood the 20 proceeding was.
But they are clearly factors which I 21 considered in making my recommendations and in reaching 22 conclusions for the Staff as part of the Staff's 23 investigation.
24 JUDGE KELLEY:
Okay.
That brings us, I think, 25 then -- let's go back to the question I put to ACE-FEDERAL REPORTERS, INC.
2tC-347 37m Nationw nic Coserage M A U6 N.M
28040.0 r-GRT 1230
\\
i 1
Mr. Kirkpatrick, which is in the nature of general 2
follow-up.
3 Are there other follow-ups for Mr. Kirkpatrick at 4
this point?
I believe Mr. Blake indicated he might-have 5
something?
6 MR. BLAKE:
No.
7 JUDGE KELLEY:
No?
Okay.
Well, I won't repeat 8
the question about the reliability of the test.
I'll just 9
put it to you.
10 MR. KIRKPATRICK:
You really asked me two 11 questions.
The first one was 'with regard to the reliability 12 of the test.
Clearly all our testimony indicates that the 7-13 test was very unreliable.
There were big problems with it.
14 I dwelled a lot on the fact that increasing the 15 duration of the test would have increased its accuracy and I 16~
think this was clearly an option during some time periods, 17 although I'm not sure that the operators realized this.
18 Certainly their engineering staff that was reviewing these 19 tests should have realized this.
Dut, quite honestly, there 20 were times when there was significant identified leakage that 21 increasing the accuracy of the test by any means would simply 22 have made it more definite that you were bound to calculate a 23 leak rate in excess of the allowable limit.
So, I think the 24 answer is that it was a bad test and that you couldn't trust 25 the results and that the operators should have been ACE-FEDERAL REPORTERS, INC.
202 347 DO N ationakle Costrage m)M NA
28040.0 RT 1231 G.g 1
suspicious of the results.
2 The last thing you asked me though, was:
Did this 3
justify them ignoring this?
I tried to give a practical 4
answer.
A very important safety consideration, I think, is 5
the amount of leakage that you have.
As long as you have 6
water in that reactor and are able to shut it down you cannot 7
have a serious accident.
There were times when that leakage 8
did exceed the allowable limit for unidentified leakage.
The 9
operators and the plant staff had no idea what that was.
The 10 whole purpose in the test is to know where that leakage is 11 coming from and to take action if it comes from an unsafe 12 place.
That leakage, I think, exceeded the allowable limit n
'~
13 at the time of the accident.
I can't say that it caused the 14 accident.
We really don't kncw where that excess leakage was 15 coming from.
But I do believe that if the procedures had 16 been followed, that plant would have been shut down at the 17 time of the accident.
And certainly the operators are 10 responsible for ensuring that they are operating a safe 19 plant.
20 I think with respect to leakage at that time, they 21 had an unuare condition.
I don't think that we can be in a 22 position to say, in NRC, that an operator can violate an 23 important safety feature of his plant and not be responsible, 24 even though I recognize that they were in a very difficult 25 position when they did this.
ACE-FEDERAL REPORTERS, INC.
8 _. e.....
- n-
20040.0 G RT 1232 v
1 JUDGE KELLEY:
Thank you.
Just a comment.
I 2
didn't mean to imply by the way I phrased my question that 1 3
thought that ignoring the test was the appropriate response.
4 We do have an issue before us, ignoring it or throwing it 5
away -- we do have a question before us, what difficulties 6
did the operators have in conducting this test?
It seems to 7
me if the test had a lot of errors in it, couldn't produce a
reliable results, that that certainly goes to that issue 9
which is essentially in my mind a mitigation issue, at 10 least.
So that that was where I was coming from in the 11 question.
12 Mr. Stier, do you want to respond to this?
13 MR. STIER:
Let me see if I can answer your 14 question in a practical way, perhaps by describing a little I
15 bit of the perceptions of the operators as we understand them
)
16 from looking at test results and from having conducted the 17 investigation.
18 It is clear that tests were run continuously.
Not 19 back to back, necessarily, but sometimes two or three or more 20 tests might be run on a shift.
And for every test that was 21 flied, for every test where the result was below 1 gallen a i
22 minute, there might have been one or two or three or more 23 tests that were over 1 gallon a minute.
24 The test in their view, therefore, because of p
25 these widely varying results, was unreliable and, I think v
1 ACE-FEDERAL REPORTERS, INC.
202-M7.)h10 Nationwide Coterage Min 3tMM6
28040.0-r"7RT 1233 U
1 from our analysis of the test, that was clearly justified.
2 An unreliable test in the sense that there was no clearly 3
consistent connection between what was coming out as a test 4
result and other evidence that they were seeing, like 5
watching the trend on the makeup strip chart or knowing in 6
some general sense what was happening with sump collection by 7
counting the number of sump starts that would occur during 8
the shift that they were serving on.
And, so, the operators 9
would see patterns in those parameters by their observations 10 of what was happening in the plant, and these test results 11 would swing back and-forth and seemed to have no connection 12 to them.
O 13 However, it is not really fair to say that it was 14 a purely random generator of numbers.
That's not the case 15 either.
Depending on the conditions that existed at any 16 period, there would be some rough correlation between 17 patterns of test results and the amount of leakage that they 10 were experiencing.
19 For example, in some periods where there was not a 20 lot of drain tank collection and where the plant was running 21 at steady state conditions, you would get a good leak rate, 22 as they would call it, on a shift, by running one, two or 23 three tests and one would come out under 1 gpm.
So it 24 wouldn't be difficult, in terms of their behavior the way 25 they were administering the test, it wouldn't be difficult to ACE-FEDERAL REPORTERS, INC.
3,-.,
m_ mm_,.
m u-
f~
{
28040.0
( ORT 1234 us 1
get a good leak rate on their shift.
2 There were other times wher plant conditions were 3
not conducive to getting a good leak rate, such as in 4
January, December and early January 1978-79, and in March, 5
February and March of 1979, when you had high drain tank 6
collection; it was almost impossible to run a test and.not to 7
do something in the course of that test to cause the result 8
to change -- it was almost impossible to get a good leak rate 9
on a given shift.
And you can see references in the logs to 10 frustration that they were feeling.
We can't get a good leak 11 rate, we only have X number of hours to go.
Got to get a 12 good leak rate.
g3V 13 I'll give you some idea, though, of variability 14 which does show consistent -- some consistency in the 15 patterns of results.
On page 110 of volume-I of my report, 16 we have the tests that were run between October 16 and 17 October 18; that is the five tests that showed leakage, 18 unidentified leakane over 1 gpm, that were not discarded.
We 19 physically have those tests.
Presumably those tests were run 20 in a sequence and they weren't thrown away so we know what 21 successive tests yielded.
And those results are as follows, 22 and I'll drop off the last three digits after the decimal
.23 l point.
Starting on October 16, with test number 149, you had 24 results of:
2.5, 2.0, 1.7, 1.2 and 1.3.
25 So you don't have wide swings b;ck and forth.
You l
ACE-FEDERAL REPORTERS, INC.
202-347 37(U Nationside Cos crage 80lk33 M 646
I 28040.0 1235 V('"RT 1
do have a fairly consistent pattern.
I'm sure that that was 2
not always the case that they were that close.
But to say 3
that it was totally unconnected with what was happening in 4
the plant and what they were observing is not really true 5
either.
There were some indications that they would get from 6
looking at these results that would trigger some response on 7
their part.
8 For example, in this time period when they got 9
these results and they couldn't get a good leak rate, they 10 went out and looked for leakage.
They knew something was 11 wrong.
They knew that their ability -- that the odds had 12 changed somehow on getting a good leak rate and that, 13 perhaps, one reason for that was the plant might be leaking.
14 So you can see people going out and looking for 15-leaks and subtracting identified leakage and trying to do 16 something to compensate for whatever it was that was pulling 17 these results consistently over 1 gpm.
18 JUDGE KELLEY:
but just a follow-up question:
So 19 there were certain times when plant conditions were more 20 propitious to the use of this test.
Is this sophisticated 21 hindsight on your part or would the operators at the time 22 have known that this is a good time to get a leak rate test 23 or a bad time?
24 MR. STIER:
I'm not suggesting that there was ever 25 a time when it was better to be using this test than some ACE-FEDERAL REPORTERS, INC.
202-347 3hD Ndtionuide C0%ffdge kn336 N46
e 28040.0
ORT 1236 l
1 other time.
I don't think the test was reliable enough to be 2
used, the way it was administered, with the errors in it, at 3
any time.
All I'm saying is from their. point of view there 4
was some rough correlation between patterns of test results 5
and plant conditions.
They would look at'the results.
They 6
wouldn't'just look at the result of any given test, I guess, 7
the best way to say it.
They would look at whether you have 8
gotten a good leak rate within that 72-hour period,.and if 9
they were getting consistently bad results during that 10 72-hour period, they would say:
Hey, something is wrong 11 here.
We'd better do something.
We'd better go look.
12 JUDGE KELLFY:
As a rough indicator of trouble, NJ 13 not an accurate indicator of leakage?
14 MR. STIER:
That's correct.
15 MR. RUSSELL:
I would also like to respond I 16 believe in testimony yesterday I stated I felt the operators 17 would perceive the leak rate test as a random number 10 generator.
I wanted to put that into perspective.
It's not i
19 uncommon for parameters measured in the plant to have some I
20 variability in their result.
But it's usually a small 21 percentage of the number you are looking for.
22 In this case, oscillations of plus or minus half a 23 gallon a minute or a gallon a minute, when you are looking 24 for 1 gallon a minute, is essentially a 100 percent 25 oscillation.
In that context I believe the operators O
I 1
ACE-FEDERAL REPORTERS, INC.
202-347 3XX)
Nationwide Coscrage
- X A3.16-6M6
_ _ _ - - - _ _. _ _. _ - _... - - _ _.~
28040.0 1237
(~}RT
(_/
l 1
perceived this as essentially a meaningless test that was 2
giving them random results.
3 At the same time, it is important to look at each 4
test and how each test was conducted and whether there were 5
actions takea to improve the chances of getting a number 6
around 1.
This becomes particularly significant with high 7
leakage, where, because of the error in compensation for 8
leakage to the drain tank, you offset this and you would 9
start getting this variation around 1.5 ga.11ons a minute, j
10 which, if you get a plus or minus 1 oscillation around 1.5, 11 you are not going to have many tests that are going to come 12 out less than 1.
s
%)
13 If operators found that by adding water during a 14 test or hydrogen or some other action they could improve 15 their chances of getting one less than 1, then they -- and 16 they knew that their actions were precluded by the procedure, i
s 17 that is the kind of thing that you do see and I think that 10 requires looking at each test and the plant conditions at the 19 time to reach those conclusions.
r 20 JUDGE KELLEY:
Yes.
I'm not even at all into the 21 question of manipulation.
I'm not talking at all about 1
22 putting in hydrogen or doing anything, whatever anybody may 23 have done.
I am talking about how the test worked when you l
24 pushed the button.
That's all.
25 JUDGE CARPENTER:
In the context of Judge Kelley's ACE-FEDERAL REPORTERS, INC.
~.- c-...
n n-
28040.0 RT-1238 p/
\\
1 question about operator feeling about the test, can you see 2
how one could feel the test was not' producing valid results 3
and fail to check the deficiency box on the piece of paper on 4
which you signed your name?
I keep hearing this month after 5
month after month.
This is going on.
People signing.a piece 6
of paper and failing to check the appropriate box, if they 7
felt as you have just described they might feel.
That, to 8
me, is a fairly narrow issue but to me a trunk of the tree of 9
what we are here about.
10 Can you see how they might have felt that it 11 wasn't -- that it was appropriate not in the sense of 12 understanding the precise words on the license piece of 13 paper, but certainly within the procedures, general 14 procedures, that if you've got a problem, report it.
I'll 15 just put it that simple.
If you've got a problem, report 16 it.
17 MR. It OSS ELL :
I can describe what their testimony j
10 was and how they appeared to have rationalized it.
You have 19 to recall that they felt that the test was required to be 20 performed once per 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />, and if they ran the test more 21 frequently than that, that they were being conservative and 22 that those tests weren't required then by the technical 23 specifications, so they somehow rationalized that that was 24 not a requirement.
You didn't need to call that a 25 uurveillance test.
That's essentially the types of answers ACE-FEDERAL REPORTERS, INC.
m,.3,.
-m-
28040.0 pR T 1239 V
1 we got.
2 It is true that they got a cover sheet that would 3
go with the test procedure that did have places for them to 4
check off whether it was a deficiency or not.
In this case I 5
think more appropriate would be an exception because of 6
problems either in the procedure or in plant conditions that' i
7 made it difficult to get an acceptable result.
But for some i.
8 reason, for this test they did not consider it one for.which 9
exceptions and deficiencies applied.
For other surveillance 10 tests they felt that was the case.
11 JUDGE CARPENTER:
That's the thrust of my 12 question.
Ilow could that be?
c 13 MR. RUSSELL:
I believe that's something you are 14 going to have to ask the operators because I'm relaying my 15 reaction to their testimony to me on these issues.
It was a 16 bi-stable.
In some cases they applied it.
The logic didn't 17 follow in other cases.
18 MR. CAPRA:
Let me supplement that a little bit.
19 In our interviews with the operators, as Mr. Stier testified 20 yesterday, his interviews apparently were consistent with i
21 ours, we did not have any operator that testified that they 22 applied the exception and deficiencies to the leak rate 23 surveillance procedure.
Ilowever, we did have testimony to i
24 varying degrees that operatora did think, at certain times, 25 som(one was looking at and taking nome action to try to ace-FEDERAL REPORTERS, INC.
I 202 347-)?ul Ne nmmkle rmerage An)4 Ne
]
i 28040.0 psRT 1240 V
1 correct problems in the procedure.
2 Many of the operators thought that the problems 3
with the procedure itself was in the computer, and, 4
therefore, they believed that the computer people were 5
workiijg on the problem.
Now, the testimony varies with to 6
what degree they thought action was being taken, but by and 7
large they didn't think that nothing was happening, i
8 MR. RUSSELL:
I might supplement that with one 9
other obuervat' ion.
As I recall, a deficiency was not i
10 completed for the licensee event report.
Recall I indicated 11 the report --
1 12 JUDGE KELLEY:
I'm a littic bit concerned about O'
i 13 how far we are getting away from what we were talking about.
1 14 These various problems are all goina to come out.
15 MR. RUSSELL:
That's correct.
They will come out 16 later.
17 JUDGE KELLEY:
The pending question was whether 10 the test was any good.
Has everybody spoken to that?
okay.
19 Any other Collow-ups?
20 MR. BLAKE:
I feel constrained at this point to 21 make an observation, not in the nature of an objection or a 22 notion fo strike.
The observation goes to Mr. Kirkpatrick'n 23 reference = to the potential for a link between this leak rate i
24 tout and the accident at TMI-2.
i 25 It la not, from my reading of the order that l
l Aci!-FEDliRAL RiiPORTIIRS, INC.
muo.,
s.
.,a cmm,,
..m u.
20040.0 f'}RT 1241 v
1 precipitated this proceeding nor anything in our discussions 2
to date, nor anything in the reports that I am familiar with, 3
that it is the intention of the board to get into that 4
connection or into that area.
To the extent we are, my 5
estimates about the length of this proceeding would increase 6
beyond the pessimistic views that I expressed earlier.
7 But it my view is different than the Board's, then 0
if we are -- as I say, that was a fairly gratuitous comment 9
in the context of a general question that Mr. Kirkpatrick 10 answered for the Board -- but if that is a part of the 11 proceeding, which in my view is incorrect, my belief, in that 12 regard -- then I guess I would like to hear that.
I would O'
13 like to know that.
Because I view that as beyond what it is 14 we are here to talk about.
15 JUDGE KELLEY:
I understood you, Mr. Blake.
We 16 really ought to have another mike down on your table.
It was 17 a little difficult to understand you in this instance.
18 Mr. Goldberg, do you want to comment?
19 MR. GOLDDERG:
I think Mr. Blake has raised a 20 valid point.
I think what Mr. Kirkpatrick was appropilately 21 doing was trying to give a complete answer to the question 22 that was posed to him and to reflect his answer to that 23 question -- his opinion.
