ML19208B162
| ML19208B162 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 07/06/1979 |
| From: | Keaton R, Marsh B GENERAL PUBLIC UTILITIES CORP., NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III) |
| To: | |
| References | |
| TASK-TF, TASK-TMR NUDOCS 7909190163 | |
| Download: ML19208B162 (31) | |
Text
f.
UNITED STATES OF AMERICA i
l NUCLEAR REGULATORY COMMISSION I
I!
In the Matter of:
l 2
IE TMI INVESTIGATION INTERVIBf
-1 3!
of Mr. Robert W. Keaton l
Manager of Systems Engineering - GPU 44 l
Si Si 7j 8t Trailer #203 91 i
NRC Investigation Site TMI Nuclear Power Plant 10i Middletown, Pennsylvania 11!
June 1, 1979 12!
(0 ate of Interview) 13i July 6, 1979 14!
(Date Transcript Typeo)
~
297 15!
(Tape Numoer(s))
16l 17l 18!
19t 20i 21; NRC PERSONNEL:
I Mr. Robert Marsh 23!
Mr. Anthony N. Fasano 24:
25
( i> E
,7 9 09190 [gg I
I llj MARSH:
The time is 1:27 p.a.
It's June 1,1979 and My name is Bob Marsh.
gj I'm an investigator with the U. S. Nuclear Regulatory Commission assigned 3f to Region III, Chicago, Illinois.
This afternoon we are located at the 4!
corporate headquarters of the GPU Service Corporation at 260 Cherry Hill 5!
Road in Mountain Lakes, New Jersey.
We have with us at this time Robert W.
6i Keaton who is Manager of Systems Engineering with GPU.
At this time I'd 7j like the other individuals present in the room to identify themselves, I
g spell their last name, and identify their position.
9I FASANO:
10 Anthony N. Fasano, Inspection Specialist, NRC.
11:
KEATON:
Robert W. Keaton, Manager of Systems Engineering for GPU Service 13l Corporation.
14!
HOBER:
J. G. Haber, Manager of Generation Division Support.
16i MARSH:
Thank you.
Mr. Keaton before we turned the tape on, I discussed 1,s,;
briefly with you this two page memo which you have in front of you.
As I indicated the memo div.. sses the purpose of NRC's investigation, the authority under which it is being conducted and the scope of the investigation.
It 40 also goes into the rights of the person being interviewed.
And on the second page there's several questions which I have requested your response to.
The first question reads "Do you understand the above?"
231 24!
25;
f
{
2 i
lj KEATON:
Yes, I do.
i 2!
3j MARSH: And secondly "Do we have your permission to tape this interview?"
l 41 5
KEATON:
Yes, you do.
I 6i MARSH: And thirdly "Would you like a copy of the tape and transcript?"
7 l
Si i
KEATON: Yes, please.
g; 101 l
MARSH:
They will be provided.
il 12!
MARSH: There is a fourth question, although it's not called out specifically at the end of the letter, it is included in the body of the letter regarding your right if you so desire to have a company representative present and it's my understanding that this is the capacity in which Mr. Hober is currently in the room?
17!
18(
19!.
KEATON:
That's correct.
20!
MARSH:
Fine.
Thank you.
t 22l MARSH: M r.
Keaton, to begin with could you give us a brief resume of your 2 31 experience in the nuclear field and a description of your duties with GPU 2 41 and then following that if you would we'd like to discuss and get in your c: 0 [2, 3D#
i
I I
3 l
lj own words your recollection of your activities in connection with the March gj 28th incident at Three Mile Island starting with your initial notification, i
3l how you found out about it, and what the sequence of events were for the 4j first several days.
Si Gi KEATON: Well, with respect to my resume, I have a degree in physics from 7!
Yale University which I received in 1957.
The first two years after graduatior i
g I worked for the Dupont Company at the Savannah River Plant on the use of g;
the nuclear facilities there, this was classified work.
I then went in 1959 to Atomics International.
From 1959 to 1965 I was at the sodium 10 g
reactor experiment, initially as a Senior Physicist and later as the Manager y
of the Engineering Section at that reactor.
From 1965 to 1968 I was the y
American representative to the Halden Reactor Project in Halden, Norway, g
where I was engaged in research in fuel irradiation, instrumentation develop-ment, and computer applications to nuclear reactors.
On my return from Norway in 1968, I became Manager of Safety and Analysis at Atomics Interna-tional as part of the fast breeder reactor program there.
I subsequently i
17!
held other positions, management level positions, associated with the 18{
Clinch River Breeder Reactor Project, and finally as Manager of the LMFBR Technology Programs.
