ML19208B110
| ML19208B110 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 07/07/1979 |
| From: | Broughton T, Mark Resner GENERAL PUBLIC UTILITIES CORP., NRC OFFICE OF INSPECTOR & AUDITOR (OIA) |
| To: | |
| References | |
| TASK-TF, TASK-TMR NUDOCS 7909190061 | |
| Download: ML19208B110 (49) | |
Text
/
I/
9 p
UNITED STATES OF AMERICA f
NUCLEAR REGULATORY CCMMISSION i
In the Matter of:
2!,
IE TMI INVESTIGATION INTERVIEW b
Of Mr. T. Gary 3 roughton Control and Safety Analyst Manager Si l
6i N
8l Trailer #203 95 NRC Investigation Site TMI Nuclear Power Plant 1 01 Middletown, Pennsylvania 11!
June 11 1o70 12 (Gate of interview) 13!
Julv 7. 1070 (0 ate Transcrip: Typec) 14; 304 and 305
'15i (Tape Nu:r.cer(s))
16i 17!
18f 19i 20!
21 NRC PERSCNNEL:
'22I i
Mr. Mark E. Resner 23; Mr. Anthony Fasano 24 25;
') ; U cvI u
)
t
'7 9 0919 0 gg j
>'/
~
]
,r i
t l
f RESNER:
This is an interview of Mr. T. Gary Broughton.
Mr. Broughton is I
2j a Control and Safety Analyst Manager, and he is employed by the General i
f Public Utilities Service Corporation.
This interview is being conducted in 3
4 Room 203 of the GPU Service _ Corporation.
The mailing address is 260 Cherry gj Hill Road, Parsippany, New Jersey 07054.
Individuals present for this 6
interview representating the NRC are Mr. Anthony Fasano.
Mr. Fasano is an 7!
Inspection Specialist with Region I, of the U. S. Nuclear Regulatory Commissior.
gj Moderating this interview is myself, Mr. Mark E. Resner.
I am an Investi-g gator with the Office of Inspector and Auditor, Headquarters, with the U.
S. Nuclear Regulatory Commission. Also present representing Mr. Broughton 10 for the GPU Service Corporation, is Mr. Alan S. Brown.
Mr. Brown is a
]
supervisor, Generation Administration.
The present time is 8:45 a.m., EDT.
l I
Today's date is June 11, 1979.
At this time I will ask Mr. Broughton if 131 he will give us a brief resume of his experience in the nuclear field.
~15; BROUGHTON:
My experience in the nuclear field began in 1966, as part of the naval Navy nuclear program.
I spent tan years in the Navy.
My tours in the Navy included those on two dif'erent ships.
One tour as an instructor 181 for two years of a land base prototype.
My final tour in the Navy on a ship was as the new construction engineer officer.
In 1976, I left the Navy and came with GPU, and worked in the licensing section.
In 1977, I became the Licensing Manager, and I assumed the position of Control and 22; Safety Analyses Manager which was generated on a reorganization of the 23i company in October of '78, 24j 25i r
b
s I
t I
l l
{
2 l
RESNER:
Thank you very much, Mr. Broughton.
This time Mr. Fasano has some 1.,
2{
questions he would like to pose to you.
3; 4l FASANO: Well, we'd like to have you describe in your own words, your knowledge of the events of March 28, 1979.
And in particular, your noti-5
[
fication and subsequent involvement of the events of that day.
i 7
BROUGHTON:
8!
On the 28th of March, about 8 a.m., Mr. Dick Wilson of GPU, advised me that there had been a trip of the TMI 2 unit, at about 4 a.m.
g Some of the infor ation that he had relative to the trip was there had been 10l a difficulty in consensate feedwater system, which had resulted in a loss 12!
of feedwater, and that had resulted in a trip of the reactor.
The exact cause of the loss of the feedwater was not kncwn at that time.
They had 13l
... at the plant they had suspected there was some additional abnormal 14i conditions, which had existed in the primary system.
Possibly that several primary relief valves had opened.
It would be normal in a case like this, to have only one such valve open.
However, they suspected that there may 17l of been more than one valve opening.
That there was a low primary system 181 pressure following this event.
Again, an abnormal candition.
Safety 19) systems had operated on a low primary system pressura, however, the level 20!
in the pressurizer went off scale high, which was an abnormal condition.
21; And as a result, the safety systems may have been terminatec.
The high 221 pressure injection pumps may have been stopped, in response to the high 231 level in the pressurizer.
There were radiation alarms received in the contzinment.
There were some radiation alarms, which had been received in 25!
s
(
3 the secondary plant.
It was suspected that there may have been a leak in 2
the B steam generator froc the primary side into the secondary side.
The t
3 steam generator had been isolated.
And at this time, we also knew that the 4
reactor coolant pumps had been secured due to low pressure in the primary.
5j It had been reported that there was an attempt to cool the primary system 6
down using the A steam generator.
And an attempt to start the reactor y
coolant pump fn that loop, the A loop.
However, when the pump was started,
{
it did not perform as it was expected, and it was not left running.
It was g
g expected that the off site releases, up to this time at about 8 a.m., had been low. This is the initial report that I had received around 8 a.m., on the 28th.
12!
FASANO:
131 As far as the infor: ration that you had, did you take any actions?
Make any evaluations, or suggestions based on that information, in particular the abnormal situations you mentioned like the high level in the pressurizer?
15 You mentioned that the... nad more than one relief valve actuate.
Was that confirmed? You can go on from there.
But we'd like if you can explain 17i what action you took with that information.
19!
BR0!!GHTON:
Based on this information, we had determined that we should try to obtain more information from the plant.
Hopefully, immediately by t
21l telephone.
And we were preparing to have some of the people from the 22!
Service Corporation go to the plant that day to cather information that would be needed to look into the initiating event anc he course of the 20 transient in more detail.
Rich Lentz, who is in the Control Room Safety 25i Oha
(
b i
l 4
1 1l Analysis Section, was asked to see if we could obtcin more information from 2l the plant, and to standby to go out to the plant later on during the day, 3j t gather the factual data that mignt have shcwn what initiated the event, t
4 and what took place during the event.
We did not make any immediate re-gj commendations from... based on that scan information at 8 a.m.
But.more Si inf rmation did start to arrive ' ner on in the morning, and as more people f
came into work that day, they became involved in the knowledge of the l
Later on that morning, perhaps, at 9 or 10 o' clock, there was a g
gj meeting of several different people, which was convened by Bob Keatin to g
discuss the information that we had to date, and to assemble a team of f
people from GPU to go to the site to provide assistance.
And to assist in f
the... assistance if any was needed directly.
But primarily to assist in the analyses of the event, initating the path, the course of the transient.
14!
FASANO:
At that time you didn't mention a site emergency, or general emergency. Were you notified of the site emergency and/or a general emergency condition?
18i BROUGHTON:
Yes we were, but not at that time.
The meeting that I just spoke of that was chaired by Bob Keating, probably ended at say, 10:30 or And just about the end of that meeting, we received word that high 50.
21!
I iodine levels had been suspected in the primary.
And shortly after that, 22l l
we received word that a site emergency and a general emergency had been 231 declared, and that the dome monitor in the containment building was read'.1g a high level.
This information all came in, I would say before 11:30, in 25i the morning.
i e.
g (
t b~
]
s '
e
t
(
5 lj FASANO:
That would be quite a few hours after a general emergency had been 1
2 called, I would think.
31 4!
BROUGHTON:
That is correct.
There was a definite delay when those events Si took place, and when we heard about them back here at the Service Center.
6i p
FASANO:
Did you discuss this? Did you wonder why that information came so late?
a 91 10f BROUGHTON: We did, but I was not involved in the... into them looking into that particular aspect of it.
We were making preparations to leave for the site.