We are not representing or 24 presenting to the Doard in this proceeding any NRC position 2S with respect to whether or not there in any connection ace-FEDERAL REPORTERS, INC.
x mn.,
smm co,,m,,
mwm
28040.0 p HT 1242 V
1 between the leak rate testing issues and the accident.
I 2
think Mr. Kirkpatrick's answer should be understood simply as 3
his answer to the question that was posed to him, and his 4
attempt to completely state his opinion in response to that 5
answer.
6 JUDGE KELLEY:
I think I can speak for my 7
colleagues, saying that the Board does not understand that 8
any issue about the relationship between leak rate ~ issues 9
that we are looking at and the accident -- that there is any
-r 10 such issue in this case.
That was your observation --
11 question, essentially?
12 MR. BLAKE:
Yes, sir,'
O 13 JUDGE KELLEY:
I think -- is that our reaction?
14 That's our understanding also.
I don't know if it 15 Mr. Kirkpatrick wants to make ary comment.
>It's not 16 necessary, 17 MR. KIRKPATRICK:
Well, that commenta ws only in 10 the context of how important I felt that lea Frat ;t;es t was
/,'*
'1 knowledge of how much leakage act}ially ei,<isted was.
(.
19 and the 20 I don't believe there is a direct connection between that e
l'
/
,g 21 leakage -- or I haven'ts -- have no evidence that,thee is a 22 direct connection between that leakage and the accident.
,r '
/
23 Only to note that, I beljeve at the time of the accident, it 24 did exceed the allowable limit.
g p
25 JUDGE KELLEY:
Mr. Blake, in that satisfactory O
ACE-FEDERAL REPORTERS, INC.
x m.nm soe. a. no m,,
yo n-
20040.0
('SRT 1243 L]
1 discussion from your standpoint?
2 MR. BLAKE:
Yes, it is.
I felt that 3
Mr. Kirkpatrick's response, frankly, to the Board's question, 4
was sufficient, even without that adder.
But I think we were 5
headed in that direction.
I had a different view and I 6
wanted to know where we were.
Thank you.
7 JUDGE KELLEY:
Are there any further questions for 8
Mr. Kirkpatrick?
9 MR. MC DRIDE:
No, your Honor.
10 JUDGE KELLEY:
Okay.
Mr. Kirkpatrick, as you have 11 heard, there is a possibility we might want to call you back 12 later for some narrow point, but apart from that the Board g-
\\'~/
13 wants to thank you for your participation and your very 14 considerable work on the issues and your role in this very 15 useful dialogue that occurred.
Thank you very much.
You are 16 excused.
17 MR. KIRKPATRICK:
Thank you, your Honor.
18 JUDGE KELLEY:
Off the record for a moment.
19 (Discussion off the record.)
20 JUDGE CARPENTER:
Returning to the next question, t
21 the next question, makeup level transmitter switching.
22 Question 272 in the Russell /Capra sequence.
i 23 Is there any instruction or guidance in any TMI-2 24 technical specification or administrative procedure that 25 required TM1-2 operators to use one level transmitter channel
(
ACE-FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coverage
- Ek 3366M6
28040.0
.r'3RT 1244 N-]
1 in preference to the other?
2 MR. CAPRA:
No, sir.
We are not aware of any 3
guidance or requirement that specified a preferred level 4
transmitter.
However, once an instrument is selected it is 5
to be used for both the initial and final data readings.
6 Section 3.4 of surveillance procedure 2301-3D1 requires, "the 7
same sources should be used when recording initial and final 8
reactor coolant system temperature, pressurizer level, makeup 9
level and reactor coolant drain tank level.
Different 10 sources" -- I'm sorry -
" differences in sources could be 11 misinterpreted as reactor coolant system leakage when 12 comparing successive readings."
73 V
13 MR. STIER:
I agree with Mr. Capra.
14 JUDGE CARPENTER:
273:
Is there any instruction 15 or guidance in any TMI-2 technical specification or 16 administrative procedure that cautioned TMI-2 operators not 17 to select an unstable leve.l' transmitter channel for use in 18 performing a leak rate test?
19 MR. CAPRA:
I believe the answer to that question 20 is yes.
Administrative procedure 1036, entitled " instrument 21 out of service control," describes the method of ccatrol of 22 readout devices which become inoperable or are strongly 23 suspected of being inoperable.
It requires that they be 24 marked, documented and controlled until repaired.
25 Administrative procedure 1036 was applied to f]
J ACE-FEDERAL REPORTERS, INC.
202-347-3X10 Nationwide Coserage 800-33MM6
28040.0
/' '72RT 1245 b
1 makeup level indication.
In part the procedure calls for 2
inoperative or suspect instrumentation to be marked with a 3
yellow "out of service" s. ticker to ensure that such 4
instrumentation is not zelied upon by operators until the 5
instrumentation has been repaired, calibrated and returned to 6
service."
There is, however, no specific -- that procedure, 7
administrative procedure 1036, is a general procedure for the 8
control of out-of-service instrumentation.
If you look at 9
the question from the point of view:
Is there something in 10 the leak rate test procedure itself, there is nothing that 11 cautions the operator about that.
12 JUDGE CARPENTER:
Mr. Stier?
('3>
13 MR. STIER:
I think that technical specifications 14 3462 and 4462 make it clear that you've got to use 15 instruments that will tell you whether you are within 16 limiting conditions for operation.
The whole purpose of the 17 leak rate test as-it is defined in those technical 18 specifications is to demonstrate that you are within limiting 19 conditions for operation.
If you can't determine accurately 20 what the makeup level is because you have an instrument that 21 is malfunctioning, how in the world can you demonstrate by a 22 test result that you are under 1 gallon a minute of 23 unidentified leakage?
So I,think this technical 24 specification does impose an obligation to use
(]
25 instrumentation that you believe to be accurate.
V ACE-FEDERAL REPORTERS, INC.
M-347-3700 Nationwide Coverage 800-336-6646
~ - _ -
20040.0
/~'7RT 1246 b
i 1
JUDGE CARPENTER:
But not explicitly?
2 MR, STIER:
It doesn't say:
Don't use a level 3
indicator that's inaccurate.
That's correct.
4 JUDGE CARPENTER:
I agree with your comment.
It 5
violates common sense to use an unstable instrument.
- Still, 6
there's no specific instruction.
7 MR. STIER:
That's correct.
8 JUDGE CARPENTER:
Thank you.
274:
One level 9
transmitter channel, referred to as LT-1, has been identified 10 as having been erratic or unstable for extended periods of 11 time.
During what periods, if any, was LT-1 tagged 12 out-of-service?
7s\\-)
13-MR. CAPRA:
If you'll bear with me on this I've 14 got essentially a two-part answer.
The reason for that is, 15 in originally performing our analysis, we did not obtain the 16 instrument out-of-service l'og required by administrative 17 procedure 1036.
Thus, originally in our analysis we did not 18 establish.the periods of time the makeup level instruments 19 were actually tagged out-of-service.
20 However, from our review of the makeup strip 21 charts and control room log during the six-month period
~2 2 evaluated, there were two periods of time where significant 23 makeup level instrumentation problems were experienced by the t
24 operators.
Between October 30, 1978 and November 6,
- 1978,
/~3 25 both level transmitters were taken out of. service at
(
\\,s' I
r ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 804336-6M6
20040.0
(^9RT 1247 V
1 different times for maintenance work and recalibrated.
.2 Between December'1, '78 and October 11, 1979 --
3 MR. GOLDBERG:
January.
4 MR. CAPRA:
I'm sorry.
-- January 11, 1979, the 5
output'of LT-1 became very erratic and unreliable.
However 6
it is not clear exactly what maintenance was performed during 7
this period.
At.least two maintenance work requests were' 8
submitted during this period.
One, as a result of level ~
9 oscillations, was submitted.on December 5, 1978 and the 10 other, due to LT-1 reading off scale high on December 25, 11 1978.
12 The work request for level oscillations indicates f3 V
13 that the level transmitter testing was completed and released 14.
for normal use on January 11, 1979.
Copies of the two work 15 requests are contained in the Stier report at volume V-A, 16 document tab 2.
In order to respond to the second part of 17 the response --
18 MR. GOLDBERG:
Excuse me for interrupting my own 19 witness, but he may have misspoken.
He said there was an 20 off-scale high reading from LT-1 on December 5, 1978.
I'm 21 wondering if he can check and be certain if that's the 22 correct date.
I'm sorry --
23 MR. CAPRA:
The correct date for the second one, 24 if I did misspeak, was Christmas Day, December 25th.
-25 The second part of the answer was in order to ACE-FEDERAL REPORTERS, INC.
Ac-347-3700 Nationwide Coverage 80fk336-6M6
28040.0
(~$RT 1248
\\-)
I respond specifically to this question the day before 2
yesterday.we did obtain a copy of.the instrument 3
out-of-service log.
While not all entries are legible on the 4
log itself, there are three periods of time identified in the 5
log that makeup level transmitter designated LT-1 was 6
specifically identified as being out of service..These 7
periods are as follows:
Between 1100 on December 5, 1978~and 8
1825 on December 8, 1978, that four-day period.
9 The next period was 0915 on December 16, 1978, to 10 1335 on December 20, 1978.
That's a five-day period.
11 The next period was from 2000 on January 5 of 1979 12 on sometime on January 10, 1979.
The specific time period is 7s0 13 not recorded in the log.
That's approximately a six-day 14 period.
15 Now, during this particular -- during the two 16 periods that I have mentioned, which is, first period, f
17 October 30, '78 through November 6,.1978, there were eight
(
i 18-leak rate. tests that were retained.
Out of those eight i
19 tests, two of them from our analysis appear normal.
There L
20 are four tests where the difference in reading between level j
l 21 transmitters during those tests varied from 13 to 15 inches l
22 and there were two leak rate tests in which the level 23 transmitter seen by the computer was essentially reading off 24 scale high at 9-plus inches.
25 During the longer period, which is December 1,
'78 ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 8m3%6M6
I 28040.0
/~ ~3R T 1249 U
1 through January 11, 1979, there were 50 leak rate tests that 2
were retained.
We believe that 16 of those tests were run 3
with the computer seeing the good-level transmitter and that 4
34 of the tests were run with the computer seeing the bad 5
level transmitter or 68 percent of the tests were run on the 6
unreliable level transmitter.
7 During the specific periods that I mentioned where 8
the out-of-service log indicates that the instrument, LT-1 9
was physically tagged out of service, such that it would have 10 a yellow sticker on the selector switch in the control room, 11 there were 16 tests run.
During that period of time 10 of.
12 the tests were run on the bad level transmitter.
7s(')
13 JUDGE CARPENTER:
Thank you.
Mr. Stier, do you 14 have anything to add?
15 MR. STIER:
Well, on the narrow question, I agree
.16 with the answer given by Mr. Capra.
If you would like me to, 17 however, since the next question deals with the propriety of 18 using that particular instrument, I can give you a more 19 general answer that would cover the same subject matter that 20 Mr. Capra's answer covered; tell you a little bit about our 21 observations of the malfunctioning of this particular level 22 transmitter.
23 JUDGE CARPENTER:
Well, the questions are l
L l
24 different becauce his next question is a very narrow one 25 about was LT-1 ever repaired.
Perhaps it would be better to u(~T I
ACE-FEDERAL REPORTERS, INC.
-ma>
s.e_m c_.,
sa -
28040.0 r ORT 1250 i
LJ 1
hold?
2 MR. STIER:
I don't know that we are going to get 3
into another~ question that would deal with the subject.
It 4
deals with the evidence that_was available to the operators 5
that this level transmitter was malfunct'ioning, and I can 6
cite you to the references in our report that deals with that 7
but I didn't want to start on a long answer without knowing 8
whether you wanted me to discuss it.
9 JUDGE CARPENTER:
Okay.
10 MR. STIER:
There were two periods that were very 11-close to one another, when we saw malfunctions in the - -in 12 one of the makeup level transmitters.
The first period where i.J 13 one of them was reading off-scale high, that is right at the 14 top of the strip chart, is covered in.our report, volume I, 15 pages -- pages 72 to 73.
26 We say in' that section that six out of nine tests 17 that were run during that period were performed using that 18 makeup level transmitter, which at times was reading off 19 scale, and there is a footnote at that point to indicate 20 which tests we are talking about.
21 The more important period, however, is the period 22 that runs from December 4th through January 10th.
That 23 period was very carefully analyzed because one of the level 24 transmitters, LT-1, was behaving in a very erratic way.
It
('}
25 was bouncing up and down, showing a very wide -- wide v
s ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6646
28040.0
~'7RT 1251 (O
1 variations in level-over very short perio'ds of time.
90, 100 2
gallons within a matter of minutes would change in the 3
indicated level from that makeup level transmitter.
That 4
period is discussed in our report, volume I, pages 73 through 5
79.
In that section we have figures that are reproduced 6
sections of the strip chart that compare the readings from 7
the unreliable level transmitter with the one that is more 8
reliable.
You can see the difference in the trace.
9 We wanted to know how many tests were run using 10 the bad level transmitter during that period of time and 11 wanted to find evidence to indicate whether people who were i
[~.
12 in the control room running those tests had reason to believe b
13 that they were using a level. transmitter that was 14 inaccurate.
The reason why that level tiansmitter, of 15 course, shouldn't be used to run a leak rate test is because 16 it makes the result unrelated to actual conditions that were 17 occurring during the test.
It has to do strictly with where 18 you start and where you end up on these wide swings in the 19 trace on the strip chart.
It has nothing to do with reality, 20 with the reality of the conditions in the plant, and 21 therefore would give you a result that was totally 22 unconnected with actual leakage.
23 MPR did an analysis of all of the strip charts 24 during that period of time, examined them very closely to try
(~N, 25 to reconstruct which level transmitter was on the strip' chart V
ACE-FEDERAL REPORTERS, INC.
202-347-3RU Nationwide Cos erage 80lk33MM6
28040.0
/7RT 1252 V
1 at any given time.
Conversely, that would tell you which 2
level transmitter was hooked up to the computer, and 3
therefore used to run a leak rate test at any given time.
4 The result of their analysis is depicted in figure 5
V-5 in volume IV-A of our report.
And we show you in that 6
figure the entire period from the beginning of January 7
through -- strike that -- from the beginning of December 8
lthroughJanuary15th, and indicate during which periods MPR's 9
best judgment is that one or the other of those level 10 transmitters is on the computer and on the strip chart and 11 which test was performed using which level transmitter.
12 In my analysis of the responsibility of each g-)
\\/
13 operator for running a given test, I looked at MPR's 14 analysis, their best judgment.
I looked at the strip charts 15 myself.
And I was somewhat more conservative than they were 16 and I used what I felt was my best judgment plus the 17 information that I.got.from MPR when I said that a particular 18 test was run using a bad level transmitter.
As I say, it was 19 somewhat more than MPR's best judgment.
It was MPR's best 20 judgment plus some measure of conservatism.
Because there is 4
21
-- in certain periods there is some question about how
.22 erratic LT-1 was.
23 There were efforts to fix it during this period of 24 time.
So, sometimes you can't quite tell how erratic that
(~}
25 level transmitter was.
Generally -- of course we don't know Ne ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800 336 4 46
28040.0 7"7RT 1253 U
1 what the trace looked like from LT-1 when it was hooked up to 2
the computer.
All we know is what the trace from LT-2 looked 3
like.
4 We also know the beginning and ending makeup 5
levels for tests run during that period.
And, so, when you 6
match.up those levels, beginning and the end of the test, 7
taken from the test printout, when you try to match that up 8
with the strip chart, I felt that if MPR said that LT-1 was 9
inaccurate and the beginning and ending points of the test 10 were substantially different from the strip chart trace, I 11 felt that that was an indication that LT-1 at that time was 12 behaving in an erratic way and so I felt that that test was 7s
(
}
13 run on a bad level transmitter.
So you are going to see some 14 slight differences between the footnotes that I-have in the 15 sections dealing with individual responsibility, tests that I 16 characterized as using the inaccurate level transmitter, and 17 those that MPR felt in their best judgment were run on the 18 inaccurate transmitter.