Last year I left Atomics International and came to GPU Service Corporation. My present position as Manager of Systems Engineering is to oversee the activities which we perform that are of a broad systems type basis such as the nuclear analysis for the reload of reactor cores, the applications of on line computers to reactors, the safety analysis of 24 present and new reactor facilities, and the preliminary engineering of new 25i plants.
f L
. g I
4 l
1!
MARSH:
Fine, thank you.
Could we take your recollections now of the March 2
28th incident?
l 3:
4!
KEATON: Yes. On March 28th, which was a wednesday morning, I was initially 3l in a meeting at a motel down the street from here reviewing certain admini-6 stration procedures that had been suggested.
John Hober, who is in the 7
room here, was the one who had arranged that meeting.
To the best of my 81 memory it was along about 9:00 or 9:30 that I was summoned to the telephone g
in the motel and my superior, Dick Wilson, asked me to leave that meeting 10!
and come here to the offices because there had been some sort of an incident at TMI Unit 2, with not much more detail given than that.
I immediately y
came over here and we sat down and started to discuss what had occurred.
13 At that point in time we had little conception of what had really gone on at the island. We knew that there had been a transient, we knew the reactor had gone through a turbine trip, and subsequently a reactor trip, we knew that the situation had not been normal, that there had been a high pressure injection, and I don't think we knew a whole lot more than that.
g I remember distinctly being told originally that there had not been any 1
181 release of radiation and so that was the information that we had.
The reason for the meeting early in the morning was to identify a team of
,c0,;
people from here to be sent out to the island to investigate the incident to find out what had happened and to serve as support to the Met Ed people in getting the unit restarted. We identified a group of people, I think 23l l
five if I remember correctly, to go and I subsequently, later in the morning, 24) met with them and with a couple of the other managers here to discuss what 25i
._n
{ b, p
~, k jr /
s
f 5
I lj their duties would be and when they were going out. Our intention was to 2j send them out right after lunch.
It was along about 11:00 I think that i
31 Dick Wilson again came down to that meeting and pulled me aside and indicated 4
that the information that was being received was the incident was more Si serious than we had originally known.
Although even then I don't think we 6
had any real understanding of what we were facing, and emphasized the p
importance of getting that team on the road very quickly, so we stopped our 8,
deliberations and the people very short'ly thereafter left.
The rest of the gj day is a little bit fuzzy frankly.
But we became aware and I think in the 10 morning both Dick Wilson and I became aware of the fact that the main g{
coolant pumps had been turned off and both of us, because of our background gl with sodium reactors I think in part, were very concerned about the fact that the pumps were not running and it was I think in the early afternoon that we met in Mr. Arnold's office down at the end of the building and he got the plant people on the telephone on his squawk box in his office and we started having conversations with them with respect to what was going In some cases I believe we were talking to the people in the control on.
I remember for example I think a couple of times that we were talking room.
to the control room, although I'm not completely sure about that.
And
, 91 1
there were other times that I know that we were talking to Jack Herbein, who is the Vice President of Met Ed, who was located ir the Visitor's 21:
Center across the river from the island.
And he was in turn on the telephone 22!
with the control room and we were getting updates of the plant status, 23 discussing with them what they were trying to do, wnat the plant condition 24l was, and I believe it is a true statement that it was largely as a result 25!
l b'
c ug I
l 6
l l
lj of the recommendation from here that late in the afternoon there was a high 2:
priority placed on trying to reestablish the system pressure to reestablish l
3!
vacuum in the condenser and then te start a coolant pump which was finally 4
done late afternoon or early evening I guess.
I think we were here until
{
about 9:00 that night if I remember correctly.
I think at that time although 5
Gi we knew there... by then we knew that there had been some radiation releases, 1
7j the releases really were small and I think that we left here on wednesday gl evening without even then any real understanding of what we were facing.
g On thursday we prepared to set up analysis in support of the island... started 10 getting activities organized here.
I remember holding meetings with iy several different groups in which I told them what I knew of what had happened on the previous days and we started talking about the kind of support that we might provide them.
At that time to the best of my memory our thoughts were directed toward restart.
Dick Wilson and one or two others I think left late thursday morning to go to the site.
But we were basically I think thinking of the type of team that would need to be set up for recovery and the actions that would be taken, there again indicating 17!
that we really didn't realize even at that point in time how really serious the situation was.
I remember meeting with Bob Arnold on thursday afternoon and having some discussions that he had instigated about the possible organization, the recovery work, and so forth.