I think the group numbered, perhaps six or seven.
And I was g
13 f the people in that group, that was preparing to leave for the site.
ne 14i FASANO:
- 0. K.
Well, can you continue now as to when you got... you arrived at Three Mile Island and what subsequently took place.
1U BROUGHTON: Yes.
Rich.Lentz and I arrived at the Visitors Center at about 5:20 p.m., on the 28th.
And when we got there, we were unable to obtain immediate access to the plant, however, the plant staff was gathering up the type of information that we thought we needed to assist in analyzing the transient.
And they were sending a representative with that information 22j out.
And that representative was George Kunder.
And he arrived at the 221 l
Visitors Center at about 6:05 p.m.
He had with him a sequence of events 24l for about 10 minutes after the loss of feedwater.
And since he had been in 25
,3 a
}
l
l
{
t
{
6 i
!lj the Control Room from about 5 or 6 a.m., on the 28th, he also provided us 2
with a brief summary of the things that had been transpiring during the i
3{
day.
4I 5
FASANO:
Can you briefly describe the information that you had available at 6
that time?
7l 1
gj BROUGHTON:
The information that we had was very consistent with the gj briefing that we had received at 8 a.m., on the 28th.
I don't believe that 0
there was any new information there, as to the cause of the event.
The 1
cause of having lost the condensate pump, for example, was still not known.
There were two causes that were speculated.
One was that a problem with I
131 water in the air system; and the second cause that was speculated was postulated was an electrical fault.
But there was no additional information 14!
that would help identify if one or the other was the contributer... the main initiating event.
The discussions about the high pressure injection initiating on low system pressure; the pressure in the pressurizer level 17, increasing to off scale; the fact that high pressure injection was secured g
when pressurier was off scale was confirmed by George Kunder.
They had suspected that they may have had a bubble somewhere in the system other than the pressurizer, perhaps in the core.
l 22!
FASANO:
Excuse me, did he say about what time they suspected this bubble?
24j 25!
l r
Q' j
l
i I
(
7 1.
BROUGHTON:
No, and I'm not sure that the conditions under which we took 2
these notes were all the best.
This was about 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> after he had been 3
there.
And I think that his later statements of what took place are probably 4;
far more accurate information that we learned from this discussion.
It was l
apparent to us, was what we really needed was the actual plant data, from 5
i 6j this time period.
So that we could independently look at this, and try to f
make sure we had the proper sequence, and the proper timing of things.
7 I
8 Based on that, Rich Lentz went back into the plant, probably around 7 or gj 7:30, the night of the 28th.
Perhaps it was even later than that.
It may 10l have been after the reactor coolant pump was restarted.
But it was in that time period. He went back into the TMI 2 Control Room, obtained copies of the post event data summaries, and brought those back out to the Visitors Center, so that we could start to look at actual plant data.
In one case, 12!
he had about 8 to 10 minutes worth of data.
I think in another case, there may have been 15 minutes worth of data.
Based on that data, we plotted up that information, and were able to confirm the fact that the system pressure was abnormally low for a loss of feedwater event.
That at about 6 minutes after the loss of feedwater, the system pressure w'as ecual to the saturation 18!
pressure for the hot leg, confirming that the bubble had shifted into the primary system.
And we also were able to confirm that there was a delay in 20:
initiating emergency feedwater.
This was the first time that we had an indication of that I believe.
22) i 23l FASANO:
You had information on letdown, also?
24!
25i
'30 0 e
s
4 i
f 8
(
1,l BROUGHTON: We had no information on letdown.
We had no information on 2
high pressure injection flow rates.
We didn't have information when high 3
pressure injection pumps were started and stopped.
But actual flow rates 4l through the letdown system, high pressure injection system, were not available.
5l There just not recorded parameters. We were not able to determine when Si emergency feedwater was initiated, but it was apparent that it was not 7j initiated... that there was not emergency feedwater flow into the steam g
generators within the first minute, as what they expected on a loss of feedwater.
g; 10!
FASANO:
11 You determined that based on the data you had looked at?
12l 13l BROUGHTON:
Yes, that was determined based on that data we had obtained at perhaps, 8 p.m., on the 28th.
15:
FASANO:
Prior to that nobody had notified you that it was off for something like 8 minutes?
g 1St BROUGHTON:
That had not been something that was identified in the original information that Dick Wilson had at 8 a.m.
It was not identified in a 20' 21l discussion that we had with George Kunder at about 6 p.m.
And in reading the transcripts from operator interviews and people on shift, it was apparent that those individual operators knew that there had been a delay.
I'm not sure that they knew exactly how long the celay had been.
But this informa-24!
tion was fairly clear that there had been a delay.
We thought initially n
s 9
~
,.y s
l
i l
l i
I.
9 1l that it might have been on the order of 15 minutes.
But additional data 2{
showed that it was about 8 minutes.
I 3i 4l FASANO:
Are you familiar with that... the emergency feedwater system to gj great detail? On the valves, the 11 valves, these are the automatic l
6i control valves of the intergrating control system.
Now, they open up when the level reaches about 30 inches in the steam generator.
I a!
BROUGHTON:
That's correct.
g 10l FASANO: When these valves are closed, are they fully closed or is there
,i d
i allowable passage for water at all times?
12!
13 l
BROUGHTON:
I believe that they are either designed to leak some, or in 14',
practice they do leak scme, but I haven't looked into the details of the valve.
The number I heard is on the order of 25 gallons a minute, as a leakage rate when the valve was shut.
17l 1St FASANO:
What I had, ad its 7 gallons a minute or greater, and of course I 19!
didn't read the final setpoint.
This would mean then if your locked valves were open, you would have started to get water into the steam generators from the emergency feed, at the time that the pumps actually got to their 22l l
full pressure.
23 24l 25; c v ]:
}
10 1
1!
BROUGHTON:
That's correct.
l 2'
i 31 FASANO:
- 0. K.
This is not your area though.
I mean your not systems i
4{
oriented?
5l 6f BROUGHTON: We don't work specifically on valves specs, or system design f
specs.
i Si FASANO:
- 0. K.
But to your best knowledge then, this... you would have g
wa e
me, even e re y u got down
?
10 11!
]
BROUGHTON:
Yes, you would.
I 13l FASANO:
Do you want to continue further, as to your involvement an'd subsequent actions?
16i BROUGHTON: After we had plotted up the data that was available on the evening of the 28th, we requested that the reactimeter, which was a piece 18!
of B&W test equipment, be delogged.
That the taped record on that be rerouted to a paper output, so that we would have additional data for a longer period of time than what was available from the plant's normal monitoring system. And in discussions with the plant, it appeared that that delogging would take place sometime during the night.
And Rich Lent::
and I arranged to be called when the delogging started, so that we could 24) come in and look at additional data.
So, at perhaps 9 o' clock that night, 25!
J c,0
i 11 l
l 1
we'd left the Visitors area, and returned again, maybe 6 or 7 the next 2!
morning. When we returned, we found that the reactimeter had not been I
3 delogged because the plant had not completed the cooldown that they anticipated 4l during the night. And we wanted to make sure that we were gathering transient 5l data, till we had reached a cooldown in depressureized condition.
The 6i actual delegging of the reactimeter, I believe, began later on, on the 29th.
I don't have the times on that. When we did come back in that 7
8l m rning, we reviewed the statements made by the Control Room operators, g
which were taken at, I believe at 3 a.m., on the 29th.
They had come back 10l to the Visitors Center.
The people there had been briefed by them on what had taken place during the day.
And they provided a brief written summary of the major items.
That summary confirmed some of our earlier conclusions.
]
Particularly, that on the increasing pressurizer level, actually had been I
taken by the operators to reduce high pressure injection flow, and reinitiate 141 letdown, and maintain pressurizer level. And there had been a delay in initiating emergency feedwater.