19 I felt that the evidence was strong, very strong, 20 that the people in the control room knew that LT-1 was 21 behaving in an erratic way.
There were work requests that 22 were issued.
There were three work requests that we felt 23 were significant -- December 5, December 25, and January 10 24
-- for the repair of that level transmitter.
We saw
(~T 25 references in shift turnover notes to malfunctioning level v
ACE-FEDERAL REPORTERS, INC.
202-347-37(X)
Nationwide Coverage 804336-6M6
28040.0
/~1RT 1254 b
1 transmitter.
And the testimony of the operators indicated to 2
us that'it was well known in the operations department that 3
that leve1' transmitter was malfunctioning.
4 Given all of that, we examined each test and made 5
some judgment about whether that test was used -- was run 6
using the bad level transmitter, and then summarized the 7
results, both in the body of volume I of our report, in the 8
sections that I have indicated to you, and with respect to 9
each operator or shift foreman in the section of the report 10 that deals with individual responsibility for all of those 11 individuals who were within'the scope of our investigation.
12 JUDGE CARPENTER:
If I may ask the follow-up 7-)
V 13 question, you referred to page 73, volume I of your report.
14 Do you have that before you?
15 MR. STIER:
Yes, I do.
16 JUDGE CARPENTER:
The first full paragraph on page 17 73 states, with reference to this malfunctioning of 18 transmitters, that another malfunction developed in one of 19 the makeup level transmitters.
It began to behave 20 erratically.
Does that mean that it fluctuated, or that it 21 just wandered, willy-nilly?
Did it oscillate up and down?
22 MR. STIER: 'It oscillated wildly up and down.
On 23 the next page is figure 1, page 74 is figure 1.
You can see l
24 a trace that shows how that level transmitter was r~T 25 f u'ic ti oni ng, and clearly has no connection with reality.
l u l
ll-l ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 8m336-6M6
20040.0 r~9RT 1255 U
1 JUDGE CARPENTER:
You then go on to say that, I'm 2
paraphrasing, because of these fluctuations that the outcome 3
of a test was left to pure chance.
4 Then the next sentence is what I'm interested in.
5 It says:
"Therefore, if the true identified leakage were 6
high, the use of the inaccurate level transmitter would 7
increase the probability of calculated unidentified leakage 8
being under 1 gallon per minute."
I don't quite see why.
It 9
seems to me analogous to the plant instability oscillations.
10 I don't see that it moves it one way or the other.
.11 JUDGE KELLEY:
How about telling us why after a 12 cup of coffee.
73U 13 (Recess. )
14 JUDGE KELLEY:
Back on the record.
Before we get 15 to the answer to the pending question, Mr. Stier had a 16 thought about sequence of testimony which we'd be happy to 17 hear.
Why don't you state it.
18 MR. STIER:
Well, of course, I have no idea what
'19 subject matter of the questions were that were directed 20 st.rictly to Mr. Russell, but I do have some idea, of course, 21 what the subject matter is of the questions directed to me 22 and to the extent that those questions deal with a test and 23 call for a comparison of our conclusions with those of 24 Mr. Russell, it would seem to me to be advantageous for us to 25 be here at the same time, responding to questions about ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage R&336-6M6
28040.0
(")JRT 1256 1
specific tests, so that if our views differ, you have an 2
opportunity to see where they differ and ask follow-up 3
questions to try to narrow those differences and to sharpen 4
what the issue is.
Otherwise -- you know, it may be a moot 5
point if there aren't any such questions.
6 JUDGE KELLEY:
Well, for example, we did make a 7
specific reference to their review of your work and your 8
review of their work.
Why don't you talk about it and see if 9
there are real differences?
That kind of questioning seemed 10 to be appropriate panel-type discussion.
We really haven't 11 specifically spoken to this but I think_you are suggestin'g a 12 similar thing.
7)-
(-
13 MR. MC BRIDE:
Did we intend this to be on the 14 record?
15 JUDGE KELLEY:
By agreement we were just leaving 16 it off.
17 MR. MC BRIDE:
I thought you said off the record.
18 JUDGE KELLEY:
Mr. Stier was going to make a 19 suggestion to me and I said why don't we put it on the 20 record.
It doesn't have to be on the record, per se -- I 21 think we could leave this off the record.
I'm not sure how 22 long it's going on.
Go ahead.
23 (Discussion off the record.)
24 JUDGE KELLEY:
We are on the record.
We had a 25 pending question, I believe.
O(~%
ACE-FEDERAL REPORTERS, INC.
202-347-3NU Nationwide Coserage 2336-(M6
28040.0
(lRT 1257 V
1 MR. STIER:
Judge, I believe the pending question 2
deals with the last sentence in the first full paragraph on 3
page 73 of volume I of my report which states, "Therefore, if 4
the true unidentified leakage were high, the use of the 5
inaccurate level transnitter would increase the probability 6
of calculated unidentified leakage being under 1 gpm.
I.
7 might add that'that in fact was the case in~ January of 1979; 8
that is, it is our judgment, based on_MPR's analysis of 9
leakage, actual leakage in the plant, that roughly on January 10 3rd, leakage exceeded 1 gpm, unidentified leakage exceeded 1 11 gpm, and remained at that level until January 15th.
12 During that time, of course, it became very O-13 difficult in running leak rate tests using the good makeup 14 level transmitter to obtain unidentified leakage under 1 gpm, 15 for all the reasons that we discussed previously, even though 16 there was a degree of variability in the results.
The 17 center, the mean of that range was at or above 1 gpm, and on 18 a regular basis, running tests using the good makeup level 19 transmitter, LT-2, results were coming out over 1 gpm.
20 Now, looking at the trace, which is made by both 21 level transmitters that is depicted in figure 7 -- I'm sorry, 22 tryi.ng that -- figure 2 on page 77, I think that's a better 23 figure to illustrate this point.
You will see on'that figure 24 a relatively smooth trace which we belle /e was made by LT-2.
25 Then during the period marked as "the test" the level U
-ACE-FEDERAL REPORTERS, INC.
202 347-1700 Nationwide Coserage 800 336-6M6
28040.0 r~7RT 1258 O
1 transmitter is switched to LT-1.
And you can see that there 2
are fluctuations in level indication of perhaps 300 gallons 3
every three to five minutes; very jagged and large changes in 4
indicated makeup level, then when the test period is over the
.5 level transmitter is switched back to LT-2 and the trace 6
smooths out again.
Of course, this is the transmitter that 7
is reading out on the strip chart, therefore the transmitter 8
that's hooked up to the computer to run leak rate tests was 9
the opposite transmitter.
So that this test would have been 10 run on the good level transmitter.
11 Now, depending on where you start that test and 12 because the level transmitter would be attached'to the 13 computer, the operator wouldn't know at what point on the --
14 on this wildly swinging level indication he was starting the 15 test, but by pure chance -- if he happened to start at a low 16 point and ended up with the final readings taken on a high 17 point, the indicated leakage would be very low, perhaps a 18 very, very low, or high negative number.
Minus four or five 19 gallons per minute.
20 If you look at the probability that that's going 21 to occur, you would get a very wide curve.
That is, the 22 chances of any result coming out would be depicted in a very 23 wide spread of results.
The variability would not be 24 concentrated at the mean.
25 Now, the other side of that is, what happens if ACE-FEDERAL REPORTERS, INC.
20M47-3700 Nationwide Coserage 80tk336-6646
28040.0 f-'7:RT 1259 U
1 you run a test on the good level transmitter?
If actual 2
leakage is high, as it was during that period of time, you 3
would have a very narrow' curve, concentrated in the center 4
around 1 gallon a minute.
The chances of getting a good 5
result under those conditions would be much reduced, and 6
that's what I meant when I wrote that sentence in the 7
report.
8 It is my judgment that had that level transmitter 9
at LT-1 not been used during that period of time the results 10 that were recorded and filed would probably not have occurred 11 and it would have been extremely dif ficult - t.o get a result 12 under 1 gallon a minute.
To give you an example of how that 7-(_
13 works out in practical terms, this level transmitter, as far 14 as we can tell, was repaired on January lith.
It was 15 functioning properly between January lith and January 15th.
16 The leakage was still above 1 gallon a minute during that 17 period of time, the unidentified leakage.
18 You can see from the tests that were filed how 19 difficult it was to get a good leak rate under those 20 conditions, with level transmitters properly functioning, 21 because during that period there were only three tests filed 22 and two of those tests had water additions.
Had they not had 23 water additions, those results would not have come under 1 24 gpm.
And a third test result probably occurred at a time
~
25 when the leakage was temporarily corrected, when the valve ACE-FEDERAL REPORTERS, INC.
~_ m m.
28040.0 rs3RT 1260 U
l' that was leaking was furmanit.ed and it reduced the leakage, 2
and that probably resulted in the test result under 1 gpm.
3 I don't think during that period of time, using 4
good level transmitters, had it not~been for two water 5
additions and a' test run under the conditions I described 6
where the leakage was actually reduced, that you would have 7
gotten an unidentified leak rate under 1 gpm.
8 MR. CAPRA:
I would like to make one small comment 9
on what Mr. Stier said.
We do, in fact, agree with 10 everything he said.
I would just like to point out for the 11 Board's benefit, if you look at those traces throughout the 12 period that the level transmitter was unavailable, and we are s
13 talking about oscillations plus or minus 10 inches or so, the 14 bottom part of the. level transmitter trace for LT-1 is 15 approximately equal in most cases with level transmitter 16 LT-2.
So that we.are always going above the level 17 transmitter during that period of time.
18 MR. RUSSELL:
I would like to also add an 19 explanation which is similar'to Mr. Stier's'and that is the 20 oscillations that we talked about before, with a normal curve 21 distribution about a mean value, we now have an oscillation 22 in addition due to the erratic instrument.
So, if you had a 23 mean value of a leak rate with the other instrument errors 24 being offset because of some leakage to the drain tank of,
(~T 25 say, 1.3 gallons a minute, and you were using a good v
ACE-FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coserage 8 @ 33tw 646
28040.0 r-1261
(_)7RT 1
instrument, you may have an oscillation about 1.3 gallons a 2
minute of maybe a half a gallon a minute, so that the area 3
under the probability curve between 1 and lower is small.
4 If, instead, you use a widely oscillating 5
instrument, the area that would be intersected with the 1 gpm 6
line, if you had a distribution of results of plus or minus 1 7
gpm, centered about 1.3, you have a much larger area under 8
the probability curve.
Therefore it is easier, in that 9~
sense, to get a result less than 1, or a good result.
10
. JUDGE CARPENTER:
Thank you.
We have time for one 11 more.brief question, 275:
Was LT-1 ever repaired during 1978 12 or 1979?
gy V
13 MR. CAPRA: ~In responding to the last question I 14 mentioned two work requests.
I mentioned one on December 5, 15 1978 and I mentioned another on December 25, 1978.
Mr. Stier 16 mentioned an additional one which was January 10, 1979.
.17 During.the break I did look in the report and was able to see 18 that work request.
Those are the three that I am aware of.
19 Also, in responding to the last question, I believe I 20 misspoke and said that those work requests appeared in volume 21 5-A, document tab 2.
It should be document tab 7.
22 JUDGE CARPENTER:
276 is also a small one.
Do you 23 have anything to add to that, Mr. Stier?
24 MR. STIER:
Just that we believe that it was 25 finally repaired, although perhaps -- we think there are some ACE-FEDERAL REPORTERS, INC.
2C347-3XU Nationwide Coverage RG33MM6
r.
__. ~.
20040.0
'"RT 1262 k"J s
7 1
indications that attempts were made during that period of i
2 time to repair it, but it was finally. repaired, we believe, 3
on January 10th or lith.
4 JUDGE KELLEY:
This seems like a good point to 3
[
5 break for lunch.
We are going to take an hour and'a half t
.6 today.
Resume at 1:30.
j
(
7 (Whereupon, at 12:00 a.m.,
the hearing ras 8
recessed, to be reconvened at 1:30 p.m.
this same day.)
9 10 11 12
,g i u i
13 r
14 15 r
?
l 16 s
17 t
4 18 19 20 i
21 22 L
23 I
I 24 I
25 1
1 l
ACE-FEDERAL REPORTERS, INC.
j 202 347-37m Nationwidc Coverage An3.W646 t
28040.0 I 3 R T 1263 N]
1 AFTERNOON SESSION (1:30 p.m. )
2 JUDGE KELLEY:
We'll go back on the record.
We 3
just worked out an und'erstanding, satisfactory to all, about 4
sequence of question presentation to Stier and Russell /Capra, 5
perhaps starting today but, in any event, basically on the 6
tests.
We'll be getting a proposed sequence from them'in due 7
course.
Perhaps it's not necessary before Tuesday morning, 8
but that's when we would need it.
9 We raised an issue first thing this morning about 10 the extent to which there should be input fron.Stier and 11 Russell /Capra on prefiled or other testimony.
I'm just 12 identifying the question.
Counsel were going to engage in
~
b
13 some discussion on that over lunch.
Perhaps you could report 14 the results of that.
Mr. Gephart?
Or Mr. --
15 MR. GEPHART:
Yes.
Well, we, I think had what I 16 could term very fruitful discuccions, Judge Kelley.
Perhaps
~17 we have completed the discussions.
I think some general term 18
-- subject to people rethinking it over the weekend, we kind 19 of know what each other's position is on these things.
20 JUDGE KELLEY:
But there are some divergencies of 21 view, I take it?
22 MR. MC BRIDE:
There is to this extent some 23 unanimity.
I'll let the Staff speak for itself but I do 24 believe there is no perception of need for Mr. Stier and his r~
25 people and Mr. Russell and Mr. Capra to sit through the U'
l ACE-FEDERAL REPORTERS, INC.
2
.- nm,
28040.0
3RT 1264
-]
'I individual operators' testimony and then resume the witness 2
stand to give their views about it.
I may be' wrong on that 3
but I think 1 understand the Staff's position to be that.
4 JUDGE KELLEY:
That's one piece of it, I gather.
5 I gather there's no support for that. approach, correct, 6
Mr. Goldberg?
7 MR. GOLDBERG:
Yes.
8 JUDGE KELLEY:
Would it be useful, now -- do you 9
want more time to consider it or are the parties prepared to 10 state how they think we should approach this?
11 MR. GEPHART:
I think we would prefer to state 12 that Tuesday morning.
Mr. Voigt, for example, is an f-b 13 important part of this.
We would want to talk with him.
I'm 14 sure Mike would, too.
^15 JUDGE KELLEY:
We've had the discussion.
I think 16 Tuesday morning is soon enough.
I do think we should, along 17 about then, hear and then very soon decide what we are going 18 to do.
If you would rather have an opportunity to consult 19 with him, we can wait until then.
20 MR. MC BRIDE:
Thank you.
21 JUDGE KELLEY:
A couple of just quick items.
I 22 did determine that it is all right to leave papers here over 23 the weekend.
The building itself is locked from the 24 outside.
That particular hearing room door isn't locked but 2 5~
the intervening door towards the men's room is, at 6:00 in ACE-FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coserage 800-33MM6
20040.0
(~7RT
-1265 V
1 the evening or something.
The NRC can't warrant security.
I 2
guess my'own judgment, for whatever it is worth, is you'd be 3
safe enough with paper if you left it here.
But that's 4
entirely up to you.
You can do it if you want to.
5 A word on Mrs. Aamodt, I asked my secretary to 6
call Mrs. Aamodt to see what the situation is with her and 7
she reports back to me that Mrs. Aamodt informed her that her 8
daughter was still quite ill.
It had been necessary to miss.
9 this week.
She will not be here on Tuesday, she is quite 10 sure.
She hopes that she'll be able to join the hearing, at 11 least before what we've called individual responsibility 12 begins.
I'm not sure whether she means by that individual k"s) 13 party witnesses or testimony of the experts'about particular 14 tests.
But, in any case, that's really about all I know.
I 15 would say, again, that she has not made any request or even 16 any' suggestion that the hearing should be halted to await her 17 return.
There is no postponement request pending.
So that's 18 what I know on that subject.
I guess we can resume the 19 questioning at this point.
20 JUDGE CARPENTER:
Continuing the same line we were 21 pursuing before lunch, the next question is 276 in the 22 Russell /Capra sequence.