I know as a fact that I did 21; not work particularly late thursday evening and friday morning I had gotten 22!
up early and appeared at the Morristown airport at 7:00 to take a flying 23!
lesson, again indicating that at this time we thought that the worst of it t
24l was behind us and ran into Bob Arnold at the airport who said he felt like 25!
6
,y n
i
s
~..
t 1-
- I i,
l 7
!lj the situation was more serious than we had realized and he was on the way i
2:
to the island.
So I came back friday morning and got on the telephone with l
3; Dick Wilson who by then had been out at IMI for what,12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> or so or i
4j maybe,s little more than 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />, and that was the first time that I Sj really got an understanding of the fact that our preb bms were not over.
i 6;
That's the first that I really heard about the hydrogen baoble that was in 7{
the reactor and we realized that the plant wasn't at that point safely shut down.
g Later on on Friday morning I had a long discussion with Gary Broughton g,
who reports to me as Manager of the Control and Safety Analysis Group.
i 101 Gary had baan one of the original team that had gone out on wednesday and had had a chance curing thursday to inspect some of t'ha data and I have nere in fn nt or me the notes that I took in the telephone call with him on fr' day morning and this was the first time that we got a really good rundown of what hac happened.
He had managed to have gone through the data
, 4y enough so that he Could Edsk through the sequence of events and then giVe Itte an update on the CLrrent status of the reaC!dr.
Ano S looking'back over this yesterday, in preparation for the discussions this morning, I s
noticed that the report that I got from nim on fricay morning is sub-stantially correct as we know it today.
And shortly after that,:enversetfon, 201 I gat on the phone and later went and taihid to in person, Mr. DiecKamo',
who is President of GP'J, and went through with him what tNe entire under-21!
standing of the situaMon was.
So it was at that point on f, iday morning J. hat I at least became fully aware of the fact that; (1) the trartsient had been very severe in terms of its damage to the reactor; and (2) th[at there 24\\
~,
was Mill a hydrogen bubble, that the piant was not in a stable coa /iguration, 2S!.
y i
'I
- b
(_
n, NQ'
[
'O i
l l
l
(
8 l
li and that we had a lot of work to do.
We... on Friday then set up here an i
2:
organization to support the work at Three Mile Island.
During the initial i
3j phases of the effort we were the only contact with the architect engineer, 4l Burns and Rowe, so that all requests for information from Burns and Rowe 51 from the site went through us.
This changed later on but initially this 6
was the way it was run.
In addition, we were carrying out a series of i
7l analytical and design tests here in support of a request from the site.
By I
g friday afternoon we had made arrangements 1 have an around the clock i
g operation here and in fact did so sta" ting on Friday. We had made arrange-ments to keep a line...a telephone iine open betwee a.M the site, to 10J g
keep telephone o,narators on du*.y over here around the clock and so forth, f
So that we were basically then sat up in the mode that persisted for a couple of weeks in supporting the site activities.
I was...I suppose more 141 or less the Manager of these at.livities and in fact my office was set up as the command station for the activities here.
I worked very late Friday night and came in again saturday morning.
Then on saturday afternoon I...Mr. Dieckamp called me over and asked me to be prepared to go to the g
site first thing Sunday morning to serve as the coordinator and' liaison for the industry advisory group which he was setting up.
So I made arrangements to have one of the other managers here take over the role that I'd been playing here.
Sunday morning I went to the site and spent most of Sunday meeting with the members of the industry advisory group, bringing them up
,c2 to date on what had happened, what the situation was, explaining to them 23j the type of acti/ities that we were hoping they could help us with, and 24i then for the next few days I was the liaison with the industry advisory 25i group.
cf
?b'
/'
i
l h
I 9
i l
1!
FASANO:
Thank you.
Maybe we could...this is Fasano speaking.
We can go 2l back to the first day, wednesday, the 28th.
Did you hava information with 3f respect to the temperature in the reactor vessel, the hot leg temperatures, 4
or the pressure,,or the incore temperatures?
Si 6
XEATON:
Yes, we did, but not of the incore temperatures. We did not see 7
any data from the incore thermocouples until much later.
We did have data g
on the hot leg and the cold leg readings or at least we had data that they g;
were offscale and in some cases they were offscale high and in some cases ffscale low.
I don't remember with any degree of certainty exactly when 10!
y during the day we first had that data.
I believe that we had some indication i
g of temperatures on Wednesday morning.
I am certain that we had them Wednesday y
afternoon when we were meeting in Bob Arnold's office because I remember us constructing tables on his blackboard.