During the rest of the day on the 29th, other people began to arrive to the site, and because I had been involved in looking at this first bit of data, I spent a fair amount of time during the day, briefing new arrivals on what we had known about the transient up to that time.
At about, probably 2 or 3 p.m., Dick Wilson and several other people from GPU arrived, and after briefing them on the transient, 21!
Dick Wilson met with us and some other people on a designated task force to 22l discuss our various roles on determining the causes of the accident.
And 23i documenting those causes. We split up into different groups.
One of the 24l major parts of the event, we attempted to get underway as soon as possible, 25!
$,c I
[
i U
I f
lj was interviews of the operators who had been involved in the plant operation 2
during the period 4 a.m. on.
And those interviews were scheduled to start 3l at 11 p.m., on the 29th.
I believe, during the day, on the 29th, we were i
4 able to start getting some reactimeter data, and had some of that data Si pl tted up.
However, the first bit of reactimeter information that was l
p1 tted only ran out 8 to 10 minutes.
So, for many things it provided no 6
additional information beyond what we already had.
It did provide some
/j f
information in a couple of areas that we did not have information.
For g
g example, the reactor coolant drain tank pressure is a parameter that is 10;;
recorded on the reactimeter that is not recorded anywhere else.
So that was an indication that there had been a discharge from the primary for...
12l I forget how long the first draft was, but I'm sure it outdates 10 minutes, indicating almost a continuous discharge from the primary, for that period 131
- 141, of time, which was abnormal, following this type of event.
And about...
probably about five or six p.m., on the 29th, I again left the sita to waiting for additional reactimeter data to be delogged.
And that was i
supposed to be taking place that evening, or during the night.
The hope 17!
was by the next morning there would be at least a half an hours worth of 18f data, so we could extend our knowledge of the transient further out at the 191 time.
Do you have any questions up until this period?
20t 21!
FASANO: Well, only that your reactimeter data was delogged on the 29th.
22l That meant all of the information on the reactimeter was availacle after 23i the first day, but not during the first day.
Is that correct?
241 2si p
c S' i
i f
[l.
13 1l BROUGHTON: Well, when the reactimeter is delogged there are 24 channels.
2l All 24 channels are dumped at once.
But, it's only delogged for a certain i
3j time period.
I believe, on the 29th, it was only delegged out to perhaps 8 4
or 10 minutes.
It may have been that it was delogged all the way out to 5l 20, and the only thing we could do was to plot it.
It was had plotting i
6j required that data maybe only... the plots only ran up to 8 to 10 minutes.
7j But at the end of the first day, I'm sure that we did not have longer than 30 minutes worth of reactimeter data. We certainly didn't have plots that g
i went out beyond that time.
g; 10t FASANO:
So for the operator, this would not be available for him during the event.
In no way would this be a help to him.
Is that true?
13I l
BROUGHTON: Well, the reactimeter is not at all useful to the operator.
14l It's not a... it's a special piece of test equipment which was in during 16l the start of test program that you use to gather information wnich is becomes part of the startup report.
It's a non-operational piece of equipment 17l We have kept the reactimeter in service at TMI 2 because of the monitoring 1Sf capabilities that we have on that plant, makes it difficult for us to look at as many things as we'd like to for as long a period as we'd like to, and l
as much detail, as we'd like to.
And based on severai events during the 21i test program, which were off design type event, we found that the plant's 2f monitoring system, was not... didn't provide sufficient information to the 23l l
allow a detailed analyses of what had taken place.
So based on that, the 24; plant elected to keep the reactimeter in service until we'd had evaluated 25!
changes and made changes in the plant's normal monitoring system.
n
", \\ )
I
- k. tlyt' i
~
a i
{
14 lj FASANO: Would you care to expand a little on that.
What parameters, and 2!
would these be essential ones, or knowledge of events and opertor action?
l 3!
BROUGHTON:
Does this regard to things,that are monitored on the reactimeter?
4 Si FASANO: Yes, that are not monitored elsewhere.
6 i
BROUGHTON:
gj There is basically nothing monitored en the reactimeter that's not monitored on the plant already.
It's a question of what kind of a g
record is kept.
For example, drain tank pressures... there's no premanent 10 record kept of that.
But if you elect to monitor that on the reactimeter, you can't keep a permanent record.
t 131 FASANO: Alright.
15:
BOUUGHTON:
And you can also monitor things on the reactimeter in a much more rapid frequency.
For example, one second intervals on the primary 17!
pressure where the post trip monitor monitors, let's say 15 second intervals.
1SI 19f FASANO:
So, on the post trip monitor could you have had the reactor 20l coolant drain tank pressure monitored... or at least data availaole to go 21:
out?
22!
i 23l j
8 ROUGHTON:
No, the plant's post trip monitor would not cover drain tank 24!
pressure.
It does cover the major parameters of hot and cold leg temperatures 25i pressurizer level,..
system pressure, power.
_ n
h\\J i
~
q
s s
l
{
15 l
1!
FASANO: Well, what I'm trying to get at, how would the operator know that 2
-he was having this more than ususal ficw into his reactor coolant drain 3j taak?
- s 4I
>},
BROUGHT 01: 'rie has prd:sure, temperature and level indication on the drain 5
r tank.
' 6; 7!
t 3
s gj FASANO: Availab}a to him?
i i
91 BROUGhTON: Available,ta ab a a panel in the Control Room and there is 10i also computer monitoring of, I believe both level and temperature, and g
curing the course of tne event, he di f see temperature changes on the f
ccmputer.
So, the rea.:theter does noi.hing more mc.iitors no points that are not already available for monitoring by soma other system.
r 15; FASANO: What locatial is this readout for J.'im, as far as the readout on 17j the reactor coolant draining systers.
Is it on his main panel?
i ISI
~
BROUGHTON:
It's on a... what we wetid consider to be 'the back of the
- 191, pnel.
He has to leave the normal control sectiqn of the Control Room and 20i walk around to the back of the panel.
If fact, op'e or more operators did 21l that in the early part of this event, to monitor that tank.
29' t
23l 1
i FASANO:
So, that they would hive had knowledge that they had an unt.sual 24j rise in pressure continuous?
s 25!
- yJ'
'i
~
6 y
( '
s s
o I
I 16 i
lj BROUGHTON:
I'm not sure it was monitored early enough to determine that 2
the pressure was abnormally high. One operator that did monitor it, concluded i
3{
that it was likely that they had blown the rupture disc on the drain tank.
This was a... I'm trying to recall this from a previous interview, but it Sj was probably within the first half hour that that Si FASANO:
So they would have had information within the first half hour that 7
they had something unusual with regard to the reactor coolant drain tank?
8 91 10:
fact that an operator looked a'.
that and evaluated it as a rupuure.
The g
drain tank rupture disc band,t artainly.
13!
FASANO: Were you familicr with Unit 1, as far as occurrences such as this?
Is this really unusual to happen? A reactor coolant tank disc rupture, or is this something that's happened?
17l 1
BROUGHTON:
I don't know of any other time that it's happened.
That 18(
doesn't mean that it hasn't.
I don't know of any other reasons.
I do know where other instances where the relief valve... the prirary relief valve has stuck coen for an extended pericd of time.
As a matter of fact, there 21; was an occurrence on TMI 2 about a year before this accident, in which the 22!
valve stuck open for 4, 4 1/2 minutes.
23) i 241 25!
Y i
c u s
i l
l 17 i
1!
FASANO:
Do you recall the date of that?
i 2!
i 3
BROUGHTON:
Yes, it was 3/29/78. And I'm reasonably certain that the 41 ruptured disc was not broken on that event.
I 5!
FASANO:
Do you know if there was an evaluation on the cause of having that 5;
71 particular valve stick open?
l 8!