Assuming that LT-1 was not tagged 23 out-of-service, was there anything to prohibit an operator 24 from using it?
25 MR. CAPRA:
Your Honor, we feel that there was.
ACE-FEDERAL REPORTERS, INC.
202 347 3700 Nationwide Coserage 8M336-6646
28040.0
/~1RT 1266
'tQ 1
We believe that if an operator knew or suspected that LT-1 2
was not providing reliable information, using LT-1 to provide 3
input.to the leak rate test, we believe, would be a violation 4
of administrative procedure 1036, and the intent of 5
SB301-3D1, the leak rate test procedure, and essentially a 6
violation of the intent of the tech spec itself.
But there 7
was nothing physically to prevent him.
8 JUDGE CARPENTER:
Mr. Stier?
9 MR. STIER:
I'm in essential agreement.
I think I 10 would put it in more simple terms.
The good sense of the 11 people running the test, knowing that one of the parameters 12 that had to be measured was the makeup level at the beginning 13 and end of the test, using an instrument that was not capable 14 of measuring that, would-certainly not be proper in running 15 the test.
16 I talked before about what I think is the basis in 17 the technical specifications for that.
But I would go even 18 beyond that.
And I think that the testimony that we 19 obtained, in a general sense, tends to support that.
20 Operators recognized that inability to measure makeup level 21 effectively would not be an appropriate way to run the test.
22 JUDGE CARPENTER:
Going onto 277:
If an operator 23 selected LT-2 to provide data to the computer during a leak 24 rate test, does that not mean that the strip chart recording 25 for the period of time during which the leak rate was being i
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage im33M646
28040.0
(~7RT 1267 U
1 performed would be erratic and unreliable?
2
.MR.
CAPRA:
Yes.
That is true, provided the 3
operator did not switch the recorders during the entire 4
test.
We believe in most situations it was not necessary for 5
the operator to know the exact level in the makeup for the 6
one-hour period of the test.
However, if the operator was 7
concerned about the exact level, he could take the initial 8
and final data -- set of data for the leak rate t'est on LT-2, 9
being a good. level transmitter, and switch LT-2 back to the 10 makeup recorder when data was not being collected for the 11 test.
Or, he could possibly switch it momentarily to check 12 the exact level.
-wU 13 I think you have seen from the one exhibit that 14 Mr. Stier showed you this morning the sample trace.
The 15 trace that we observed, while it was erratic and unreliable 16 from the point of view of running a leak rate test, we don't 17 believe that it was that erratic to monitor the general level 18 in the makeup.
19 There were several leak rate tests that were run 20 on the good level transmitter, meaning that for the one-hour 21 period the strip chart that the operator was observing did 22 show the bad trace.
It appeared that it was not an 23 encumbrance to the operator to operate the plant in that 24 manner.
25 There are also several samples of what I mentioned ACE-FEDERAL REPORTERS, INC.
mm.3m
~.o_ m o_,.
,_,,_,.,_m
_ _ _, ~
20040.0
~)RT 1268 (U
1 last about operators switching the level transmitter to check 2
the indication on the other one during the period of the 3
test.
4 MR. RUSSELL:
I would like to add, also, that the 5
assumption that we made in answering this was that the test 6
is being done on the computer.
You still had the option of 7
doing the hand calculation and using the patch panel.
8 JUDGE CARPENTER:
I think those assumptions pretty 9
clearly apply in the construction of the question.
10 Mr. Stier?
11 MR. STIER:
I agree with Mr. Russell and 12 Mr. Capra.
I would like to add one other thing:
It was also 13 po'ssible to obtain a digital readout of makeup level from the 14 computer so that even though the computer -- even though LT-1 15 might have been on the computer, they would not be totally 16 without information about makeup level.
17 As Mr. Capra indicated, the readings for the test 18 are only taken at two times, the beginning and the end.
It 19 would be posrible to switch back to the good transmitter for 20 some period during the course of the test, during that 21 one-hour period.
22 JUDGE CARPENTER:
278 continues:
If an operator 23 wished to monitor the makeup level during the period when a 24 leak rate test was being performed, would he not need to use r
25 LT-1 for the leak rate test and use LT-2 for the strip l
ACE-FEDERAL REPORTERS, INC.
l x.e _ m m _,.
1 1
20040.0 r'7RT 1269
~
U 1
chart?
I believe you nearly responded to that, previously.
2 MR. CAPRA:
That's what I was going to say.
I 3
think we pretty much covered that in the previous answer.
4 The important aspect of the surveillance procedure is that 5
the initial and final data set which is collected would be 6
collected on the good level transmitter.
There would be 7
nothing that would violate the literal procedure or even the 8
spirit of the procedure, in such a situation where you have 9
only two level transmitters and one of them being out of 10 service, for the operator to switch back and forth, 11 periodically, during periods where the data was not being.
12 recorded.
13 MR. STIER:
I agree that we've covered most of 14 it.
It may be of interest to you, if you saw the photograph 15 which I think has been supplied by the Staff, of the strip 16 chart.
On that photograph you'll see a little dial 17 immediately below it, by which you can switch from one level 18 transmitter to the other.
It is a very easy change to make, 19 JUDGE CARPENTER:
279 continues in the same vein:
20 Would you agree that many operators believe that the TMI-2 21 leak rate test was unreliable and that the test results were 22 essentially meaninglass?
23 MR. CAPRA:
I believe that's a fair 24 characterization of the testimony that we did in fact hear 25 from the operators during the interview.
There was only one ACE-FEDERAL REPORTERS, INC.
m-c_,.
^-
r
, t:
I i
20040.0
/~,, 1270
./~' ~3 R T f'
\\_/
n 1
operator that I can recall who testified to us that he did 2
not experience any difficulty runt, og the leak gate test.
s 3
JUDGE CARPENTER:
Mr. Stier"!
f 1-t i cl 4
MR. STIER:
I think that that fairlysumNarizes 5
the views of the majority of the people we interviewed.
e 6
JUDGE CARPENTER:
280-If an operator fe[t that
'/
2
,f.
7 the leak rate tests were essentia'lly meaninnakas, would he
/'
8 not prefer to use.LT-1 for the c'emputer and keep LT-2 9
connected to the makeup level re' corder so that the operator J
U 10 at the panel would have reliable /ata available to him?
11 MR. RUSSELL:
I believe we haye ulreadytanswered,
that but I would like to supplementa,ot.,}Thatis
[ 1-tt)at the 12
,f 3 V
e,..
13 operator is required to operate the fac.iLity in accordance
'd 14 with'the regulations, the license.anp f$cill,tV procedures.
}
15 Using an erratic level transmitter an input'ios the computer 16 for a required surveillance test would ' violate the facilit.y.
q 17 procedures.
^)
18 He had a procedure, AP-1010,, which permitted hin '(.;
s e
19 to file an exception if plant conditions precluded the r
20 perfornance of the test such that he wouId have/ a formal g
r 21 mechanism for getting resolution of the'prcblems thejvwere 22 having, either with the instrument or with the procedure.
23 That was not done, as we had pre.rously testified.
24 JUDGE CARPENTER:
Mr. Stier?
r~N 25 MR. STIER:
Well, ' the ques tion as written asks u
ACE-FEDERAL REPURTERS,%INC.
my.g,
- y. 9_
y
-A y
J.
l,, '
'.328040.0
}; (1RT
/~
1271 i
/
I' [,
I whether operator -- an operator, under these conditions, 2
would prefer to use LT-1 for the performance of leak rate 3
tests.
Actually the question does not ask about his 4
preference for using LT-1 for the performance ~ of a leak rate i
S test.
It just -- it asks would he not prefer to use LT-1 for f
6 the computer.
It is implied that he would be using the 7
computer for the leak rate test.
That is, it would be on the 8
computer during the course of the leak rate test.
Let me 9
answer that in pieces.
. r.
(
10 First of all, I think that clearly there would 11 have been a preference.
While the makeup level indicator was
-12 malfunctioning, while LT-1 was malfunctioning, there would 7s
- \\.
)
13 have been a preference to keep that on the computer so that 14 your better level transmitter would read out on the strip 15 chart.
That does not mean that while you run leak rate 16 tests, you shouldn't reverse that.
17 If an operator felt that the leak rate test was 18 meaningless, he might very well decide to leave it on the 19 computer and run a meaningless test by using a meaningless 20 level indicator.
21 He might also, since we are talking about what 22 people might have intended, he might also have wanted to 23 leave it on the computer to increase the odds on getting a 24 good leak rate, if he understood that he couldn't get a good f
25 leak rate using the properly functioning level indicator
{'>T
~-
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 801133MM6
28040.0
(~"RT 1272
()
I while he was having some problems getting a good leak rate 2
test.
3 Either way, whichever' motive it was, it still does 4
not satisfy his responsibilities under the tdchnical 5
specifications to use the leak rate test to document 6
leakage.
7 JUDGE CARPENTER:
Thank you.
I find your answers 8
interesting.
To me the question is so speculative as to go 9
beyond the pale of technical analysis.
Thank you for your 10 answers.
11 MR. STIER:
Well, I tried.
12 JUDGE CARPENTER:
281:
There are frequent
'~'
13 references in the record of thi~s proceeding to operators 14 having switched level transmitters for the performance of 15 leak rate tests.
Do you have any documentary evidence that 16 any operator ever began a leak. rate test by using LT-2 for 17 the computer and then switching to LT-1 prior to the 18 completion of the test, or vice versa?
19 MR. CAPRA:
NO.
We don't have any documentary 20 evidence.
However, there are a substantial number of tests 21 where level transmitters'were switched just prior to the 22 beginning of the test and then switched back just.after the 22 test.
There are also a limited number of tests where the 24 level transmitters were switched momentarily during the
/"3 25 test.
V ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 85336-6M6
28040.0
(~!R T 1273
%-)
1 While the Staff doesn't believe that for any of 2
the tests which were retained that operators began tests on 3
one level transmitter and then switched level transmitters.
4 for the end of the test, we believe that such a situation 5
could be one explanation for one of the leak rate' tests which 6
we'll discuss later, which is leak rate test number 69 by NRR 7
designation, and leak rate test 89 for Mr. Stier.
8 The question here is whether your initial data set 9
is taken on one level transmitter and then switched-to, let's 10 say, a higher reading level transmitter for the final data 11 set.
We have don't believe that that occurred in any of the 12 tests that we evaluated with this possible scenario as being f.3 i
)
\\_/
13 an explanation for the unusual strip chart trace that we'll 14 discuss when we get to those tests -- that particular test.
15 JUDGE CARPENTER:
Mr. Stier?
16 MR. STIER:
We didn't find any evidence of that 17 behavior in our investigation.
However, we do know that 18 there is at least one test that was indicated by the Staff 19 that might have had a switch of level transmitters occurring 20 during the course of the test.
We just don't know.
We are 21 not saying it didn't happen.
We just didn't find enough-l 22 evidence to support that conclusion.
23 JUDGE KELLEY:
Is the source of this a Hartman 24 allegation, or do you know?
l
' (~
25 MR, STIER:
Yes.
U l
ACE-FEDERAL REPORTERS, INC.
I 202-347-3700 Nationwide Coserage 800436-N4
20040.0
,r gRT 1274 LJ 1
MR.-CAPRA:
I believe that's the next question, 2
sir.
3 JUDGE KELLEY:
Sorry.
So it is.
4 JUDGE CARPENTER:
-282:
Has any operator (other 5
than Hartman) ever admitted that he deliberately changed 6
channels for the level transmitter while a leak rate test was 7
in progress?
8 MR. CAPRA:
I don't believe any operator, 9
including Hartman, has indicated that.he deliberately changed 10 channels for-the level transmitter while the leak rate test 11 was in progress.
I believe that Hartman testified that this 12 was a possible method by which leak rate tests could be 7s 13 manipulated.
14 JUDGE CARPENTER:
Mr. Stier?
15 MR. STIER:
I agree.
16 JUDGE CARPENTER:
As I recall, you previously 17 testified the procedure specifically tells one not to switch 18 during the test?
19 MR. CAPRA:
No, sir, the procedure states that the 20 initial data set and the final data set shall be collected-21 using the same instrument.
22 JUDGE CARPENTER:
Right.
23 MR. CAPRA:
What this question is assuming is that 24 you are using a difference in level transmitter readings,
(~3 25 such that you would start the test on the low-level reading
()
ACE-FEDERAL REPORTERS, INC.
202-347-37m Nationwide Coserage 8tn346M6
28040.0
,r-7RT 1275
(_)
1 and you would end up on the high-level reading.
2 JUDGE CARPENTER:
Such a switching is specifically 3
indicated as being undesirable, in the procedure?
4 MR. RUSSELL:
That's correct as it relates to the 5
beginning data set and final data set.
But the procedure 6
wouldn't preclude you from switching during the test when you 7
are not collecting data.
8 MR. MC BRIDE:
Judge Carpenter, I believe the next 9
question has already been answered.
In the interests of 10 shortness of life, I might suggest that I think both sets of 11 witnesses have already answered the question.
12 JUDGE CARPENTER:
The Board will accept your O
V 13 suggestion.
14 JUDGE KELLEY:
Off the record.
15 (Discussion off the record.)
16 JUDGE KELLEY:
Back on the record.
We are now 17 going to have Mr. Stier.and the MPR witnesses and Mr. Russell 18 and Mr. Capra, in a panel, on a series of questions, 19 initially put to Mr. Stier and one set of questions into l
20 Russell /Capra and another -- and we have agreed on a l
21 sequence, but suffice it to say that right now we are going 22 to begin with the question to Russell /Capra, number 84.
23 MR. GEPHART:
Excuse me, Judge, we had some 24 follow-ups, I believe.
/~
25 JUDGE KELLEY:
Let me just finish stating this
(_)3 l
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationuide Cmcrage 8043346M6 L
O 38040.0 k'BRT 1276 1
proposition, then we can do that.
We are going to start with 2
284 for Russell and Capra, then go to 316, skipping 294 to 3
301 for now.
With that understanding, we have, I think, 4
three follow-ups that we will do on the preceding questions.
5 JUDGE CARPENTER:
Mr. Stier, the first follow-up 6
question reads:
If an unstable level transmitter was feeding 7
into the computer and was not tagged out of service, was 0
there any way that an operator would necessarily --
9 "necessarily" underlined -- know that the computer 10 transmitter was in fact unstable?
11 MR. STIER:
Well, with emphasis on the word O(-
12 "necessarily," I would have to say no.
That is, there was no 13 way that he would necessarily know it.
14 If you take that word out of the question, there 15 was, and we have testimony that there was, general knowledge 16 en shifts of the condition of the level transmitter.
That 17 is, word would pass around the few people who were there in 18 the control room as to the condition of the level 19 transmitter, and if somebody wanted to check, even without 20 switching the level transmitter back to the strip chart, as I 21 indicated, I believe you could get digital readouts from the 22 computer of the level that was being measured on the one 23 hooked up to the computer and there were times when it is 24 very clear on the strip chart that a switch would be made, i
25 You see spikes on the strip chart that we believe were made ACE-FEDERAL REPORTERS, INC.
2024474701 Nationuide Coserage 800-336-6646
f38040.0 1
BRT 1277 1
by operator -- by an operator simply switching over te the 2
other transmitter to check, to see how it was working.
And 3
the. needle would go way up, indicating.that.it was not 4
working properly, and then he would switch back.
5 So, not necessarily, but I wanted to clarify the 6
answer and indicate that there were ways with which, in 7
relative ease, one could find out.
8 JUDGE CARPENTER:
Yes?
9 MR. RUSSELL:
I would like to point out there were 10 also some tests that actually printed out values for makeup 11 level that were plus 99.9 inches, basically the top of the 12 tank, when the makeup strip chart indicated that was not the 13 case and it was clear that the instrument was then out of 14 service, based upon the data recorded on the surveillance 15 test sheet itself.
16 JUDGE CARPENTER:
On the computer printout?
17 MR. RUSSEI L :
On the computer printout.
18 MR. CAPRA:
Also, if you look at an individual-19 test sheet, in other words an extract of the makeup strip 20 chart for an. individual test, it is difficult to really 21 understand that, since you look at a test sheet it may -- may 22 not indicate any period of instability during the part that's i
23 extracted for exhibit purposes with a rate test.