15' FASANO:
How ' bout pressure?
16i 17 XEATON:
Yes. We also...again I'm not sure when we first had it 'ut a
certainly by Wednesday afternoon at least we had pressure data.
19, 20l l
FASANO:
Was there a correlation made between your TH possibilities and 21 your tem...and your pressure?
e 23l KEATON:
On Wednesday afternoon very definitely because we quickly... as soon as we got the data grabbed the steam table and looked to see if the 25i l
-s -
p';(I
'l I
i f
I
{
10 3}
conditions were saturated in the hot leg and they were and this was one of 2;
the reasons that we expressed back to the site what we felt was the importance 3;
of trying to elevate the system pressure and get a pump started again.
4!
5 FASANO:
Okay then this goes back to probat,1y late in the afterncon, you 6;
made a suggestion?
I 7!
l XEATON:
The first time that I remember for sure that that war done was 8
gf middle to late afternoon, yes.
10!
if FASANO:
At that time do you recall what the status of the reactor was or 12 what the site people were doing? Did you have information on that point?
i 13l p}
KEATON:
Our first contacts with the site...the amount of information we g
got was a little confusing which is I don't think very surprising...that the atmosphere in the Control room at that time was fairly hectic and our unw rstanding sort of grew as a result of telephone calls in the afternoon.
I can't say honestly that we instantaneously had all of this information but rather that it grew as we managed to understand what was going on.
20:
FASANO:
Were you in contact at all with Bacccck and Wilcox, B&W?
21 I
22i KEATON:
I was not personally and I don't remember for sure whether there 23l was anyone that contacted B&W at Lynchburg.
We did have some discussions either directly or second party discussions with Lee Rodgers who is the S&W representative at the site.
)
c j, t
f 11 lj FASANO:
Did you have any information early in the day on the boron analysis 2l that was made in the early morning?
31 KEATON:
To the best of my memory I did not see any boron analysis until I 4!
5 was...until auch later.
I Gi 7l FASANO:
Also the source range or the intermediate range activity as it was g
going up early in the morning?
9f
^ "
0
- * ""I"9' "
- 10]
11:
FASANO:
Specifically, it was reactor coolant pumps that you had infor-f mation on?
141 KEATON:
It was...the first thing that we had information on was the temperature and pressures in the hot leg and cold leg, that's right.
17!
FASANO:
And the reactor coolant pump?
191 XEATON:
And the fact that the pumps were turned off, yes.
21; FASANO:
You say that you were quite concerned on the turning off of the 22' reactor coolant pumps.
'das this early in the morning that you originally 23i had this concern or was it later on?
20 251 A
e o-
i
~
12 lj KEATON:
The first time that we heard that they were off, the comment that i
2:
I made at the time was that if I were out there I would be trying to restart 3
the pumps.
I believe that was in the morting.
I might comment that we 4j later found out...we did not know at that time but we later found out that 5
in fact they were trying to restart the pumps and had been unsuccessful.
6j We didn't know that at that time.
4 71 FASANO: Why would you be so concerned on having the reactor coolant pumps g
restarted?
I mean what was the basis that you had?
I mean you mentioned g;
10 y u were n the sodium loop type?
u:
]
KEATON: Yes.
13!
FASANO:
How did you make that correlation to the pressuri::ed water reactor?
14 15i KEATON:
Simply that the temperature conditions in the loops were such that we didn't feel like... that you could guarantee that you could cool the core by natural circulation and therefore some form of force ccoling would seem to be indicated.
191 20l FASANO:
Okay.
So did you have enough information to evaluate whether natural circulation was in effect or not?
22!
23l KEATON:
No.
Certainly we did not have that.
We had enough information to 241 make us nervous and that's all I can really honestly say.
25{
1,b
<w
l t
13 1
FASANO:
So in that situation you'd rather have it.
2; 3
KEATON:
That's right, that's right.
4i, gj FASANO:
Okay.
You mentioned that you were concerned about vacuum pressure 6
in the condenser or that the establishment of vacuum in the condenser.
7f What was the basis for this concern?
8 XEATON:
This in fact was a reflection of what we had learned from the gg 10l plant that they were trying to get the condenser vacuum reestablished and they felt that was part of what they needed to do in order to reestablish the force cooling in the primary loop and so in my comment I'm really reflecting what the plant told us that they felt was an important thing.
And we agreed with them in tne sense that that would be a good heat sink.
15!
16i FASANO: Would there be any concern here... were you concerned about these 17i were knowledgeable of the steam generator possibility of having shocked the 181 steam generator? Did you have any information? Do you recall?