BROUGHTON:
Yes, it was a loss of power to, I believe, one or more of the g
vital power supplies was lost, due to the design of the circuitry, loss of 10l that power supply caused the valve to open.
12l I
FASANO: And remain open?
- 131, 14)
BROUGHTON:
And remain open.
15:
16; FASANO:
So it didn't stick open, it just remained open?
17,;
ISI BROUGHTON:
That is correct.
And then when power was restored, the valve snut.
It was either the restoration of power, or perhaps closing the bar valve, I don't recall in enough detail to say.
22 FASA@:
I guess that was a i.icensee Event Report, I think, we did receive
,3l c
a copy of that.
24) l 25j
(\\
.3 l
r S ',
l
I 18 1!
BROUGHTON:
Yes, I'm sure that was an Event Repcrt.
2:
FASANO:
You do analy:e these Event Reports? Just off the subject, slightly.
31 4j Si BROUGHTON:. We don't normally analyze Event Reports, unless in conjuction 6
with the evaluation, there is some analyses required.
7!
FASANO:
Such as an ECCS actuation?
8!
Si BROUGHTON:
10:
An ECCS actuation due to low pressure, is not something that we could necessarily analyze.
w l
FASANO:
Do you want to continue?
I have some other particular questions, but I think if you could continue...
14!
15i 16:
BROUGHTON: Alright.
On the evening of the 29th, this would be somewhere 8 o' clock or later, I was also involved in briefing the people who had just 17; come to the site, on what had taken place.
And, during those discussions, 18t we received information from the Control Room, that there might be a problem with gas concentrations in the primary system.
There was a concern about the waste gas system, being near capacity.
And as a result, several of the 21!
people who had come in that day, went into the site, and one or more went 221 into the Control Room before midnight on the 29th.
I think my n' ext involve-23!
ment at the time, would have been the following morning, Fricay morning.
241 And I believe, that when we came in, at 7 or 8 on Friday mornirig, that we 25l Ai
) '
t
.~
{
19 i
1:
had additional plots of reactimeter data.
Perhaps, at the thirty minute l
point. And it was our intention to begin more detailed anlaysis of the 2
3 event that morning.
However, we began to hear about additional events q
which we had no knowledge of before.
Scme of the people who had been g
conducting interviews during the night, for example, had learned of a 6
p ssible hydrogen explosion in the containment on the 28th, around 2 p.m.
7 That was I believe, that came out of a discussion in an interview.
I g;
believe, it also came out of the some of the data which had been plotted up during the night.
It was not reactimeter data.
It would have been data gy available on other strip charts, which had probably been taken out of the r
]
plant during the night.
12!
FASANO:
You say during the night, that was the 29th?
131 14:
BROUGHTON:
Aright.
From 0001 on the 30th up to 7 or 8 o' clock in the 15, morning.
Exact source of all that information, I'm not sure, but there were several different sources that indicated the same thing.
181 FASANO:
Excuse me, the first time that the idea that there was hydrogen and eventually hydrogen explosion, as far as you knew, On the 30th?
21!
BROUGHTON: That's as far as I know.
I'm sure other people had been aware
.2' of that during the night.
I'm not sure what recogni::ation there had been 23i of that on Thursday, the 28th, anongst the GPU people that were there the 24!
Non Met Ed people.
It also became apparent early in the morning on the 25i 28th, the N,\\"
h
~
i 20 i
l!
FASANO:
Excuse me, the 28th?
i 2!
3{
SROUGHTON:
The 30th, I'm sorry.
Friday, the 30th, that some acticn had to 4j be taken with regard to gas in the primary system.
That the waste gas g
system was at or near capacity, and that the continuous letdorn would 6l continue to aggrevate *. hat situation.
An;. there was the activities in the tanks, that were too high to allow discharge.
So, I began to become more f
involved with the plant operations.
And the discussions with people at the g
g plant and with Dick Wilson and others from GPU were there, we began to look 10l into ways in which we could degasify the system alternate ways to degasify the system, other than the letdown.
i 12l
{
FASANO: Degasify the primary system?
i 141 BROUGHTON:
Degasify the primary system.
So, very early in the morning on Friday, probably 9 or 10 o' clock, I guess that's not early.
Perhaps, 9 or 10 o' clock Friday morning, we began to strongly recommend that the letdown system be secured to prevent putting additional gas into the waste gas 18t system.
And that steps be taken to begin to degas the system through the
, 91 1
pressuri::er. About this time, it was determined that additional technical 20l support on a continuing basis was needed to assist the Met Ed staff.
We organi::ed ourselves into two shifts, to provide continuous support.
We had people up in the Control Room, that was our direct communication link with 2 31 i
the operating staff. And we had people located onsite in the Station 24l Superintendent's trailer, to be our communications link with the Control 25i
\\
t J
a l
l i
I l
l 21 i
li Room.
And then offsite to B&W and GPU facility here at Mountain Lakes, and g{
Met Ed, Reading, and anywhere else we needed to conduct those types of i
3j communications to get technical support.
So, we got into that organi:ation, certainly before noon o'n Friday.
And I think to some extent that type of 4:
Si support had started the night before when we sent some people in to assist 6
in the question of the gas, prior to the 29th.
And at about, I believe it 7i was about 2200 on Friday, the 30th, we did begin to degas the primary g;
system through the pressurizer.
9f FASANO:
101 Had you made the estimates as to the content or the quanity of noncommencables?
g 12' BROUGHTON:
I did not, but those estimates had been made.
And they were made around midnight on Thursday night.
I believe, it was Bill Lowe and
,4 1
Jim Moore, wr.o were in the Control Room involved in that procedure. Jim g;
Moore is the next person you'll be talking to.
There was some quantative 17;
... first of all there was a qualatative determintion that they were not condensables in the system around midnight.
And very shortly after that, 18i there was some pressure changes made in the system to try to evaluate the
,95 1
magnituce of that noncondensable bubble and that fairly early on Friday
,0 c
morning, they had some numbers for that.
22l 1
RESNER:
At this time, we'll take a break to change this side of the tape.
23I It is now 9:28 a.m.
24; 25!
y c%\\
l l
l 22 l
RESNER:
This is a continuation of the interview with Mr. T. Gary Broughton.
2{
The time now is 9:29 a.m.
l 31 f
4j FASANO:
- Gary, Si BROUGHTON:
I finished my last statement.
6 7!
FASANO: Alright.
Does that finish your recall up to the 3 day period now?
g 9I BROUGHTON:
O That would cover Thursday into Friday.
11:
]
FASANO:
Maybe we could go back and then cover some of the initial information you had give.
As far as your information, did you have... you mentioned that you had abnc emally low pressure in the primary system.
How did you arrive at that? idow this is at 8 o' clock in the morning on the 28th.
16, BROUGHTON:
Yes, that was a report that we got. We had no data to look at 17!
and arrive at our own conclusion as to what had taken place.
That was merely the report that we had received from the plant.
20; FASANO:
And I guess that you also stated that there may have been more 21l than one relief valve actuated on the pressuri::er?
I 23!
BRCUGHTON:
Yes, that's information that we got from the plant.
We did not 24!
make that evaluation ourselves.
And as a matter of fact, we found no 25:
o
^\\)
J'
( O. k i
6
l 23 l
1l information that would support that more than one primary relief valve 2
never did open in the transient.
We did determine that there had been 3j leakage from one of the relief valves, the primary relief valvu, prior to y
the event, that may have been on the order of 6 gallons a minu+.e.
Exactly
{
which valve that was, I'm nt.; sure that that's been determined
- Mether it c
6 was the electrically operated valve or the two closed valves.
i 7
FASANO:
Has it been subsequently determined, or is it still an unknown?
8i 91 BROUGHTON:
I do not know the answer.
But, I'm sure there is someone who 101 g
does.