The only 24 way you can really see the pattern that developed and O
25 understand fully how much this level transmitter was switched ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coverage 8533MM6
( 38040.0
BRT 1278 1
back and forth is.to actually take the entire roll of strip 2
chart and roll it out and follow it,~ hour by hour, day by-I 3
day, and see where all the switches were made.
4 There are occasions where I'll see the level 5
transmitter was switched to put the good level transmitter 6
onto the computer for a one-hour period and then switched 7
back.
There is no leak rate test retained for that period.
8 After you see a number of these, you believe that 9
the reason that they were switched was to perform a one-hour 10 leak rate test but no-test is retained; yet later in the same 11 watch when the bad level transmitter is on the computer, 12 there is a good leak rate test that is retained.
By being 13
" good" I _mean less than 1 gallon per minute.
14 There are at least one, and possibly more 15 handwritten notations on the strip chart which show the 16 handwriting switched to LT-2 for leak rates, beside where the 17 erratic trace is printing out.
18 JUDGE CARPENTER:
Turning to the second follow-up 19 question:
Did one of the level transmitters begin to exhibit 20 fluctuations, again, in March of 1979?
21 MR. STIER:
Is that directed to me or to all of 22 us?
23 JUDGE CARPENTER:
All.
24 MR. RUSSELL:
Could you read the question again, O
25 please?
ACE-FEDERAL REPORTERS, INC.
202-347-371U Nationwide Coverage No.3%fM6
. ~ _. - - _. -.
., _. _,~.-__-_ _ _
[ j8040.0
\\-BRT 1279 1
JUDGE CARPENTER:
Yes.
Did one of the level 2
transmitters begin to exhibit fluctuations, again, in March 3
of 1979?
4 MR. CAPRA:
I do not recall that case occurring.
5 It may have, but I don't recall that from our analysis and I 6
don't believe that was the case for any of the leak rate 7
tests conducted during March of
'79.
8 JUDGE CARPENTER:
Mr._Stier?
9 MR. STIER:
I don't have any knowledge of that.
10 JUDGE CARPENTER:
Additional follow-up question:
11 If an operator or supervisor were to testify that perhaps the
.O
's) 12 nost important indicator of' plant conditions available to him 13 in the control room was the makeup strip chart level 14 indicator, would that in any way justify hi's desire to have 15 available to him the immediate and long-term, underlined, 16 reliable trend,. trend underlined, in makeup level?
17 MR. CAPRA:
Based on the testimony of operators 18 that we interviewed, makeup level was in fact an important 19 indication that they did look at and did observe.
But as we 20 discussed in our earlier testimony, we don't believe that 21 that was the reason for not putting the level -- the good 22 level transmitter onto the computer for at least the data 23 taking part of the leak rate test.
We didn't have anyone 24 exprese testimony that the reason they did not switch was O
25 because of the reason in the question posed.
ACE-FEDERAL REPORTERS, INC.
202 347 3ht)
Nationwide Coterage MXL336-Ne s.__.
28040.0 RT 1280 1
JUDGE CARPENTER:
Mr. Stier?
2 MR. STIER:
I agree.
3 MR. MC BRIDE:
Judge Kelley and Judge Carpenter, 4
the problem with the answers was that the question 5
presupposed some testimony which the investigators may not 6
have heard.
It struck me that they answered the question 7
based on what they heard.
What we are trying to do here is 8
create a record that will then be available to the board when 9
operators and supervisors testify.
10 I wonder if the question could be restated, in 11 essence, as a hypothetical.
It would make no difference on g
12 the record if there were no such testimony in the individual 13 responsibility phase.
But the purpose of the question was, 14 in essence, as a hypothetical, so that if there is such 15 testimony by an operator or supervisor the Board will be able 16 to evaluate that testimony.
17 JUDGE KELLEY:
Can you restate it as one 10 question?
19 MR. MC BRIDE:
Y as, sir.
It an operator or j supervisor were to testify that perhaps the most important 20 f
21 j indica tor available to hin in the control room were the 1
22 l makeup level indicator, would it in any way justify his 23 desire to have that, the immediate and long-term trend of 24 that indication, available to him on the strip chart when he (7_,)
i 25 walked into the control room?
ACE-FEDERAL REPORTERS, INC.
l 202 147-37m NoonwiJe Coserage KM 33MM>
( )8040.0
\\-BRT 1281 1
JUDGE KELLEY:
Do the witnesses have additional 2
comment when the question is phrased in-that manner?
3 MR. RUSSELL:
It is true that that is an important 4
instrument to have indication.
In fact we require that at 5
least one such instrument be available in order to operate 6
the facility.
The nexus of that to leak rate testing is 7
where I don't agree with the question or the premise.
There 8
are other mechanisms that can be used to perform the 9
surveillance test.
One could, for instance, do a hand 10 calculation and not use a computer,'such that you would not 11 even have to switch it for the two-minute data collection at 12 the beginning and the end of the test.
13 I don't believe that a supervisor would be particularly concerned about a short period df time, on loss 14 15 of data, when you look at that in context of what they were 16 doing and what the months of strip chart records indicate.
17 JUDGE CARPENTER:
Mr. Stier?
18 MR, STIER:
As I understand the question, it did 19 not have to do with a leak rate test.
It had to do with the 20 desire on the part of a supervisor to have the strip chart 21 available.
That is, a good makeup level indicator reading 22 out on the strip chart available to him.
I think the answer 23 to that is, certainly,.yes.
Clearly, they felt that the 24 makeup strip chart was a very important instrument.
But 1 O
25 agree with Mr. Russell's answer insofar as it indicates that ACE-FEDERAL REPORTERS, INC.
202J47dho Nationwide Coserare 80M W6M6
{lg28040.0 BRT 1282 1
you could still have the good makeup' level indicator 2
available on the strip chart and use it to run a good leak 3
rate test.
4 JUDGE KELLEY:
Thank you.
5 MR. MC BRIDE:
I thank the Board for its 6
indulgence.
There was a good-faith representation that you 7
may in fact hear that testimony.
8 JUDGE KELLEY:
The Board understands, then that 9
this would bring us to question 284 to Russell and Capra, my 10 copy of which is in the Xerox machine.
11 (Discussion off the record.)
1
'n 12 JUDGE KELLEY:
We do have one additional question 13 trom the Staff, follow-up question on the preceding set of 14 questions?
15 JUDGE CARPENTER:
We have one Staff question to 16 Russell.
Would having the " good" makeup level transmitter 17 hooked up to the strip chart, even for the important reason 18 of having this available and accurite, justify not using it 19 for a leak rate test to show compliance with the 1 gallon per l minute limit on unidentified leakage?
20 j
21 MR. MC BRIDE:
I believe that's been asked and l
22 j answered.
i 23 i
MR. GOLDDERG:
I know Mr. Stier answered it but 24 l I'm not sure the Staff witnesses have answered that.
n
[
25 l
MR. RUSSELL:
The answer is yes.
Excuse me --
i ACE-FEDERAL REPORTERS, INC.
s_. <-
~~
n i
20040.0
'~BRT 1283 1
JUDGE KELLEY:
You answered it yes?
2 MR. RUSSELL:
Wait a minute.
I need to make sure i
3 I. understand the question.
It is not acceptable to'use that 4
for as a reaso.n for not performing the leak rate test, using 5
an instrument which is operable.
6 MR. CAPRA:
So the answer is no.
7 JUDGE CARPENTER:
Having run out of 8
permutations --
9 JUDGE KELLEY:
Which brings us to question 204.
10 Judge Bright will pick up at this point.
11 JUDGE BRIGHT:
We are going to a new section A) k-12 called " strip chart time error."
13 284 is:
What source of time did you use to 14 determine the "real time" of any given event?
15 MR. CAPRA:
This subject of strip chart timing 16 error, we've got a series of 10 questions on it.
Looking 17 over the entire series, I think it covers the full range, so 18 I will limit my answer specifically to the question here.
19 In performing the leak rate test analysis, there 20 were basically three times it had to be considered.
One was 21 computer time, the second was makeup strip chart time, and 22 the third was log time.
23 By makeup strip time I'm talking about the time 24 there is a preprinted time of day, ranging from midnight to 25 the following midnight on the strip chart itself.
ACE-FEDERAL REPORTERS, INC.
2(C Nationwide Coserage 80t>33M6M
,_. 347 37m _. _ _. _.,_., _ _ _ _ _,
.. ~
!8040.0 RT 1284 1
In log time I'm referring to the control room 2
operator log -- I'm sorry, the control room log or the shift 3
foreman's log, which is the time that was recorded in the log 4
by the operator that the event occurred.
The computer tine 5
that was printed on the leak rate tcst sheet for the start 6
and completion time of a test is ccnsidered -- was considered 7
by us as clock time or real t i ra e.
8 Computer time and log time were considered.to be 9
the same with the except. ion.of occasionally there was an 10 obvious logging error.
Ani the strip chart time was adjusted 11 to computer time or real time.
So the real answer in that 12 all of the times were based on the computer time itself, the 13 time of the leak rate test.
14 JUDGE BRIGHT:
Vr. Stier?
15 MR. STIER:
I would like Dwight Harrison would 16 have to answer that.
He's the one who made the adjustments 17 and I think it's important to have a comparison of the way 16 the Staft did it to the way he did it.
19 MR. HARRISON:
I believe our techniques are l
20 essentially the same and I agree with Mr. Capra on how he i
l that.
I would comment that in general the t.imes that il defined 21 22 i the operator puts in the logs there's -- I don't know if you 23 call it an error or not, but ho tends to round thinos off to 24 the nearest five minuten in most cases.
I think that's not
-r 1
25 really relevant.
I agree with him.
ACE-FEDERAL REPORTERS, INC.
202147.l?m NatenwiJe ( m erage RIL lt6 t6v,
[ )8040.0
BRT 1285 1
MR. STIER:
Judge, I might suggest that in 2
understanding how our records of. test analysis should be 3
understood, MPR has included at the beginning of Appendix K 4
of volume IV, an explanation for the way strip charts have 5
been marked by them in order to make adjustments for time 6
error.
You'll see that as time was established, they drew aa 7
arc in a dotted line with an arrow, showing you where you are 8
on the chart, in terms of computer time.
9 I just wanted to call that to your attention so 10 that as you look through the test data you'll understand why 11 the markings have been placed on there by MPR.
12 JUDGE KELLEY:
Can I just ask, I may be jumping 13 ahead, but "real time," I'm not sure what that means.
Is 14 that Eastern Standard Time or Greenwich Mean Time?
What kind i
15 of time is this?
16 MR. CAPRA:
Whether it's Eastern Standard Time or i
17 Daylight Savings Time will become important later for one 18 particular test.
19 No, we are talking about basically the time that 20 the computer has in its data base.
That time is printed out 1
21 at the top of the leak rate test and the beginning and ending 22 data times are also printed out on the leak rate test sheet 23 itself.
We are assuming that that is real time.
24 JUDGE KELLEY:
Which in --
. O 25 MR. CAPRA:
Clock time.
l t
i ACE-FEDERAL REPORTERS, INC.
202 347 3700 NIionnide roverage kutO)MM6
('38040.0 bdRT 1286 1
MR. STIER:
In Harrisburg, Pennsylvania.
2 MR. CAPRA:
Clock time and computer time should 3
ideally be the same.
There was a responsibility of the 4
midnight shift to synchronize recording instruments, clocks 5
of that nature.
6 JUDGE KELLEY:
Did it print out an hour, a minute 7
and a second or what?
8 MR. CAPRA:
It printed out to the second.
9 JUDGE KELLEY:
To the second.
10 MR. HARRISON:
Let me clarify one thing, your 11 Honor.
There were many times that the time on the strip 12 chart is not master clock time, by some fairly.large 13 amounts.
Sometimes as much as one or 13 hours1.50463e-4 days <br />0.00361 hours <br />2.149471e-5 weeks <br />4.9465e-6 months <br />.
I don't know 14 the reason for that other than probably the operator felt it
. 15 wasn't particularly important.
He's interested in the littic its piece --
17 JUDGE KELLEY:
He's interested in the interval, I 18 assume.
19 MR. CAPRA:
The procedure itself, let me. read it.
20 What I'm reading from is shift relief and log entries, which 21 is administrative procedure 1012, item 3.1.7 states:
"Each f
22 recording instrument shall be checked on the 11:00 to 7:00 23 shift for correct timing and legibility of marking."
The way 24 that was routinely carried out, at least what we have
, O 25 observed, is the strip chart time was not necessarily reset l
ACE-FEDERAL REPORTERS, INC.
m,. m,
~ _ u. ~,,.
m n_
c-.,
,-.--.._.,,.,.-.--._-..,-,.,-,-,.-..-,,---._,.,.,..,,-,,_-,.n..
38040.0 BRT 1287 1
so that the printed time at the bottom of the strip chart 2
corresponded to real time.
But usually on the mid-watch, at 3
midnight the strip chart would be either printed by hand or 4
with a stamp, the date would be affixed to it.
There are 5
periods of time where the strip chart was physically moved to 6
try to match the printed time on the chart with clock time.
7 JUDGE BRIGHT:
If I understand it, you are saying 8
that your testimony encompasses a number of these questions.
9 10 of them?
10 MR. CAPRA:
No, I'm sorry.
What I meant was you 11 could get started on this question and answer many things 12 about strip chart errors.
I was just trying to answer the 13 specific question that was asked.
I believe that the range 14 of questions will get into -- as we progress through the next 15 nine questions here -- will answer essentially how the timing 16 was done and the analysis was done.
17 JUDGE BRIGHT:
Oh.
Okay.
Let's just plow ahead l
10 question by question.
If you have already answered it, 19 please lot us know.
20 285:
Is it a fact that the operators at TMI-2 had 21 at least three clocks available to them in the control room 22
-- their wristwatches, the wall clock and the computer's own 4
23 clock -- and that some error in determining the real time of 24 any given event may have occurred simply because an 25 inaccurate clock was chosen?
1 ACE-FEDERAL REPORTERS, INC.
x mnu s.sm. umm m,,
mm-
(~'2R T 38040.0 1288 1
I think you have answered that, haven't you?
2 Because you are going by computer time and that's actually 3
printed right on the strip chart?
4 MR. CAPRA:
That's correct.
The biggest 5
difficulty was determining the period on the strip chart that 6
represented that computer time.
In other words, matching up 7
the trace on the strip chart with the printed variables on 8
the leak rate test result itself.
We didn't find a big 9
problem with matching computer time.with log entries, once we 10 established what the period was.
Most of them plotted on the 11 strip chart very close to what they should have.
12 JUDGE BRIGHT:
Mr. Stier?
13 MR. STIER:
I think our experience was the same.
14 We matched the strip chart time to computer time and then, 15 with some margin for error, tried to line up logged events 16 with the computer time to try to figure out when events 17 occurred in relation to leak rate tests.
18 But, as Dr. Harrison indicated, you.have to make 19 some allowances for that logging time because people tended 20 to round things off; the time that was put in might have been 21 put in substantially af ter an event occurred.
That might be 22 a reason for sone error.
23 JUDGE BRIGHT:
286:
Is it a fact that the time an 24 event was shown on the strip chart depicting the makeup level O
25 was often not the same as the real time of the event?
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationmide Coserage MXk336-6M6
- 8040.0 BRT 1289 1
MR. CAPRA:
Yes.
I have testified that is the 2
case and this fact was taken into account in performing the 3
analysis.
4 MR. STIER:
We've answered that and our answer is 5
essentially the same.
6 JUDGE BRIGHT:
287:
Is it a fact that your 7
investigation included an effort to determine the real time b associated with alleged water or hydrogen additions (from 8
9 changes in the makeup strip chart) to associate those events 10 with the one-hour period in which a leak rate test occurred?
11 MR. CAPRA:
Yes, that's true.
'J 12 JUDGE BRIGHT:
Mr. Stier?
13 MR. STIER:
Yes.
That's true.
If you would like 14 a further explanation of how that was done, I could have 15 Dr. Harrison do that.