20f KEATON:
That's a good question.
I honestly do not remember.
I certainly got the information on friday morning and I'm not sure wnether I had it 22!
j before then or not.
23t 24l 25!
b^
c. c,en
l I
1 14 i
1!
FASANO:
Okay.
2:
l KEATON:
I sort of doubt it.
The best of memory is that I was surprised 3
when I heard about it later on.
4l 3
FASANO:
Uh huh.
Now, you mentioned the high pressure safety injection is 6.,
7 not normal on the...so that raised a concern also in your mind.
Did you 81 know that ECCS seems to get actuated...at least high pressure safety injection g
seems to get actuated when you have a feedwater initiated trip at the TMI 2?
10 At least in the earlier year of 1978 I think you had a number of high gj pressure injections.
Are you familiar with this?
i 12l KEATON: Yes i am and in fact I think in part my initial reaction was associated with t 9 fact that I was very familiar with what had happened 1El last year at the time when we had a transient in which the main safety 16l valves had stuck open and resulted in overcooling the primary system, depressurization and a high pressure injection and that particular incident.
17!
Although it turned out to be fairly severe economically, was not severe at all in terms of damage to the core or anything like that.
Our initial reaction I think on Wednesday morning, when we heard about tne TMI transient i
that it had a high pressure injection was that it was probably another one 21l of the same type of thing that the reactor was alright but that we were 22}
23f going
- .o have to do a lot of analysis in order to really understand what had happened and so.that was our original reaction.
The people that we 24j selected to go out to the site on wednesday were based upon that type of 25!
c08 3
{
15 l
1!
thing. Well, okay let's go out and help them analyze the transient.
It 2
wasn't with the thought that we had massive core damage or something like l
that. We didn't suspect that until much later.
3l 41 gj FASANO: The point I think I'm really getting at is then you wer ' knowledgeable 6
that there had been at least two, maybe three, maybe four high pressure 7{
injections in 1978?
8 ;
gj KEATON: Yes, we aware of that and at least in the case of the one which 10 had resulted from the safety valves being stuck open. We had done very 1[
detailed analysis of that transient.
u!
13j FASANO:
Yeah, or.e would consider that abnormal?
14lg KEATON:
Yes, sir.
And as you know I think the plant was shut down for several months and we had to change all the safety valves as a result of that.
18t FASANO:
Yeah, che more normal, trip where you have a run back and your run back doesn't hola apparently...you did get some high pressure injection.
20;.
So in a way this plant seems to be designed to have an easy high pressure injection on a related trip ta the turbine reactor.
Would you say this is true?
231 24j 25i i
P t
f l
f 16 1!
KEATON:
I don't believe so although I will tell you honestly that I l
2l haven't studied it real carefully, but I believe that it's true that 3i normally we would expect to ride out a turbine trip in fact I'm a little l
41 more sure of my ground then I'm sounding.
Our analysis and B&W's analysis 5;
would indicate that we should be able to ride out simply a turbine trip
~
6; without even tripping the reactor.
If the initiating event is a loss of 7
feedwater flow, then our analysis would indicate that we would expect the 8!
reactor to trip but we would not expect to get high pressure infection.
Sl 10l FASANO:
Uh huh.
l 11i 12 KEATON:
And in fact we have very recently, in the of course of the last 13 two or three weeks done :r!me additional analysis which confirms that statement.
14 That we should be able to ride out a simple loss of feedwater flow without 15 getting a high pressure injection.
1Gi FASANO:
Even with what you said..
2205 or 2255?
17, ISI KEATON: Wait a minute.
I'm talking about high pressure injection.
7g 201 i
FASANO:
Yeah, but I'm thinking of you usually use your electromotive valve to an...
23l KEATON:
That's right.
24f 25l a,
J c ug
I i
l 17 1
FASANO:
...in conjunction with this runback?
l 2l i
3j KEATON:
That's correct.
And I'm talking about the case.here that valve 4!
functions properly, that is that is opens but then it also closes when it's Sj supposed to and in that case the pressure comes down and levels out in the l
range of about 2000 psi well above the HPI injection setpoint of 1600 psi.
6 7!
g,I FASANO:
Oh huh.
Have you in your service... as a service function here, gj have you analyzed the transients? Has your group analyzed the transients f r the 1978 events?
101 11:
KEATON:
For the 1978 evcat?
l 131 1*, j FASANO:
Yes.
15' KEATON:
The only one that we analyzed in real detail was the case where the safety valves stuck open.
181 FASANO:
The other ones I guess happened...one was in March.