12!
FASANO:
Your initial information also talked about the high level in the g
pressurizer, which was unusual, also.
With this information, the initial information, were you able to make any conclusions as to where you could have provided operations with any helpful conclusions, or suggestions?
16i i
17!
BROUGHTON:
We had some other people here at GPU who remained here.
And established an open line with the site to do those types of things.
And exdctly what time they started doing that, I'm not sure.
I believe it was around noon or shortly thereafter.
And some of the people that stayed here were in a position then after gettting quantative assessments of the plant, these were all fairly qualatative.
When they got some qualatative assess-23!
ments, they were making specific recommendations, as to things that snould 2 41 I
be done.
For example, charging to make the system solid.
That recommen-25i dation was made from here sometime the afternoon of the 2Sth.
., 1 a
,n
./
o I
i
{
24 I
I 1:
FASANO:
That would be... you were down on site, so therefore, you weren't 2
here...
i 31 1
4; BROUGHTON:
I was on my way to the site.
Si FASANO:
Did you get information as to the reactor coolant pump flow 6
7 degradation? You mentioned that they told you the pump was turned off.
g Was there any question as to why they were turned off? And as anyone reviewed the reactor coolant pump degradation and what they should have g
someone?
10 11!
BROUGHTON:
Our initial information as to why the pumps were turned off was
,3(
due to icw pressure.
That was what the first report dealt with.
1 14; FASANO:
Nothing to do with flow?
-15i 16i BRCUGHTON:
No.
Nothing to do with flow.
18(
FASANO:
If you had the information with regard to ficw, would this hdve alerted you to anything?
21l 22!
SROUGHTON:
I think it would have alerted us to something.
But it would 23l certainly be a secondary indicator of an abnormal condition in the system.
24l We did not go after that indicator immediately, and so I don't even know 25i
{-.
[
25 t
1:
that's
... I don't think that was picked up on the post event monitor.
2j So, we really would not have had any flow data from the primary until we 3
delegged the reactimeter.
It would have been about the 29th.
4j 5
FASANO:
Would the operator have this information available to him?
6i BROUGHTCN:
Oh, yes.
The operator has that information.
And it was one of 71 f
the things that he was looking at.
He also has vibration information.
He g
was concerned about that. And generally, he was concerned acout net g;
10f p sitive suction head.
u FASANO:
But this information was not transmitted to you by 8 o' clock, on the first day?
14!
BROUGHTON:
No.
We were a little confused about the information we got on the pumps. We understood that the pumps were shut down due to Icw pressure.
?
17i FASANO:
Low pressure.
What pressure?
19i SRO' GHTON:
Low primary system pressure.
Don't know. We had centrally no J
,40 numbers attached to this description.
21:
22f FASANO:
Now that you know the numbers, what pressure would that have been?
23!
24!
25i
<V'-
- \\ 3
,+
i
]
l
\\
l 26 i
1!
BROUGHTON:
Well, knowing the numbers, it now looks like the net positive 2
suction head for the pemp was... that limit was violated about 6 minutes 3j after the trip.
So, that if checks had been re.ade to determine adequate net 4
positive suction head at that time, or anytime subsequent to that, it would Si have shown that inadequate, as far as the suction head.
This was not the 6
nset of a low net positive suction head was not late in the development of f
the transient.
But the action that was taken with the pumps was, I think aa hour and 15 minutes, was whan the first pump acticn was taken.
g 9I FASANO:
g Have you seen previous... well, you mentioned that you did see one, back on the 29th, the year before, the 29th of March, 1978, I guess 11:
you must had a traumatic pressure drop there, too.
12',
13l BROUGHTON:
Yes, there was a traumatic pressure drop.
The initial conditions 141
,l were different, in that that happened during zero power physics testing.
15 So plant was in hot standby, at 532.
And they were at normal operatir; pressure, when the valve opened the pressure dropped d out 200 psi / minute.
And the pressure dropped, I think it was checked about 1,'00 psi.
19i FASANO:
I see you have an event in April, also, where you had ECCS actuation.
20f 21l BROUGHTON:
Yes, and in April of '78, we had a stuck open safety valve.
22l And we did perform a very detailed analyzes of that event.
To make sure we 23 understood the initiating event for the cooldown.
And that there was 24) nothing else that took place during the transient, that either contributed 25i Ql
- q. A]s J
4 27 i,
1!
to it or had been obscured by it, the major effect of the valve sticking.
l So, we did a fair amount of analytical work relative to that event.
2 3i FASANO: Were your pumps violated then, also in positive suction head?
Si BROUGHTON:
I don't believe that it violated the net positive suction head 6i curve, because we were cooling down quite rapidly.
And the pressure reduction gj that took place, was small relative to the cooldown that took place.
I g
believe that in all cases, we remained in compliance with the net positive 10; suction head.
There other tech specs limits that we did violate.
Fuel pressurization and some others, and they were all looked at, most of them lli
]
had not been looked at by us at GPU, but looked at by B&W.
13 FASANO:
I guess you had two more events, a year maybe three, November and l
December. We had ECCS again, at low pressure.
Did you look at those also, the License Event Reports?
i 17l BROUGHTON:
We got in.olved in those, but we did not do any detailed analysis on them. We got involved in changes to the logic for initiating sodium hydroxide on an ECCS.
21l FASANO:
That would be the first one and the second one, right? The March 22' 29th, and your April...
23 24I 25i
.),p>
l l
28 l
lj BROUGHTON:
In April that sodium hydroxide injection.
Yes.
I 2!
l FASANO:
3; That's been changed, so that you had sodium hydroxide only going i
4!
in at low level on the BWST or...?
51 BROUGHTON: That is correct.
It provides a time delay between the initiation 6
7 of the ECCS and the initiation of the sodium hydorxide.
8l FASANO:
g Were you given information on your Th at all, being off scale, during that first report at 6 a.m.?
10 11; i
BROUGHTON:
Not at 8 a.m.
And I'll check here briefly to see if that's 12!
l something George Kunder mentioned.
I don't think that it was.
I have a 131 note here that the cold legs were low and the hot legs were rated at 700.
l This may nave come from a discussion we had with George Kunder, or someone 15i else in the Visitors C nter.
161 17l FASANO:
This would be about... after 5 p.m. then?
181 191 BROUGHTON:
After 5 p.m.
As a matter of fact, I think that the sequence 20!
that we got it in was that it was after the bulk of the radiation alarms had come in, and after they had declared a site emergency.
225 I
23l l
FASANO:
But, I mean you got the information late in the evening?
24j 25[
1)9 \\
c>'
1 i
i 29 i
lj BROUGHTON:
Oh, we got the information at 6 p.m., or later.
I 2!
FASANO:
How about the A cord thermocouple readings? Do you have any g
4j knowledge that there were checks being made early in the morning on the
[
thermocouple readings?
6i BROUGHTON:
No, I did not have that information.
8!
FASANO:
You mentioned the steam generator leakage.
Did you get involved 9(
at all as to the cause? Or did you get further information, as far as the steam generator problem? You did mention concern about off gas, could you 11:
explain your involvement in that area?
f 131 BROUGHTON:
Yes, we looked into that a little bit.
Again, we didn't really have any data to draw sound conclusions on.
One of the things that we postulated was knowing that there was a delay in initiating the emergency feedwater, that perhaps, the initiation of emergency feedwater, may of been enough of a thermal shock to the generator, combined with other things, 18l that perhaps that that was the initiating event for a tuce leak in the 191 generator.
But that sas about as much thought as we were able to give it, 20t without having some more detailed information.
21; 22l i
FASANO: When you say other things, wnat other things do you mean?
231 24!
25j
't 9 6-c0!
i
I 30 i
1!
BROUGHTON:
Perhaps abnormal... We knew that the pressure on the secondary 2
side of the steam generator was low, as a result of boiling it dry.