16 MR. RUSSELL:
That's the next question.
17 MR. STIER:
Oh, Okay.
I hadn't seen the 18 questions yet.
Would you like that done now or would you 19 like us to wait for the next question?
20 JUDGE BRIGHT:
Well, if you have something you 21 l would like to add, please do it.
1 22 MR. STIER:
I understand it will be covered.
23 MR. HARRISON:
I believe it will be covered in 24 subsequent renponses unless you particularly want ne to 7-( '
25 discourse on timing charts.
I think it would be Lest to wait l
ACE-FEDERAL REPORTERS, INC.
1 202 347-1700 Nanonwide Coserage W 3 3t> (M6
!8040.0 RT 1290 1
until the next question.
I believe, in fact, that we both 2
used essentially the same techniques and both faced 3
essentially the same problems and I have no rcason to expect 4
us to have particularly different answers.
5 MR. STIER:
He just wants Mr. Capra to give the 6
answer.
7 JUDGE BRIGHT:
Let's try the next question and 8
see.
9 288:
Is it a fact that the makeup strip chart was 10 often hours off by the time it was reset?
If so, is it 11 therefore possible that there is some significant uncertainty 12 in determining the real time of events depicted on the makeup 13 strip chart?
To use a hypothetical example, if the strip 14 chart depicted an event occurring at noon, when it occurred 15 in real time at 8:00 a.m.,
and a second event was depicted on 16 the strip chart as occurred at midnight, when i t occurred in 17 real time at 10:00 p.m.,
did you assume that every " hour" 10 between noon and midnight on the strip chart actually depicts 19 70 minutes of real time for that period?
20 MR. CAPRA:
The answer to the first part of that
{
21 question is yes.
There is some uncertainty.
And the answer 22 to the second part of the question, for this particular 23 l example, is yes, for the time periods quoted.
But, in order l
24
- to actually plot the logged events on the strip chart, the 7
/
I l
\\'
25 error rate in minutes per hour of strip chart time was ACE-FEDERAL REPORTERS, INC.
20347-3 TO N,uionwide Cmerage 8a L3%W46
I i
i i
[)8040.0
~BRT 1291 I
developed to relate real time to chart time, 2
Using the example that was posed in the question, 3
at clock time 0000, the makeup strip chart' error is plus four l
l 4
hours.
At clock time 2200 the makeup strip chart error is i
5 plus two hours, If you accume a linear rate of change during 6
this time period, the makeup strip chart error between 0800
[
7' and 2000 is 120 minutes divided by 14 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br />, or 8.57 minutes 8
per hour.
Thus, for example, at time 100, the makeup strip.
i 9
chart er.ror would be plus three hours and 26 minutes, which 10 would correspond to 1527 on'the makeup strip chart.
l i
11 However, in actually performing the analysis, we I
12 didn't find in most situations that an error rate was 13 present.
In other words, the error was not necessarily 4
14 changing with time.
What we found in tne majority of cases i
h 15 was that there was an offset between real time and makeup 16 strip chart time.
That error may have been -- nay have 17 varied anywhere from five minutes to 90 minutes to two 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />.
However, from test to test to test, we found long L
19 periods of time where the -- there was no error rate, just i
20 the constant offset in error.
l 21 JUDGE BRIGHT:
Mr. Stier?
22 MR. HARRISON:
Let me -- I'll answer that part.
l 4
23 MR. CAPRA:
I=m sorry, I didn't finish my answer.
j
.24 I explained the error there.
()
25 JUDGE BRIGHT:
Okay.
1 ACE-FEDERAL REPORTERS, INC.
202-347 37m Nationwide Coserage mn 1)WA
]
- ~.
I'10040.0
\\~$RT 1292 1
MR. CAPRA:
In actually performing the analysis, 2
the entire strip chart for. each one-month period was used.
3 Where necessary, all log entries were plotted, which had an 4
effect on makeup level.
That is, for example, scrams, water 5
additions, feed and bleed operations, startups, shutdowns, 6
power level changes, chart time markings or reset by the 7
operator.
In all, hundreds of points were plotted for each 8
month.
Therefore, in most cases, there is high confidence in l
l 9
determining the makeup strip chart time that was actually 10 used in the analysis.
There are a few periods of time where 11 the chart time is somewhat uncertain.
12 MR. STIER:
I would like Dr. Harrison to respond, i
i 13 JUDGE DRIGHr:
Dr. Harrison, could you speak up?
14 I'm having a difficult time hearing you.
1 15 MR. HARRISON:
I believe that what Mr. Capra has 16 described is essentially the same technique that we used, and 17 you do, indeed, look -- if you will look in the volume IV-B 18 of the Stier report, actually it's IV-C, we have, before we i
19 go into the long dissertation on each and every test, there 20 is an example strip chart called figure K-2, which shows --
21 which I think illustra tes the kinds of things that you have 22 to do.
l 23 You basically have some things on the strip chart i
24 that you can fairly well identify, particularly batched 4
25 additions, which you can then correlate with the log, i
I ACE-FEDERAL REPORTERS, INC.
mm
__. o_,,
(')BRT 8040.0 1293 1
particularly with the CROs log, and because they have certain 2
interval -- there may be a two-hour and three-hour, you can 3
find these and match these things up, this sequence of 4
events, and conclude how nuch the time of the clock that the 5
operator is using to log the event, and the chart, what their 6
difference is.
And you do this for essentially the entire 7
period of time.
You use as much data as a you can get your 8
hands on, And that you then. continuously find the error in 9
the chart.
10 As Mr. Capra described, and I agree with him, most 11 of the time what you found is an offset.
The chart is off 12 some particular amount that is the chart was moving, to the 13 best of our ability, at the requisite inch an hour, 1 14 believe.
Something of that order.
i l
15 There are a few times where there seemed to be 16 something which caused a constant error to be put in, some 10 percent or so, in some cases.
In those cases you'd have to 17 l
7
)
18 do what was described, as far as constructino some kind ot 19 rate of chance of the errer.
But that occura -- that's not a
.I l'
20
,very common kind of error in ceneral, If it is ot't 15 i
21 ninutes today, it will be off 15 minutes six hours from that i
22 and eight hours from that.
l 23 1 eccentially acree with Mr. Capra's commente on 24 the subject.
I think, in our analyses of each individual l
(:)
l 25 tect, where we believe there ic significant uncertainty in ace-FEDERAL REPORTERS, INC.
xwm w w. w~~,.
nm
[ N 8040.0 DERT 1294 1
the timing of the chart, which can occur when there aren't 2
very many events to look at and there aren't very many 3
additions, then I think I have noted that in our analyses, 4
that there is some uncertainty.
5 JUDGE BRIGHT:
How are those charts driven, by. the 6
way?
A synchronous motor?
7 MR. HARRISON:
It's a synchronous motor, integral 8
to the timing mechanism.
There's a lot of machinery 9
associated with this device and it drives it through some 10 gearing arrangements which drives the chart.
Chart drives 11 are a mechanically interesting device which sometimes has 12 l troubles.
I 13 JUDGE BRIGHT:
It would seem to me you wouldn't 14 cet a built in lag on something like that.
I mean the paper 1
15 is running clower than the manufacturer says it is going to 16 or --
4 l
17 i
MR. COLE:
They nay not have set it exactly right at midnight, you know, when he set it it could have been ld l
19 lclightlyoff.
20 HR. RUSSELL:
Probably the easiest way to explain I
21 where you may have a changing error rate is if you misa l
22 getting the paper s},recket holes over the spockets on the I
23 platen.
Iike on your tyt>ewriter, if it slips then the 24 ocatim) nechanian is goinq at the proper speed but the paper O
25 l may not be advancing at the proper speed.
I I
4 I
ACE-FEDERAL REPORTERS, INC.
m m.n..
- ,-a, ne m,.
-nw.
-..___,,.-,,_,..,,,-,.--.._.~,_,,,_,_-._,--,._,.-_-__,-...,_,_..v,-__..
g:RT 8040.0 B
1295 1
It's like computer paper, you put it in and once 2
you set it up, if you set it up a few minutes off with 3
nothing else changing, it will stay a few minutes off 4
throughout.
And I think that's basically what we have tried 5
to describe.
6 JUDGE KELLEY:
Let's take 10 minutes.
7 (Recess.)
8 JUDGE KELLEY:
Back on the record.
Judge Uright 9
will resume the posing of the ques tions and we'll quit by 10 4:00.
11 MR. GOLDDERG:
Judge Kelley, in answering question t
12 280 I believe Mr. Capra made a misstatement, which he would 13 like to correct at this time.
14 JUDGE KELLEY:
Very wel1.
Go ahead, Mr. Capra, la MR. CAPRA:
There in ore period of time in my 16 anuwer where 1 meant to say 2200 and I said 2000.
I have 17 conferred with the court reporter and he will correct that in 10 the transcript.
To try to explain where it appeared in my 19 I answer is very difficult.
20 JUDGE KELLEY:
Is that 10:00 and 0:00?
l 21 MR. CApHA:
2200 in 10:00 p.m.
i JllDGE MELLEY:
It's a time seterence.
j 22 23 JUDGE BR IGitT :
Shall we continue with 209:
In it 24 l possible that, instead at the paper in the atrip chart i
25
- recotder turninq unitormly more quickly than real time, t.h a t Aa-FliDliRAt RIFORTliRS, INC.
- o: w n.,
s _ m_,,
..,m,m
3
.I38040.0
\\-dRT 1296 1
in reality deviation of the paper time from real time 2
occurred because of such disruptive events as the paper 3
sticking, an operator advancing the paper, an electrical 4
short, or other reasons?
5 MR. CAPRA:
Yes, that's possible.
In a few cases, 6
as we testified earlier, the makeup strip chart error was not 7
linear.
However, in those particular cases we used hundreds 8
of points in each nonth to try to affix the makeup strip 9
chart error for any particular time.
10 In addition, during periods when LT-1 was 11 providing reliable information, the beginning and ending 12 values of the makeup level that were printed on the 13 surveillance test sheet were also used to help fix the points 14 on the strip chart for the period of the leak rate test.
1 15 JUDGE BRIGitT:
Mr. Stier?
16 MR. STIER:
Our approach was essentially the same
]
17 and I think the point Mr. Capra made was a good one, where 10 the level transmitters appeared to be operating properly, 1
19 that was a means of checking the location of the Icak rate 20 test on the makeup strip chart printout, matching up levels.
21 JUDGE BRIGilT:
Question 290:
Does the range of l
22 possible explanations for the difference between paper time l
23 and real time suggest any uncertainty in your conclusions 24 that water or hydrogen additions occurred one minute, or 25 three minutes, or five minuten before the end of a one-hour ACE-FEDERAL REPORTERS, INC.
-n.,
~ _,a. o_,.
g8040.0 BRT 1297 1
leak rate test?
2 MR. CAPRA:
I testified earlier, for some tests, 3
chart time is somewhat uncertain.
However, for most tests an 4
error of one to five minutes is not significant.
In most 5
cases where exact timing is significant, such as water fa additions during the last few minutes of a test, there were" 7
additional factors that we considered.
These factors include 8
whether the time of the event was logged in the control room 9
log; whether an addition was entered into the computer by the 10 operator and included in the calculation; whether data sheet 11 4,
which is the operator-caused changes portion of the leak 12 rate test procedure was filed with the test; whether the
\\
13 makeup level indication matched the strip chart; and, also, 14 testimony of operators who may have testified that they did, 15 in fact, add water during particular tests during the last 16 few minutes of the test.
17 JUDGE DRIGHT:
Mr. Stier?
10 MR. STIER:
Well, in those instances where MPR had I
19 any doubts, any hesitation to fix an event on the strip 20 chart, they noted it in their cover sheets.
They approached 21
! this kind of analysis very carefully and very i
?2 conservatively.
And unless they were certain, they didn't 2) say that they were certain.
24 I,e t me oive you an example ot the kind of thing
,m 25 jthat can happen, just 30 you rea l l.:e where those ace-FEDERAL REPORTERS, INC.
I
- o: m. n.,
s,..,,,.,a., -
m m o,
G t
/
a p
s
( 9 8040.0 v 3RT 1298 1
uncertainties might be.
There's one, test I can recall in 2
which there may or may not have been a small water or [
3 hydrogen addition and there was, clearly, a rather large 4
water addition either immediately hefore or just 'at thy t
5 beginning of the test.
If the water addition was mady just 6
at the beginning of the test, the starting level of,the/ test 7
would have been lower than it would hi e been had tbc tost --
O had the water addition been made befe e
.nd.
Wherethatkindofuncertaintypecurred, sat is, 9
10 you don't know whether thetestwasmadeohawateraddition 11 or just af ter, you'll see that noted ve}y arefully on the 12 cover sheets that MPR provided foe those tests.
e[
thinka's}generqlind.i. cation,after 13 MR. COLE:
I tests and cop 0arin l' thbrtMdth ours, 14 going through all the l
15 thiswasnotapointofanysignificantdev[, tion.
16 JUDGE BRIGHT:
Very well.
29f How much i
17 uncertainty is there in your calculition of the real time' 18 associated with an event depicted on the strip chart?
19 MR. CAPRA:
We didn't actually-calculate the
- (,
20 confidence bands or uncertainty for the 161 tests that we 21 evaluated because each test is somhwhat different.
H3 wever )
22 we did do a comparison of the makeup strip chart times used f 23 by NRR and also used in the Stier report, Out of our 161 4
24 tests, 104 of the test times agree enctly; four tests, the 25 times agree within one minute; 10 tests, the times agree ACE-FEDERAL REPORTERS, INC.
202 Nationwnte Coserage
- 1k 33MM6
_147 37tio. _ _ _ _.. _ _ _ _ _ _ _, _.. _,,. _. -, _...., _ _ _ _.. _ _. - - - _ _ _ _ _ _.
... _ _ ~ -.
9: 8040.0 BRT 1299 1
within two to three minutes; for 17 tests, the times agree 2
within four to five minutes; and there are 26 tests in which 3
the times are greater than five minutes.
4 Of all of the tests that involve a time 5
difference, there are only four cases where timing was a 6
significant factor in calling a test either valid or invalid 7
in our terminology.
These tests are NRR test number 6, 77, 8
96 and 111.
We have questions posed to us on each of those 9
four tests, which we will get to later.
10 JUDGE BRIGHT:
Mr. Stier?
11 MR. STIER:
My response would be essentially the
,'i '
12 same as Mr. Capra's.
13 JUDGE BRIGHT:
292 says:
Have you assigned a 14 confidence factor to your calculations of real time?
If not, 15 why not?
If so, what is that confidence factor?
16 MR. CAPRA:
As I testified to, in response to the 17 previous question, calculating a statistical confidence 18 factor for each test was not done.
We have considered the 19 uncertainty of the makeup strip chart time for each test and i
20 l considered other evidence to establish personal confidence l
21 l that we have, in fact, got the correct time and we did not 22 rely on any of these particular tests, in which there is a 23 timing question, to draw any conclusions regarding willful i
24
! manipulation of the tests.
n l
25 i
JUDGE BRIGHT:
Mr. Gtier?
ACE-FEDERAL REPORTERS, INC.
l 3_.,
,_ m r-mo -
\\
918040.0 BRT 1300 1
MR. STIER:
Yes, I don't think you can establish a 2
general level of confidence for matching up real time with 3
chart time.
Each test is unique.
In some situations you can 4
be completely confident because events match up very nicely 5
and because the beginning and ending points of the test, that 6
is makeup level, match up very precisely with the strip 7
chart.
In other situations, you have less confidence.
8 As I indicated, the degree of confidence for each 9
test is indicated.
That is, if there is any doubt, the 10 extent to which there is doubt in the minds of MPR has been 11 noted in their evaluation of the test.
~;
12 MR. RUSSELL:
A similar case for NRR.
If there 13 was lack of confidence in timing, our doubt, we threw the 14 test out.
We didn't consider it, per se, as it affected any 15 of our conclusions regarding operators.
16 JUDGE BRIGHT:
293:
Would you agree that, in a 17 period of significant RCS -- reactor. coolant system --
18 leakage (such as immediately prior to the TMI-2 accident), it 19 is likely that water additions might (or indeed, should) have 20 occurred almost immediately after a leak rate test?