I guess the one in April is the one you are talking about.
21:
KEATON:
The one that I'm talking about is in April.
That's correct.
22l l
23l l
FASANO:
And again there was one in November and then another in December.
24l 25i
- t cu,8, 3tu t
h
t 18 lj KEATON:
That could be correct.
I understand that the relationship between
- i 2;
our analytical people and the operating utilities has been that we would i
3{
get in and analyze those events only if we wore requested to by the operating 4j utility.
In the case of the April event, where the safety valves stuck gj open, Met Ed requested us to come in and help them.
In the other cases 6
they did not.
7f g,
F_ASANO:
I see.
Do you actually...well in doing this do you...in doing g
your analysis do you get information from other like plants such as the Davis Besse?
10!
11!
g KEATON:
Yes, we do where it's available. We'd look at Davis Besse, 13i Crystal river SMUD or Arkansas.
i 14!
FASANO:
Had you seen the...I guess it was the 77 transient that Davis Besse had?
17l XEATON:
Only after the fact not before the fact-19I FASANO:
Okay, so as far as the following days, the two following cays, do you get involved at all with the environmental concerns or health physics 21!
concerns or...
23!
KEATON:
No. We were of course aware of some of the radiation readings and I in particular, occasionally was a transmitter of that information to top management but only as a messenger basically.
~0 l cy B
[
19 l
!lj
'FASANO:
In your analysis of the transient then in April, I guess you did 2j look at the level of the pressurizer?
3!
4; KEATON:
Yes.
Si 6i FASANO:
...and I guess it did go below...did at that time go below zero?
7!
KEATON:
It went below the indication range, yes.
Our analysis indicated 8) gj that we in fact never did completely void the pressuri:er even though it 10j was below the indicated range.
The analysis of the April event also indicated i
yp in that case we did draw a steam bubble in the reactor deme.
i 12'
]
FASANO:
In the candy canes or in the reactor?
14!
KEATON:
No, in the reactor dome.
That transient was scmewhat different from this year's transient in that it was a transient resulting frem overcooling of the primary system and so the fluid in the steam generators
,7[
in the candy cane was substantially below the temperature that was left in 181 the top region of the reactor.
So in that case the flashing would occur in the reactor first.
It's not like the transient of this year where the hot t
legs were overheated.
21i 22 FASANO:
In your analysis that you said that you did... I guess with 38W where you should remain up around 2000 if the electrematic valve actually 24}
operates properly should be able to take at least single feecwater trip and 2Si l
c 0 9, 32v i~
p
I i
l 20 i
f be able to run back... a double feedwater trip apparently you are not able i
2j to take and still remain on... remain on...
3t i
4; XEATON:
Oh, excuse me.
I may have misread you.
I think it's possible Sj that even with a single feedwater trip that we tripped the reactor.
^
l 6i 7)
FASANO:
Oh really?
I 8:
gj KEATON:
But I would...my point was that we would not expect to get high 10 pressure injection.
11!
FASANO:
I see...so okay.
So then it's...you would get your reactor trip.
As far as...it seems that the operators usually turr. On a seCond makeup 14l pump, almest automatically close off letdown in anticipation of a...one the increase in pressure and the decrease and in that anticipation to minimize that effect.
17j i
KEATON:
And also I think in anticipation of a drop in the pressurizer 18f level.
191 20l FASANO:
Yes.
21t 22l' MARSH:
Break in for just a second for a second to turn this tape 23{
over, *.he time being 1:58 reading 475 on the meter.
24}
25i'
.n\\
5L n
i c SQ',
i l
l 21 i
1]
MARSH: The time is still 1:58 p.m., June 1, excuse me, go ahead.
2l i
3li FASANO:
In your analysis did you take this into account?
I mean is 4l this...would this be included in your procedure to have this kind of 5
sequence of starting the second pump, opening your 16 valves so that you i
Gj get flow greater flow and subsidize the mass within the reactor vessel?
7Y I
Si KEATON: Yes, it is our intention to include that, in fact we try to i
g include as accurately as we can exactly what we anticipate having happen, 10!
yes.
11:
FASANO: Okay. And at least the transients that I've been privileged to look at it appeared that you do still get a dropping.
Even with these actions you do get a drop in pressure and it looks like you do hit the high Fessu e e
np n.
N w, n y ur ana ysis you wouM stay above it if 15 you stayed at 2000, in actuality though it dcesn't appear that that's the r
case.
At least what I've seen maybe because of...maybe special circumstances?
17j 181 XEATON:
It might be, I'm not sure which analysis you are referring to.