And 3j the... we did not know what the primary side pressure was at that time.
4{
It could have been a situation where there was more... much more primary gj pressure, than secondary pressure.
The emergency feedwater system does Si come into the spray, directly into the tube bundle area, rather than having i
7 any preheating in the downcover.
So everytime you do initiate T.hrough a gj dry generator, that is a shock... it's a known shock to the system.
It's g
an event to be minimized in frequency.
10!
FASANO:
The mitigation of that shock, if you had your control valves opened, would this be minimized?
131 BROUGHTON:
I would certainly think so.
Especially having some leakage through the control valves, you still have steam passing through the generator, during that part of the transient.
Since that time, we have extracted from the reactimeter, the data surrounding the time when the B 1,/ !
coolant pump was restarted at about 6:45 or 6:50, in the morning.
191 20!
~FASANO:
The B coolant pump?
21i BROUGHTON:
Yes, the B reactor coolant pump, 0.K., was started and run for, 22' I believe, 19 minutes.
23{
And that whole sequence in there closely... within a very short time frame, there is a dramatic change in primary and secondary 24) side pressure.
There's a dramatic change in readings on many of the secondarj 25i n,
s cy '.
l l
l 31 l.f plant activity monitors.
And that certainly looks like an area which 2j should be investigated more fully, to determine if that caused a transfer l
3; of activities from the primary to the secondary. We pulled that data out 4j now, and that's data which will be looked at, with regard to finding out Sj what initiated the transfer of activity.
There have been several other i
6 mechanisms proposed in which activity could have gotten into the 8 steam 7
generator. We have not completed looking at all those proposed ways.
I 81 FASANO:
g So as far as leaking steam generator at this time, is inconclusive?
10t BROUGHTON:
I'd say it was still inclusive as to how the activity got into g
the generator.
FASANO:
But you do think it's damaged?
-15; BROUGHTON:
I couldn't say.
My personal opinion is that there was a i
primary to secondary leak.
But, I'm not in a position to support that with 17l' data, although the data exists, it just hasn't been analyzed.
lSt 191 FASANO: Were you notified... well, I guess, subsequently to that there must have been some discussion on what a hammer event which happened somewhere 21!
in the early five minutes, or so.
Was this... were you knowledgeable of 22l this?
23!
And has this been looked at with the regard to the steam generators, or condensate system?
24!
25l N}?
s a
C
\\
~
i
(
32 t
1.'
BROUGHTON:
People have been looking into tne water hammer question.
I did 2j not hear about the water hammer until quite late in the discussion of the whole event.
I think the first time that I heard about it, was either 3
4 reviewing a transcript, or perhaps reading the annotated sequence of events.
Si FASANO:
Do.you know if anyone has to come to any conclusions on that area?
6 You have not?
81 BROUGHTON:
I have not.
One of the people who works for Gary Capadono, of g
the GPU Service Corporation has looked into that scme.
Vern Monroe looked into it a little bit.
And there's an attemp being made right now, to pull all that information together, and draw some sort of a conclusion.
There's
]
still differences of opinion amongst people, who have looked at it, as to what could have happened, and what did happen?
,4 1
15.
FASANO:
In your early information at 8 o' clock, did the operations personnel talk to you at all about their low levels on the acron sarp'es?
IS:
BROUGHTON:
Yes, George Kunder mentioned that.
I neglected to discuss that when I went over wnat he briefed us on.
They did see an increase in the 201 intermediate range power.
And that was in the period of, I believe sometime 21:
between 6 o' clock and 6:45, the morning of the event, this is based on what 22!
I got George Kunder looking at that in more detail now.
We can explain 231 much better what those pulses are.
But, from his preception, they suspected 24i that they might have had a boron dilution accident.
And the sample for 25i l
a f
i
I I
i I
i 33
!lj boron and came up with some low numbers, I believe, on the order of 400 to 2
700 ppm boron.
Those sample results combined with the intermediate range, 3
power increases resulted in them emergency barating.
They did not mention increases in source range level, when I first talked to them on the nignt 4i ej of the 20th.
And my recollection of when the sample results... well, the 61 discussion at that time, was not detailed enough to determine why the pcwer 7l range indication may have come down before they got the boron in.
But we g
did not explore this in great detail, because we didn't have the data.
We q
thought we should take a good look at the data and try to assist them.
10j That the action emergency borate was effective.
If it wasn't effective, at least it was not detrimental.
I guess, that's the way they looked at that.
11, y
We did not at that time correlate the attempt to run the B reactor coolant 13 pump with the reduction in intermediate and source range power levels.
14!
FASANO:
g Had you subsequently done that?
16 SROUCHTON:
We ha e subsequently done tnat.
It was just our inability to v
get our hands on real plant data at that time, to deha into this much further.
Several of us who looked at.. who knew '_ hat the intarmediate e
e ee 20!
nc ease, though, suspected that was dae to core voiding rather than a boron dilution accident.
And I think that subsequently shown to be the most likely cause.
22!
i 2 31 FASANO:
You say these detector are... what... three feet in length, 24l approximately?
i o-
{
34 l
BROUGHTON:
I'm not sure.
There were at least two different traces.
One j
being ther source range, and one being the intermediate range, which showed 2
similar effects.
And there located slightly differently, and they are I 3
gj believe slightly differant detectors.
We have sama pecple here that are f
looking into that.
There in Gordon Bond's nuclear fuels group.
g 6i j
FASANO:
This wouldn't be your area, then?
i 8!
BROUGHTON:
g No, we're not looking into this directly as to the magnitude, and so forth.
O!
We are trying to correlate what we're seeing in terms of power level with other events taking place in the plant, assuming that that's an indicator of voiding and there's a good correlation.
l 13!
FASANO:
You mentioned Gordon Bond. Was he involved at all icoking at this 14:
the first stages? Or just now...
1.
1Ci BROUGHTON:
I'm not sure how much he was involved in it.
I certainly know he had some involvement, but to what extent, I don't know.
He remained here at this office for the bulk of the time, and he had spent some time at 19i the site.
20i 21!
FASANO:
- 0. K.
So, you learned about.
0.K., did you correlate to reactor 22l pressure, or to... in starting and stopping of the pumo, as far as the 2 31, changes on your counts?
24'.
2Si p
<'y I
3 I,
[
35 1
BROUGHTON:
Not at this time, because we didn't have the data.
Subse-2:
quently, we have. We think there's a good correlation between times when f
we either had no core flow, or had a good chance for core voiding.
And the 3
4 increases in the power and source range indication. We could pretty well
)
correlate decreases in those indicaticns with actions that were taken to 5
Sj produce flow through the core.
Like starting a pump.
Like initiation of high pressure injection.
Like operation of the electromatic relief valve.
gj Our analysis in that area is still not complete.
9l FASANO:
- 0. X.
OI Did you... at that time did you come up with any conclusions as to the cause of low baron, I mean, operations apparently gave you a...
their opinion... has... does this agree with what you concluded?
l 13!
BROUGHTON:
Well, at the time we were talking to them, which was about six 14' p.m., they no longer had high levels on the source and intermediate ranges.
They had gone through many evolutions during the day, which would have put additional borated water into the plant.
For example, discharging the core i
171 flood tanks, partially.
And they had also charged a substantial volume from the barated water tank, during the day, which was at a high level of boron, additionally.
It was on the order of... it was greater than 2,000 ppm, in the borated water tank.
So, when we got this information at 6 p.m.
2 11 on the 28th, we were not concerned about the shutdown condition of the 22!
core.
23l 24i 25!
G r
.,v 7-
6 l
l 36 1:
FASANO:
Has there been any subsequent review of cause for giving such a 2j 1 w baron concentration?
3l BROUGHTON:
I don't know who's looked at that detail.