If so, 21
- is it therefore possible that, given that there is come l
22
! uncertainty in your calculation of real time, your conclusion 23 that certain water additions occurred one to five minutes 24 before the end of a one-hour test may simply depict water
(
/
5 25
, additions immediately after a test?
I ACE-FEDERAL REPORTERS, INC.
n w.,-
~ _
c-
~,-
I (m18040.0
' '$!RT 1301 1
MR. CAPRA:
I don't believe that's the case.
2 Beginning on March 3, 1979, our analysis shows that water was 3
added during all of the last 20 leak rate tests.
That's our' 4
test numbers 137 through 156.
In al] but four tests the 5
water additions were included in the test calculation by the 6
operators.
In the other four tests, the NRR times agree 7
exactly with the times shown in Mr. Stier's analysis and, 8-therefore, there was no timing error during the periods of-9 significant leakage.
10 JUDGE BRIGHT:
Mr. Stier?
11 MR. STIER:
We are quite certain about the timing t'%
k-12 of those water additions.
13 JUDGE BRIGHT:
We have reached another --
14 JUDGE KELLEY:
This is the section, as I 15 understand it, where we jump.
We skip -- we don't ask 294 to 16 301, and take up with 302?
17 MR. MC BRIDE:
That's correct.
But with 18 follow-ups.
19 MR. GEPHART:
Did I understand Mr. Capra's last 20 answer to indicate'that in four tests there is no uncertainty 21 about the timing because he agrees with Mr. Stier's 22 analysis?
That's what I thought he said.
23 JUDGE KELLEY:
That's what I thought he said.
I 24 was going to ask him about that.
O
\\"#
25 MR. GEPHART:
Okay.
ACE-FEDERAL REPORTERS, INC.
21c-3473700 Nationwide Coserage FN336-6M6
f 38040.0
BRT 1302 1
JUDGE KELLEY:
We'll get to that.
2 My question was I thought if one matched up strip 3
chart time with real time or computer time and if that's the 4
case it was fine, but I thought you said it matched with 5
Stier and so it was okay.
Is that what you said?
6 MR. CAPRA:
That is what I said.
I was using that 7
as additional verification that an independent analysis also 8
arrived at the same time frame.
I also agree with 9
Mr. Stier's comment that during the latter period which I 10 mentioned, the strip chart -- there were so many additions 11 being made during that period because of the high leakage 12 from the top of the pressurizer, that it is very easy, or was 13 very -- relatively easy to determine the exact time of events 14 because of the frequency.
15 During earlier periods, say in the October time 16 frame, there were much less water additions made and the l
l 17 frequency by which you have points to plot to gain 18 confidence, there was some more uncertainty to that period of 19 time.
In the latter period.
l l
20 JUDGE BRIGHT:
We have a couple of follow-up
-21 questions.
22 For Dr. Harrison, you stated that the operator. log l
typically rounded off to the nearest five or 10 minutes.
i 23 was 24 But that was not relevant.
7_s d
i 25 Is it a fact that, for a significant number of ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6646
[ 18040.0
BRT.
1303 1
about operator actions just before or just after the 2
beginning or end of' the test, and would the difference of 3
five or 10 minutes be relevant to those allegations?
4 MR. HARRISON:
Where we are involved, some 5
allegations as far as five minutes before.the e.nd of the L
-6 test, or things close to the end of the test, in all cases 7
there are other things, independent from the timing of the 8
chart, which one uses, as I believe Mr. Capra did discuss.
9 And I agree with him on that -- to decide whether, in. fact, 10 an addition was within the test or not.
11 In particular, in the case of the March thing, 12 there were actual entrier by the operators that they did in 13 fact put the water in or not, based on the calculation.
I 14 don't think in any case that you have -- that we did solely 15 depend upon the timing of the chart as we described in 16 matching up additions, and what have you, to decide.
17 MR. STIER:
If I could add a little something to 18 that, in.those situations where it became important to
~
19 corroborate that a water addition occurred just before the 20 end of the test, in many instances, particularly the most 21 critical, we had entries in the leak rate test indicating 22 that the operator made the water addition.
So he accodnted i
23 for it and we'll explain the significance of that, of course, I
24 as we.get~into the tests.
But, in effect, he timed it t
('\\
l 25 himself.
1 l
l ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-33M646
8040.0 RT 1304 1
In addition to that, we had operator testimony i;.
2 the most critical cases that the water addition was 3
intentionally made shortly before the end of the test.
So my 4
confidence level, given the analytical work that MPR did, 5
plus the additional corroboration that I had, was very high 6
in those instances.
7 MR. RUSSELL:
There's another part of the 8
technical analysis that I'm sure we'll come to later, that 9
hasn't been mentioned, that has to do with the predicted 10 change in makeup level that would have been observed, based 11 upon the gross leakage that was occurring at the time.
So I
\\ _
12 you also have one additional piece of information.
For 13 instance, if the gross leak rate was such that the makeup 14 level should have dropped 8 inches during the period of a 15 test and the makeup level did not drop 8 inches, but instead 16 dropped 5 inchs, you would have a 100 gallon difference.
17 That aspect was approached differently by way of 18 calculation between NRR and MPR.
They calculated the gross 19 leak rate.
We calculated an expected change in makeup 20 level.
They are essentially the same.
That was also used to 21 help tie down the details of a particular test in addition to I
22 l the timing that we were talking about before, the log entries 23 and the other evidence.
24 So, for the tests where water additions occurred l
25 that Mr. Capra cited, we are quite confident in the time of 6
ACE-FEDERAL REPORTERS, INC.
202 347 37to Nationwide roserage 80lk336-6M6
38040.0
BRT 1305 1
the water additions with respect to the test interval.
2 MR. STIER:
If I might cite another table in our 3
report, and I know you have a lot of tables to look at, but 4
let me add one more to'it --
5 JUDGE KELLEY:
I think that's helpful.
Keep doing 6
that.
7 MR. STIER:
MPR has prepared a table, VI-1,'in 8
volume IV-A, which lists all of the water additions-that 9
occurred during tests that we were able.to identify, and 10 indicates whether they were logged or not and whether they 11 were considered in the calculation.of leakage or not.
It
'T t
12 gives the date and the test number and so forth.
13 JUDGE. BRIGHT:
I have another question for 14 Russell /Capra and Stier.
In your investigation, did you 15 attempt to correlate logging times, between the CRO_ logs and 16 the auxiliary operators logs, for the logging of hydrogen 17 additions?
If so, were there substantial differences in time 18 between these separate logs?
19 MR. CAPRA:
The NRR analysir did not do that.
The 20 times we got for auxiliary operators logs was the times from 21 the Faegre Benson report.
We did not have the auxiliary 22 operators logs that actually reported the time of hydrogen 23 addition.
We did have auxiliary operators logs, but not the 24 ones that would record that information, f
k 25 JUDGE BRIGHT:
Mr. Stier?
ACE-FEDERAL REPORTERS, INC.
202-347-3700.
Nationwide Coverage 8M336-646
8040.0 RT 1306 1
MR. STIER:
Let me give you another table 2
reference.
Table VII-1, which is found in our volume IV-A.
3 It lists all of the hydrogen additions that were logged and 4
indicates whether they occurred during a test.
S In one of those columns you will see which log 6
made reference to the hydrogen addition and the time that was 7
logged for that hydrogen addition.
And in only a few cases 8
do we indicate that both logs, both the CR0 log and the A0 9
log, contained a reference to.the hydrogen addition, and the 10 time noted in each log is indicated on this table.
You'll 11 see that there are some differences in the times that were i
~>
12 logged in those few cases where it is logaed in both.
13 MR. GEPHART:
He said "some differ."
I think the 14 question was:
Were there any with substantial difference?
15 JUDGE KELLEY:
Okay.
Maybe you could elaborate.
16 MR. STIER:
Well, on October 20,
'70, we have from 17 each log, the same time is logged.
10 On October 26, the CR0 logged it at 0936, the AO 19 log, at 0930.
We have matched those times up with 0916 and 20 0910, respectively.
21 JUDGE KELLEY:
Will we be getting into these 22 particulars in due course, anyway?
23 MR. STIER:
I don't know if these tests are going 24 to be on the rests we will specifically gut into here.
There
\\
25 are only a coulele of others.
February 5,
'79, CR0 log is ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cm erage mb336-(M6
[ 78040.0
'- B R T '
1307 1
0445 and the A0 log is 0450, a five-minute difference.
And 2
we have matched that up against what we think is the correct 3
-- are the correct chart times.
4 on February'15, CRO log is 2100 and the.A0 log is.
5 2050.
6 On February 21,
'79, the CR0 log is 1540, A0 log 7
is 1535.
8 On March 1st, CRO log is 0130, the AO log is 9
0100.
A 30-minute difference.
10 And the next one is a five-minute difference on 11 March 6th, and the final one on March 23rd is CRO log 2031, r~s.
(
12 the Ao log is-2050.
I'm sorry, did I say five minutes?
On 13 March 6, it's not five minutes, it's 15 minutes.
You can get 14 a sense of the difference.
15
. JUDGE KELLEY:
All right.
16 MR. MC BRIDE:
Judge Kelley, I wonder if I might 17 just suggest something to you that the Board'might be 18_
thinking about when~they hear from the operators.
This is 19 not testimony but just a question for you to remember and 20 that is, I think you'll find that some of the operators, at 21 least, will confirm that the reason there might be these 22 discrepancies-is not just.that their wristwatches weren't 23 sychronized between the auxiliary operator and CRO, but 24 rather the auxiliary operator might wander by the tank and 25 add.some hydrogen and then when he got to a phone'later --
ACE-FEDERAL REPORTERS, INC.
202 347-3XU Nationwide Coserage 80 4 336-6646
[
28040.0
'BRT 1308 1
because they are physically separated by some distance --
2 call the CRO, in the control room, and tell him.
3 Just in order to underst,and how it could be 19 4
minutes apart or something like that, you might just remember 5
that that is a possible explanation for'this sort of thing.
6-That was one of the reasons why we had thought that a visit 7
to the site might help' the Bo'ard a little in understanding 8
the physical separations between the auxiliary operators and 9
the control room operators, on occasion.
10 JUDGE KELLEY:
Very well.
We'll bear _that-in 11 mind.~
/D k-)
12 Now, we are going to' skip from 302 -- we are 13 skipping from 294 to 301, and pick up with 302.
Go ahead.
14 JUDGE BRIGHT:
A new section, feed and bleed 15 operations.
16 302:
Is it a fact that the occurrence of a feed 17 and bleed operation cannot always be detected by examining l
l 18 the makeup strip chart?
19 JUDGE KELLEY:
Could we start with sort of a 20 thumbnail description of a feed and bleed operation?
i l
l 21 MR. CAPRA:
What we are referring to here is the 22 reactor coolant system, including the makeup and purification 23 system, is essentially a closed system.
However, reactor 24 coolant is constantly purified a certain amount, so that some f_s b
25 of the reactor coolant is taken off of the reactor coolant ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6646
(
18040.0
'"BRT 1309 1
system, sent through a cooler, purification filter, 2
demineralizer, and returned back to the makeup tank.
There 3
is a makeup pump which takes the suction on the makeup itself 4
and pumps that coolant back into the reactor coolant system.
5 So that it is a closed cycle.
6 Normally, the makeup -- the rate of makeup and the 7
rate of letdown, which is the coolant coming out, are the 8
same.
There are some periods of time where you may want to 9
deborate the reactor coolant system for reactivity control 10 purposes.
During those periods of time, the coolant is let
-11 down but it is let down to a different place than the k~l 12 makeup.
It is let down, maybe, to a reactor coolant bleed 13 tank or the deborating demineralicers, in which case, to keep 14 the makeup from constantly decreasing, you need to add makeup 15 water to the makeup itself.
16 There are several sources of-water in which you 17 can do that.
You may add demineralized water.
You can add 18 boric acid.
You can add water from the reactor coolant bleed 19 tank:
Several sources.
20 So the basics of feed and bleed are you are 21 feeding in a certain amount of makeup from one source to the 22 makeup and you are letting down to another tank beside the 23 makeup.
You can control -- there are variable controls to 24 control letdown flow and makeup flow.
~
25 Now, to respond to the question the answer is ACE-FEDERAL REPORTERS, INC.
202-347-3RM)
Nationwide Coverage 800-336-6M6
f)8040.0
BRT
-1310 1
yes. As stated in the NRR report as enclosure 1, page 7 and 2
8, feed and bleed operations did not' always have a 3
distinctive. trace on the makeup strip chart.
At times, water 4
was added, fed, and then removed, bled, from the reactor 5
coolant system in that order.
6 At other times the reverse order was used.
7 Because water could also be added and removed from the system 8
at the same time, these operations might be undetectable on a 9
strip chart but if the rate of removal and addition was equal-10 and the operation was not logged in the control room log.
-11 So, essentially,Jif you are going to set up a feed ni( /
12 and bleed operation you may increase letdown or increase the 13 rate of makeup at..different rates.
The idea is to get them
-14 synchronized so you are letting down and making up a't the 15 same time.
But the order in which you do that may vary.
16 If you are looking at a strip chart itself and you 17 increase letdown before you increase makeup, you are going to 18 see'the level in the makeup drop.
If you increase makeup E
19 before you effect letdown, or increase letdown, you are going 20 to see a rise in the makeup level.
21 If you essentially do them at the same time, keep 22 the feed rate and bleed rate equal, you'll see nothing.
23 That's why you cannot always determine from the trace that 24 you have a feed and bleed operation in progress.
O O
25 JUDGE BRIGHT:
So you have to get that from the ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Cme: age 80h336-6646
lh!8040.0 BRT 1311 1
log?
2 MR. CAPRA:
You don't have to get it from a Ing 3
entry.
But what I'm saying is absent a log entry you may not 4
be able to determine that such a situation is going on.
5 JUDGE BRIGHT:
Mr. Stier?
6 MR. STIER:
I would concur with what Mr. Capra 7
said.
I would like to refer you to a section of our report 8
that deals with feeds and bleeds, provides a general 9
description of the practice and indicates MPR's feelings 10 about their ability to identify feed and bleed operations on 11 strip charts.
That's in volume IV-A, on page VI6.24.
12 I might add with respect to log entries, we found
~
13 you couldn't always tell from the log entry that a feed and 14 bleed had occurred.
In instances the operator would log 15 "added water," and when you look at the strip chart clearly 16 it was a feed and bleed.
They wouldn't always characterize 17 it as a feed and bleed operation, so when you saw an entry, 18 you couldn't even depend on that.
19 MR. CAPRA:
I'd agree with that.
One other thing 20 I might mention is that during periods of time where we do 21 believe feed and bleed operationc are evident on the strip 12 chart, there's usually a fairly distinctive trace, where the 23 makeup level might drop rapidly for a short period of ti me 24 and then show a rapid increase again, essentially about to s
l 25 the level where it was.
We believe that that's indicative of ACE-FEDERAL REPORTERS, INC.
i l
3c-347-3700 Nationwide Cos erage 80h3346M6
,# h
'( 18040.0
B R T.
1312
~1 the operators, again, taking some period of : time to adjust 2
feed rate and bleed rate until they are equal.
3 MR. RUSSELL:
It is also important to review the 4
logs to see what's happening to reactor power and what other
'S conditions are going on in the plant and what the power 6
history has been so you can tell whether you are on a Xenon 7
transient, whether you are, for instance, burning up Xenon 8
because you just started up and gone to high power after a 9
shutdown before, or vice versa.
Because that will give you 10 an indication that operators were needing to feed and bleed, 11 and either borate or deborate, in order to keep the 12 controlling group of rods within the control band that was 13 required by procedures.
So-you have to look at a number of 14 indications and information in order to be able to determine 15 whether feed and bleed was in fact going on.
16 JUDGE BRIGHT:
303:
Is it a fact that feed and 17 bleed operations served the legitimate purpose of increasing 18 or decreasing the boron concentration in the reactor coolant 19 system?
20 MR. CAPRA:
Yes.
Boration ~and deboration often 21 were accomplished:by' feed and bleed operations and they were 22 a routine part of operating the plant.