The analysis that I'n. describing that we have done has not been a safety type analysis where you look at everything on the most conservative side, 21{
22!
but it's rather been a best guess analysis of what we wotiid most likely anticipate to happen and in that case we would not anticipate going to the 23 l
high pressure injection even if we have loss of all feedwater.
I'm not 24l talking of course about a feedwater line break but simply a loss of *.he 25i
.feedwater pumps.
I n -
N/ b J'
Q (U($
I i
22 f
FASANO:
I guess there was one when your doing temperature coefficient I
2]
measurements during the pre-op, the ascention to power testing, where I-3l guess you were I don't know... what 90 percent power...
4l j
KEATON:
Uh huh.
S 6!
7j FASANO:
...and p s were at six degrees higher than normal on your...
81 KEATON:
Uh, hun, ytt.
g, 10l FASANO:
g
...Tavg and there you got I guess a feedwatar initiated trip?
121 KEATON:
No.
If it's the one that I'm thinking about it was a temperature g
pressure ratio trip.
15i FASANO:
Okay, that's correct.
17:
KEATON:
And that is...your statement is correct with respect to a trip from that. We also analy::ed that one, I'm sorry I didn't remember that one w"cn you asked me earlier. We did some :nalysis on that and our analysis indicated that in the case of a reactor trip with that urbalance between the pressure and the temperature, an unbalance that is sufficient to trip 22l you on a pressure temperature mismatch, that in those cases we would expect 231 to get high pressure injection.
That's correct.
But the case that I 24!
referred to a few minutes ago was where we were...are operating with a 25i normal temperature in pressure and lose feedwater.
D2
&Y
/\\
23 i
i lj FASANO: And then in ths incident on March 28th would you then consider the l
lack of feedwater, emergency feedwater, as being the trigger here or the 2
3 difference?
41 5t KEATON:
No, sir.
Our analysis indicates that as far as the plant behavior 6l is concerned the fact that we did not have emergency feedwater made no real y
difference in the transient that occurred.
It...of course the behavior of g
the reactor during the first 8 minu es was a little different from what it g,
would have been with mergency feedwater.
But the fact that the emergency 10 feedwater was turned on after 8 minutes...and so the cooling was reestablished
]
in the primary system.
After something like about 20 minutes we believed 12:
that the reactor conditions were the same as they would have been had the feedwater been initiated when it should have been.
131 14' FASANO:
So then the drcpping pressure, that caused the high pressure injection, then you would evaluate that as being caused by the stuck open valve?
18I KEATON:
Very definitely.
Very definitely.
In fact, the initial impact of 191 the lack of emergency feedwater was to stop the pressure decrease at a little higher pressure than it would have otherwise.
21:
22f FASANO:
Also the height of the pressorizer? That only went down to 158 231 l
inches. Would you think that might have been contributary to keeping that 2 44 up as well?
2c; C N f,
)n[._ M 4
I
l I
I I
24 f
KEATON: That's a very good question.
I think the answer is that it might I
2j have been and I want to look at some analysis that we've recently done to 3
lo k at that.
It's possible that the pressurizer level would have dropped 4
more had the emergency feedwater been initiated when it should have been.
Sj FASANO: How familiar are you with the emergency feedwater system?
In 6
l particular that is when you have your block valves open and it comes on as 7
g you get a trip, I guess with loss of feedwater your emergency feedwater g
pumps come on...
101 KEATON: Yes.
12ly FASANO: They come up to pressure and then at 30 inch level, approximately, on the steam generators, you have your integrated control system take over and it's in automatic then these valves would open.
Are you familiar with any set leak rate for these particular valves on...so that there is some water coming in immediately?
17i 1St KEATON:
I don't remember the number.
I am aware of the fact that there is 19l such a bypass around the valves but I don't remember the values.
20!
21f FASANO:
Okay.
What would be the intent of this? Do you have any idea?
22l 23j KEATON:
I'm sorry I really can't...
2 46 25l
- *)
i 4
- I i
i 25 1!
FASANO:
But you do remember or you do think there is in the back of your gj mind that there is an automatic starting of water ficw even before you get f
down to the 30 inches?
3 4I g
KEATON:
It is the best of my memory that I have been told that that is 5;
correct.
7!
)
FASANO:
Okay.
I don't think I have anymore questions at this time but g
maybe you would like to, based on the experience that this has given you g;
10j and given us, maybe you would like express some ideas or give some advice as to what you think might benefit all of us as far as operating nuclear power plants.
13!
KEATON:
Well, I think it is clear that we have learned several lessons.