The plant's suspicion 4!
Si was that they may have had some flashing in the letdown system.
Si 7j FASANO:
They take a samole within two hours after shutdown.
Right?
i Usually.
Did you have any knowledge as to when that sample was taken? Or g,
gj do you have any knowledge?
If I remember right, it was about 75 minutes after, that they got the results.
So that's about 5:15, or so?
0l u:
}
BROUGHTON:
They may have started to take it that early.
I 11!
I FASANO: Well, the results, I believe, at least we have an indication that 14' the results were in about that time.
15 16; BROUGHTON: About 7 o' clock is the time I remember.
171 18t FASANO:
7?
19I 20:
BROUGHTON:
This is not an aree that I've looked into the most.
Not much 21, detail...
22.!
i 23!
l FASANO:
- 0. K.
I was just trying to fix a better time, if we could.
24,1 25l
., q,
3" oi
( 'J
i I
37 ij BROUGHTON:
I'm sure that's in the transcript of someone like Bill Zewe or, 2j perhaps, the man who actually took the analysis.
31 i
4j FASANO: Well, do you have any other things you would like.to bring out, as gj far as the... your involvement, at this time?
1 6i BROUGHTON:
I don't believe so.
7 81 FASANO:
g Mark, do you have anything else you would want to discuss at this 101 n!
RESNER:
Yes, I'd like to go over the two page advisement document, which y
Mr. Broughton was given at the outset of the interview.
This document explained the purpose, the scope and the authority with which we conduct
,y this investigation.
And it also apprised Mr. Broughton that he is entitled to a representative of his choice to be present during the interview.
And on this particular occasion, Mr. Allan Brown, whom we introduced earlier, 17,1 has acted as his representative.
And in no way is he compelled to talk 18I with us, should he not want to.
There were three questions on the second page of this document.
And I'll verbally confirm those for the record.
Question one, do you understand the above? Mr. Broughton has checked yes.
Is that correct, Mr. Broughton?
23l BROUGHTON:
Yes.
24l 25i 3
^
O,\\1
I i
l
{
38 I
i 1!
RESNER:
Question two, do we have your permission to tape the interview?
2{
Mr. Br ughton has checked yes.
Is that correct, Mr. Broughton?
l 31 y
BROUGHTON-Yes.
Si RESNER:
Question three, do you want a copy of the tape? Mr. Broughton has 6
checked yes.
And we will provide him with a copy of the tape through Met Ed's representative, William Burlee, at sometime in the near future.
Is 8
that correct, Mr. Broughton?
g 10f BROUGHTON:
Yes.
12!
RESNER:
- 0. K.
If no one has any additional comments, this will conclude the interview with Mr. T. Gary Broughton.
15-FASANO:
This time I would like, if you would, take some time and give some
... well, express some of your ideas as to how a Service Company might have been more helpful, or in the future, what problems you had in dealing with 18t your involvement, and if any... if you want to bring up any lessons learned, 19!
so that we in the NRC can be helped... you and industry can be helped, and 20!
others, possibly, could be helped from wnat we've learned in experience?
21i 22l BROUGHTON:
I'll confine my remarks to how we might have been able to 23!
assist better in the early stages of the event.
One of the major problems 24l is trying to provide assistances and having high quality data from the 25i
., % x DJ
.1 m
c i
I
I f
39 lj plant, in a very short time delay.
Several of the events that we've gj analyzed on the operating plants, we have... there have always been some f
unknowns in what took place in the plant, simply due to the fact there was 3
4 some parameters which turned out to be key that were not part of whatever g
monitoring system they were using.
This particular event, for example, the 6i inventory in the primary system is a key item.
And there is no monitoring f
of... nothing recorded permanently with regard to the letdown flow, high 7
l pressure injection flow, makeup flow.
The things that normally would cause g
gj you to be changing inventory to primary.
So that makes it very difficult 10f to accurately determine exact.'y what took place, so you can find some of the real unknowns, like void frictions in the system.
Things that are not possible to 13; 141 15i 16l 17l 181 191 20i 21:
22l i
2b 24i 25i
'i u I
I l
40 i
lj monitor.
So, I think from the standpoint of how to bring more technical 2j expertise to bear, in a short period of time, that certainly one of the 31 keys is in the providing high quality data in a short time delay.
4!
5; FASANO:
Just to go back a little bit on the inventory.
The information i
6 available to the operator and to any of the staff members that work within 7j the Control Room, at that time, now would you say it would require inference 8
n the part of the personnel in the Control Room to come up with an inventory on the problem?
g, 101 BROUGHTON:
I think from the standpoint of operational instrumentation, that there are always some improvements that can be made.
The basic I
operating parameters of the plant were available to the operators.
The 131 were sufficently well displayed that they had knowledge of them, and based on my understanding of what they have said in their transcripts, they understood what the values of the basic parameters were.
For example, ;he abnormal pressure which existed was known to the operators, and it was 17; recognized as being abnormal.
Ar.d that in itself, should be... an operator 18:
should be able to use that in conjuction with these procedures and under-19!
standing of the plant to take actions, that be in the direction of restoring 20t that pressure without having to infer some of the more detailed or perhaps, 21!
second or third order questions, like what is the cause for the abnormal 22 I
pressure.
If he needs to get into an inventory balance to determine that, 23t I
that perhaps he could be provided wi+h an indication that would eventually 241 allow him to do that.
But, to base a decision to raise the pressure on 25i
'h-)
(~s
i f
l 41 1:
information that might come from an inventory balance, I don't think is 2
necessary, or desirable.
3l 4
FASANO:
One of the parameters seems to be the pressurizer level.
You did gj mention that it was abnormally high.
And apparently stayed high.
How...
6 would you ttiink the operacor training needs to be changed, or improved, or f
whatever? With just 'mprovement of indications, one would then be better ff? You know, be able to cope with the problem easier?
8 9!
BROUGHTON:
No, I don't think that indication was a factor in... in the 10:
events that took place at TMI.
I think the indication was adequate to j
indicate what action should have been taken.
For example, the procedure
]
which addresses low primary system pressure, requires that the high pressure injection be..dtiated, if both pressure in the system is... in stage 2 condition, it says you have to have pressurizer level above the low level alar = point. And you must also, have system pressure above the low level alarm point. Which is 1,600, 1,640 psig, and then failing to having both of those parameters above those alarm points, indicates there's a break in the system.
So, even in the failure of one instrument, for example, if the pressurizer level instrument had been reading abnormal, it was probably reading normally.
But, even if it's been reading abnormally, the system 21, pressure cannot be maintained, then that's still an indication of a break 22l in the system.
I don't think that indication in itself is an area that is 23l l
a prime contributor to the event.
I think the training of the operating 24!
staff, to ensure that they take action to correct the anomalies.
If they 25
.Q s J
(G\\
e
l i
42 I
lj are unable to correct them... for example the staff was unsuccessful in 2
controlling pressurizer level.
They made attempts to reduce it, but they
)
never really exercised control over it.
I think of the fact that they 3
4 lacked control of the pressurizer level, the system pressure was abnormally Sj 1 w, and their attempts to raise it had been improper.
And had that type f
f inf rmation been relayed out of the Control Room to people who provide 6
q technical assistance, the basic accident would have been identified a lot f
earlier.
I don't think we should be relying on people outside the Control g
Room to prevent these kind of things.
People in the Control Room should be g
able to prevent them themselves.
In our discussions with the staff all g
during the day, it took a long time to get the real problem into prespective.
]
The fact that it was a problem with pressure in the system.
And when the l
staff's been asked about that, some of their testimony that they have not 13l l
addressed very well, why they did not react to the low system pressure.
I think training of the operators is certainly a key area that relates to this event.
Much more so, than insturmentation.
17i l
FASANO:
So then, the training is an important facit that should be improved.