As Mr. Russell 23 stated, changes in boron concentration were necessary to 24 account for power level changes, load changes, fuel burn you 2S have, and changes in Xenon concentration within the core ACE-FEDERAL REPORTERS, INC.
201 347 3700 Nationwide Coserage 800-D6-6646
[)8040.0
'"BRT 1313 1
while keeping the control rods within their assigned limits.
2 TMI-2 operating procedure 2102-2.1, power operations; and 3
2103-1.2, solubles on concentration control, both address 4
this subject and were used by the operators on a routine 5
basis to operate the facility.
6 JUDGE BRIGHT:
So you answered 304 at the same 7
time?
8 MR. RUSSELL:
We have additional information on 9
that one.
10 JUDGE-BRIGHT:
Oh, you do?
11 MR. RUSSELL:
Yes.
p k-12 JUDGE BRIGHT:
Mr. Stier?
13 MR. STIER:
I agree with their answer to 303.
14 JUDGE BRIGHT:
Well, let's try 304:
Is it a fact i
15 that feed and bleed operations were performed routinely?
16 MR. CAPRA:
I guess.this is more in terms of 17 cmbroidery rather than additional information.
Yes, feed and 18 bleed operat-ions were performed r~outinely.
However, if you 19 context that in terms of performing feed and bleed operations 20 during a leak rate test, I just want to remind the Board when 21 we discussed this earlier that feed and bleed. operations or 22 boration and deboration are one of the routine operations 23 that are cautioned against to the operators in procedure 24 2301-3D1 as not doing that while performing a leak rate O
25 test.
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 8h336-6M6
I
~
)
/.,
t 18040.0
SRT 1314-1 JUDGE BRIGHT:
Mr. Stier?
2 MR. STIER:
I agree.
3 JUDGE KELLEY:
Any special reason why-you might 4
have to do.that or you'd be under some strong operational 5
reason to feed and bleed while a leak rate was going on?
6 MR.. CAPRA:
That certainly is-possible.
As 7
Mr. Russell explained, keeping the control rods, the 8
controlling group'within the operating band, is a legitimate 9
operational concern.
And the controlling group is normally 10 maintained anywhere from 90 to 100 percent out of the core.
11 If you got to a point where rods were at their
\\
12 outer limit or 100 percent, I believe operators have 13 testified a big red light would come on and tell you, you 14 were at your outer limit.
In which case you would have to 15 adjust the boron concentration of the reactor coolant 16 system.
In situations such as this, we believe if the 17 operational necessity came up and it was unforeseen,-that 18 would be a case where you had an operator-induced change and 19 you would either invalidate the test or you would account for 20 it.
21 MR. RUSSELL:
There are conditions for Xenon 22 peaking, which constitute, in quotes, "The design basis" for 23 your ability to borate or deborate the plant, such that you 24 would be frequently making feed and bleed operations-in order gg V
25 to control the boron to control the rods in the right ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 80M36-(M6
O.8040.0
NRT 1315 1
position.
You.see evidence of this in the strip charts 2
following start-ups, after periods of shutdown where the 3
shutdown period is relatively short, where you do have a 4
Xenon peak followed by a rapid Xenon burnout.
And during 5
that period of time, because of the need to keep the control 6
rods in their controlling band, it would essentially preclude 7
doing a surveillance test because of the operational need.
8 We did look at the individual tests, to determine 9
in our judgment whether there was an operational need that 10 required a feed and bleed during tests and we found that that I
11 was not the case.
./~T
(-)
12 JUDGE KELLEY:
Thank you.
13 JUDGE BRIGHT:
Mr. Stier?
14 MR. STIER:
I have nothing to add.
15 JUDGE BRIGHT:
305:
In light of the frequent 16 performance of feed and bleed operations and the frequent i
17 performance of leak rate tests and the fact that one operator 18 was typically on the panel while the other operator performed 19 surveillance tests, is it possible that TMI-2 operators 20 ununwittingly performed a feed and bleed operation during 21 some portion of the leak rate test?
22 MR. CAPRA:
The answer is yes, that was a 23 possibility.
However, there were a large number of tests 24 during which such operations occurred or during operations in O
25 which such occurred, including feed and bleed operations.
ACE-FEDERAL REPORTERS, INC.
202 347-37(0 Nationwide Coverage 80 4 336-6646
1
/x i
18040.0
\\~BRT 1316
-1 We'll get into this in more detail.
But the operations such I
2 as feed and bleed were, in fact, contrary'to' surveillance 3
procedure 301-3D1.
It was this fact of the numerous leak 4
rate tests in which operations which were prohibited by the 5
procedure were conducted that led us to conclude in the NRR 6
report that operators may not have always established the 7
proper initial conditions for running the tests.
8 And we did not believe, at least from our review, 9
that feed and bleed operations were utilized to manipulate
~
but rather it was a result o.f running many, 10 leak rate tests, 11 many tests back to back -- not necessarily back to back, but 12 running so many leak rate tests that, since it was such a 13 frequent thing, they didn't pay as proper attention to it as 14 they, perhaps, should have if they were performing the test 15 less frequently.
16 Did you understand that, sir?
17 JUDGE BRIGHT:
Yes.
It's talking about here, I I
s e
18 think, communications in the control room.
I have been in a 19 Jot of control rooms.
I realize that there is quite a bit of 20 space in them.
But I also never discovered any problem with 21 a guy going to -- not specifically.
I haven't done that.
22 Dut it would seem to me to be no problem that the fellow is 23 getting ready to do a leak rate.
He knows there shouldn't be 24 a feed and bleed at the time.
He hollers:
Hey, Joe, I'm O
25 going to start leak rate tests.
And that would be no problem ACE-FEDERAL REPORTERS, INC.
3 C-347-3700 Nanonwule Coserage 8co 3W%46
f 28040.0
'BRT 1317 1
at all with communications.
Do they don't talk to each other 2
at all in there?
3
'MR. CAPRA:
That was one of the thinge, when we 4
questioned the. operators, we talked about communications and 5
how shifts were organized.
Normally the individual who had 6
the panel who would be in control of feed and bleed 7
operations, water additions, hydrogen additions to the 4
8 makeup, was not necessarily,.and usually was not, the person E
9 who performed the surveillance test.
However, as you can see 10 from when we describe the pictures of the control room from 11 the exhibit, the Bailey. computer was only a few feet away T
~
12 from the auxiliary panel where such operations should occur.
13 Not necessarily when the leak rate test was started that the 14 other operator was standing at that panel, but it's basically 15 a confined space.
Most operators said that they did attempt 16'
-- make attempts to inform'the contr.o1 room -- the panel 17 operator'when they had initiated a leak rate test.
18 That was one of the reasons why, when we were 19 ques tioned about requirements of the shift relief and 20 turnover procedure, that I said it was important, maybe more 21 important for this particular test than others, other
~
22 surveillance tests, that the starting time was logged, i
23 because that would be an indication to the operator, the 24 panel operator who was, in fact, on the panel at the time and O
l 25 maintaining the control room log, that would be a reminder to ACE-FEDERAL REPORTERS, INC.
202-347-3?(Mt Nationwide Coverage 8(uk33M*46
( tRT)8040.0 1318 1
him that the leak rate test was in fact in progress, and not 2
to perform these evolutions.
3 MR. RUSSELL:
I would like to comment, because 4
this was a rather significant issue.
It gets back to some 5
earlier questions <on how difficult was it to perform a leak 6
rate test and what were the communicatio'ns on the shift.
7 It was quite easy to start one.
RCSL -- go.
8 Carriage return.
And the computer was doing its thing and 9
the switching and tagging operator, who is the individual 10 Mr. Capra was talking about, would go off and do other s.
11 things; an hour later it would come back and type lines and 12 say:
Tell me what you did.
13 Well, if there were miscommunications back and 14 forth between the two as a result of doing this test so 15 frequently -- three, four times a shift in some cases or more 16
-- it is possible that an evolution occurred during the test 17 whi~ch would invalidate the test that didn't get 18 communicated.
19 Because of that uncertainty, and because we didn't 20 have any testimony from operators that they used feed and l
21 bleed to manipulate a leak rate test, we did not reach the 22 conclusion that any tests were manipulated by that method.
23 The standard -- minimum standard that I sort of used was:
l 24 Yes, you had to not only show by analysis that something was l
25 done wrong by the procedure that is questionable, but you had i
I 4
ACE-FEDERAL REPORTERS, INC.
ac -
s._ me m m,,
~~
m
~, _ _. - -. _ - _.__,-_._,
m 4.
=
( )BRT 8040.0 1319 1
to have some type of testimony from an operator that said:
2 Yes.
That's the way I did it.
i 3
We didn't have any operators testify that they 4
used feed and bleed operations.
It was a potential way that 5
was identified by Mr. Hartman.
We had no corroboration that 6
anybody else did that.
So we don't really feel that the feed 7
and bleed operations were other than lack of attention to 8
detail in conducting the test and potentially l
9 miscommunication between the operators.
10 JUDGE BRIGHT:
Mr. Stier?
11 MR. STIER:
Our approach was virtually identical r
12 to the description of the way the Staff approached the review 13 of what we categorized as feed and bleed operations, and l
14 whether they were used as devices to manipulate tests.
We, 15 also, did not conclude.that anybody intentionally used a feed 16 and bleed operation to manipulate a test result.
17 I suppose there are two ways you can look at a 18 test in which a feed and bleed operation occurred and the 19 test result was filed and not properly accounted for in the 20 test.
One is that there was a miscommunication.
The other i
2:1 is, depending on the result and how far out of line the 22 result-may have been with any indications of leakage, that 23 the person who performed the test disregarded the result and 24 filed the test anyway because the result happened to be under i
Ov l
25 1 gpm.
Because when a. feed and bleed occurs, depending on t
I i
ACE-FEDERAL REPORTERS, INC.
202 34?.3700 Nationuide Cmcare 800-3% 6M6
. -~
~
d (q 8040.0
$RT 1320 1
lwhere the test begics and ends, and what effect the feed and 2
bleed has on the performance of the test, the result can be 3
way out of line with the expected result.
4 JUDGE KELLEY:
There wasn't any kind of automatic l
5
' signal such that when you pressed the leak rate test button 6
-- I don't know, what, a light went on saying:~ Leak rate 7
test in progress?
Or something built into the system that_
l 8
would prevent you from feeding and bleeding?
Or you would be 9
reminded -- nothing like that?
10 MR. CAPRAt Nothing automated.
We did have 11 testimony from one particular shift where the operator said qk/
12 that they-used a 3 by 5 card when they had a leak rate test 13 in progress and the operator would put it on the control 14 panel.
15 JUDGE KELLEY:
That's the kind of thing I had in l
16 mind; but nothing automatic?
17 MR. CAPRA:
That's correct.
18 MR. MC BRIDE:
In your judgment the next question 19 I may have been answered, but if you choose to put it to the i
20 witnesses, I would suggest that you delete the word "to" in
(.
21 the second line.
It's a typographical error.
22 JUDGE DR.IGHT:
Would you clear up one thing for 23 me?
I have somehow gotten the idea that the person doing the i
24 surveillance test, he has at his control -- I knew he has
_O 25 hydrogen injection because 1 saw the button rign: there next ACE-FEDERAL REPORTERS, INC.
3c 347-370)
Nationwide Coserage M k 34 u.46
l l
1 I
l' I
(sJ0040.0 i
1321 ERT i
1 l to the little chart What about makeup warer?
3 I
2 l
NR. EUESELL(
We can explain the difference 3
between cwitching and tagging --
4 JUDGE BRIGHT:
Does he cont'rcl that, too?
5 HR. CAPRA:
!!o, sir.
Normally there are usually 6
three operators, control room operators on a shift.
Uaually 4
l 7
ene was a trainee under instruction and the other two were, finfact, licensed control room operators.
E 9
One of the operators would have what they call 10 l"the panel."
That was the -- when we were looking at the 11 control room, the view of the control roon, the actual
-O 12 operating panels in which he would control reactor 13 operations, steam generator operations, actually be 14 manipulating the controls of the reactor and the secondary 15
- plant.
l 16 He was the same operator who would maintain the 17 control room log.
He was the same individual who would, i
18
[,under normal cases, add makeup water, conduct feed and bleed 5
19 cperations, add hydrogen to the makeup.
That was a normal 20 l' part of his duties.
I 21 The other operator on most -- most shifts called 1
22 him the switching and tagging individual.
He would be in 23 j charge of switching and tagging, conducting log readings, i
24 taking 100 readings, whatever.
25 One of the routine procedures that he would do is 1
ACE-FEDERAL REPORTERS, INC.
?>c-347 MV)
NiionwiJe Cmerag 1%10 33 Mew 6
- ___._.- _._ __, _ _.._ _.. _ _ _. - _. ~.
l l[ 18040.0
'"$RT 1322 i
1 run the leak rate test.
So, normally they were -- the 2
individual running the leak rate test was different. than the 3
normal person who would add water.ar add hydrogen during a i
4 period of a leak rate test.
5 Now, because of the close proximity of the Bailey 6
computer to the controls for adding water and makeup, there' 7
are tests that we have reviewed which the panel operator is 8
the same individual who conducted the leak rate test, or at 9
Icast ci'gned the leak rate test.
10 MR. MC BRIDE:
Judge Bright, in the preface to 11 your question you stated that you were aware that the Gk-)
12 operators were the ones who could add the hydrogen because 13 you saw the switch in the picture, but I would appreciate it 14 if you would inquire of the witnesses whether, in fact, the 15 operators were always capable of adding hydrogen by virtue of 16 that switch.
17 JUDGE BRIGHT:
That seems a fair question.
18 MR. CAPRA:
Okay.
I testified on that earlier.
19 There were periods of time when the auxiliary operators would 20 have to add hydrogen rather than the control room operator 21 because that automatic capability was not there.
There were 22 tie.es where there were leaks in the hydrogen lines or in the 23 manifold, in which case they kept the hydrogen addition flask 24 isolated, and it would require the auxiliary operator to go i
'25 and open the isolation valve for the hydrogen addition flask ACE-FEDERAL REPORTERS, INC.
I wcr~
I 4:
8040.0 RT 1323 1
in order to have the capability to add hydrogen to the 2
makeup.
The exact periods of time in which it was isolated 3
or not isolated, we were not able to determine or did not 4
determine during our review.
5 JUDGE BRIGilT:
306 and we'll be through with this 6
section, and it ought to take a very short answer, 7
gentlemen.
306:
Is it a fact that all of the TMI-2 8
eperators who have been asked have denied using feed and 9
bleed operations as a means of manipulating leak rate test 10 results?
11 MR. CAPRA:
Yes.
/
12 MR. STIER:
Yes.
13 JUDGE DRIGHT:
Thank you, gentlemen.
14 JUDGE KELLEY:
Any follow-ups on this section?
15 MR. MC BRIDE:
tio, thank you.
16 JUDGE KELLEY:
No follow-ups for anybody?
17 It's pushing on to 4:00.
We are going to 18 reconvene here, then, Tuesday morning at 9:00 with the same 19 cast.
Is there anything else anybody wants to bring up 20 before we break?
21 Thank you very much, see you Tuesday morning.
22 (Whereupon, at 4:00 p.m.,
the hearing was 1
i 23 adjourned, to reconvene at 8:30 a.m.,
on September 16, 1986.)
24
(
)
25 ACE-FEDERAL REPORTERS, INC.
I mm.m
~.,m < -
CERTIFICATE OF OFFICIAL REPORTER
^
(
This is to certify that the attached proceedings before the UNITED STATES NUCLEAR REGULATORY ~ COMMISSION in the matter of:
NAME OF PROCEEDING:
INQUIRY INTO THREE MILE ISLAND l
UNIT 2 - LEAK RATE DATA FALSIFICATION 1
DOCKET NO.:
LRP PLACE:
BETHESDA, MARYLAND O
DATE:
FRIDAY, SEPTEMBER 12, 1986 i
were held as herein appears, and that this is the original transcript thereof for the file of the United States Nuclear Regulatory Commission.
(sigt)
/r- /
(TYPED /
i JOEL REITNER Official Reporter ACE-FEDERAL REPORTERS Reporter's Affiliation, INC.
i
- 0 t