18i You mentioned... you did look at Event Reports, and some of those Event Reports had certain similarities.
One of the things that seems to be one 20!
of the parameters seem to be the pressurizer level going below, below even 21!
site on there, quarter.
In other words, you got your low level indication.
And I believe, this happened at least twice.
But with the rapid cooldown, 23l I could see where you would just get shrinkage.
Then does it appear that 24l also, it seems a common thing that's done is when you have your low pressure,
%3' C
l
s s
s
/l 43 i
l 1:
and you get ECCS, the ECCS is really terminated in the;,e events, as well as 2j always, say, an opening before the actuation of the high pressure injection, L
3!
and opening of your 16 valve, and starting on the pump up, to bring up 4
inventory.
Actually, it's really to bring up the double for the pressurizer.
5; Does this seem unusual to you, to have at least on a number of the trips 6l recorded to have ECCS actuated and the need for an operator to start his second pump on a trip event? I believe, there is only one where you had a g'
rapid cooldcwn, whic.' might have been considered an accident.
The other g;
three or four were just transients, but there was no break.
10; BROUGHTON:
The first question.
Is it unusual to start a seccod high 11 pressure injection pump following a trip?
It is probably not a desirable g{
thing t.o hwe to do.
It would be nice if the design would permit riding through a trip without having to take those actions.
However, yith the
,4 A
system cooling down from 582, where it starts to 550, where it's going to wind up after a trip, there is definitely a shrinkage that takes place, and 1c_:
the operator's response to that, I'd say automatically now, witn the first two steps being to isolate letdown, which is an inventory less, and to start a second pump.
So, if the whole system was designed so that that 191 didn't have to take place, I think that would be better.
The fact that it does have to take place, I don't think is necessarily unsatisfactory.
With 21; regard to initiating ECCS following a trip, I think that that is definitely undesirable.
And as in the case where that has happened, we believe... I 23i think all the ones that we have looked at, we have concluded that it would 24j not have happened, had there not been some other anomaly.
In other words, 25i t[ Q )
))N
,x: I
i 44 I
1l if you have an initiating event, like a loss of feedwater, and you have a 2j reactor scram following that, you would not expect that the pressure would.
3 drop low enough to start an ECCS system.
So I don't think that is something 4;
we expect to see after a trip.
In the case of the cooldown, that was 5l another anomaly that took us well below the 550, that we expected to stabalize that.
Having the ECCS come on was probably quite beneficial in 6i i
7; that case.
Because there was a contraction, and that makes up for the i
gj contraction from getting into the system.
91 10i FASANO:
I have no further questions.
11I RESNER:
This concludes the interview of Mr. T. Gary Broughton.
The time is now 10:07 a.m.
17 t
lo RESNER: We have some additional c::mments that came to light after the 15<
16;!
conclusion of the first tape, therefore, we are going to continue the interview with Mr. T. Gary Broughton.
The time now is 10:13 a.m.
i 17!
18f FASANO:
As we were talking, would you make some comcarisons to Unit I as far as the differences and why, maybe, there's certain things happened on Unit II that were not evident on Unit I?
21, 22l BROUGHTON: Although we haven't made a detailed comparison of how Unit I 231 responds to a transient versus Unit II, there are some differences in the 24!
two plants that would cause a slightly different response curing transient.
25i i
C Q}!
1
[
45 i
t lj For ecample, the physical size of the primary system is nearly the same on 2j both units and yet Unit II is operated at a higher power level... transients i
31 or accidents that cause a loss of secondary heat removal capability this 4
results in a greater power mismatch during the first part of the transient 5l and it can cause the transient to be a more severe transient. In the case 6l of...the power level was one item and the second item is the average tem-7J perature in which the two units operate.
Unit II operates at an average g
temperature of about' 582 and the average temperature on Unit I is I believe 578.
g, So the elevated initial temperature for the transient causes a greater yg contraction in the primary system on TMI II than for a similar transient on g
TMI I since the systems both stabilize out at about the same temperature after a transient.
That has the effect of causing a icwer pressuri:ce y
level and a lower primary system pressure for the same event on Unit II
]
that you'd see on Unit I.
15 FASANO:
So this would explain the abnormal lcw levels on your pressurizer to some degree? Not this event but prior events.
,7 18t BROUGHTON:
It's certainly a factor in why certain TMI events have led to lower pressurizer levels and lower system pressures.
There are other factors, also, for example, the recent pressure on the primary or on the steam system safety valves and so forth.
So, it's not simply a difference related to the primary but the primary factor are definitely strong influences.
241 25}
l h
c y'
f 46 1
FASANO: Would a larger pressurizer compensate or would be a main con-2 sideration to minimize or, at least, make the transients comparable?
I 31 4
BROUGHTON: We believe that the larger pressurizer would certainly make the gl pressurizer level and pressure changes during the transient smaller, and i
Gj it's very likely that a slight increase in the pressurizer size on TMI II q
could give it a similar transient response to TMI I.
We will be doing some l
studies at GPU to determine what the benefits might be from a larger pressurize 8
r.
g I don't know that the goal would be to make TMI II respond in a transient 10l manner like TMI I.
I think the goad really is to see how much the transient response of both units could be improved by changes in the pressurizer that y,
.i...
could be accomplished that would be compatable with the existing designs.
r 13i FASANO:
One of your earlier transients. I believe you were aroung 90%
power level and you were running a temperature coefficient measurement, I think, in your start-up program and you were about six degrees over your average temperature at that time, and then I think...I think you lost one feedwater pump.
191 BROUGHTON:
That's correct.
20!
21:
FASANO:
And you were then on a run back and, I believe, fou got down about 22!
60% or so and got a trip.
There again this would seem to be indicative of 23l what you're saying, the temperature effect was magnified in fact, I think, 24!
in that instance...for the plant to take its normal design course.
Is that 25i correct?
q 0
^
{O L
l l
47 I
1 1
BROUGHTON:
In that particular incident the trip of the reactor resulted 2j from a pressure reduction below the allowed pressure for the distinct hot j
31 leg temperature and the exact cause of the pressure reduction, we have not 4
determined but we did find some malfunctions with the spray valve which may 5
have caused the spray valve to stay open and continue the pressure reduction I
6i below the proper stabilizing point.
The point made about a large pressurizer 7f though is relevant because...to the extent that a larger pressurizer would g
cause fewer operations of the spray valve or minimi::e the pressure oscillation i in a transient like this.
It might be sufficient to prevent tripping on gg 1 ss of one feed pump.
During that transient we also had power operated 10 relief valve actuations.
The power operator relief valve opened and shut g
twice before that scram...and the control of the system, the pressure control of the system was basically a result of what the power operated 13 9
relief valve, the spray valve did, more so than what the pressurizer was doing.
A larger pressurizer might cause you to rely less on things like spray and other techniques.
17 FASANO:
The spray would be just...at your high pressure...it's in automatic...it lSi would then spray.
20t BROUGHTON:
R1ght.
It comes on at an increasing pressure anc then it would 21 spray the system down to a few psi below that pressure.
23l FASANO:
In trying to avert your...electromatic from opening probaoly.
24!
i 25i
- p. -
,Y l
I cO
f I
1 48 l
1 BROUGHTON: That's correct.
2 3
FASANO:
So, again that was not effective, well, it was effective, but not 4!
effective enough.
Alright, I think...is there any other thing you'd like gl to mention now and we can conclude this.
Si 7
BROUGHTON:
I don't think so.
8!
FASANO:
Okay.
Thank you very much.
g, 10f h
RESNER:
This does conclude the interview of Mr. T. Gary Broughton.
The time now is 10:20 a.m.
g i
131 14:
154 15i 17l 18f IS!
20!
21; 22!
1 23j 24l 25i A
C9-a