I expect to be testifying before the ACRS next week and answering similar questions.
16i I guess that in retrospect that it's clear that there were several things that contributed to this accident.
It appears, I think crystal clear, to us that the stuck open relief valve vas the mechanical initiator of the accident.
Had that not occurred t!.de is no reason why the plant would not have followed a very normal sequence of events in taking care of itself.
It acpears to me on the basis of what we have been 22l able to see so far and our investigation is by no means complete, that it 231 appears that the possibility of a loss of coolant in the pressurizer, 24l whether from the stuck open relief valve or broken line or other cause, has not really been adequately considered in terms of defining the expected 25i o
c zg 36.0 h
,e -
I
l
~
t
[
i f
26 l
f behavior of the plant and the actions that we would expect the operators to I
2 take in the event of this.
And looking at the plant procedures for TMI 2 i
for example.
The procedures really address loss of coolant accidents which 3l 4j are accompanied both by a loss of pressurizer level and a loss of system 5
pressure, which of course is what you would expect to see if the leak is i
6l anywhere other than in the pressurizer steam space.
So I think perhaps 7{
that the industry as a whole had not carefully enough considered the fact i
gj that this would appear to be an unusual behavior on the part of the operators.
g Frem the standpoint of the operators I think we can say that they did not 10 recognize what was happening to them.
They controlled the plant based upon h
pressurizer level which is what they have done in previous events such as the ones that you described, and which in those cases was the correct thing to do.
In this case it was not the correct thing to do and in some of the training that needs to be done I think we will all want to be emphasizing to the operators the importance of controlling the system pressure if it has dropped down to saturation pressure.
Another thing that has come out of this is that in circumstances such as this, the operator is presented with a maze of information.
All of which is demanding his attention immediately.
And at least in the case of TMI 2 it is probably fairly generally true 19t there's nothing in the way that the information is presented that particularly calls the operator's attention to the most critical information.
He's left to sort out for himself what is the most critical and what is not the most 22!
i critical and some of the things that were of great interest to him are 23l things that are not even displayed immediately on the control panels.
So, 24j it is putting a heavy load on the operator and so in thinking about this 25i C () h
\\
h f
I i
27 I
1!
basically our conclusion in thinking about the lessons are that:
(1) we 2
need to make sure that we take any steps that we can to avoid having this i
3j particular incident happen again, protection against the stuck open relief 4
valve; (2) I think we need...we feel that we want to make sure that we do 5l everything we possibly can to train the operators in the different types of 6l unusual conditions that might result to make sure that they react as quickly 7
as they can; and (3) I think we need to do everything we can to improve the gj way that we present information to the operator.
91 FASANO:
On the...just go back a little...you mentioned if you have a stuck 10 lij pen valve located at the pressuri:er or a broken pipe if you like in the g
upper section of the pressurizer.
In that event...now we go back to the y
reactor coolant pumps, do you think that operating them or not operating g
them may have had an impact on the actual final condition?
15 KEATON:
This is one of the things that we are trying to analyze and we don't really have a final answer to it yet but it appears likely that in fact had the operators not turned off the coolant pumps when they did that the pumps may have very shortly stopped pumping anyway and so the actual 19t action of turning off the coolant pumps, and again I'm not sure, but I believe that our analysis will show that that actual action of turning off the coolant pumps was not the critical item.
23l f
FASANO:
Okay.
The...did you...well just to get back to that you...then 2 46 the starvation of the primary system of mass of makeup water...I gather 25i i(}
g (Q,
bE 1
r l
i
(
28 I
lj then might have been the other critical point in this particular situatio.,
2 and how could the operator be more knowledgeable in realizing, that he 3
shouldn't turn off his high pressure safety injection or increase his 4r letdown...I think in both...he was doing both at the same time...with a 5l leaking... is this correct?
6i 7j KEATON:
Yes.
I think that is exactly correct and I think that what you gl are mentioning is in fact the critical area as far as the operator's gj actions are concerned. We believe basically that there are two ways that 101 the core damage could have been avoided.
One is to...is for the operators f
to have realized that they had a system leak, realize that the pressure had g
dropped to the saturation pressure in the hot leg, and use the high pressure injection flow as a method of controlling system pressure rather than controlling pressurizer level and to have worried about the pressurizer level nly nce they reestablished a pressure substantially above the 15l saturation pressure in the hot legs.
The second thing would have been to 17; have recognized where the leak was and to have stopped the mass loss by closing the blocking valve downstream of the pressurizer relief valve.
With respect to your specific question one of the things which we are 191 procosing to do on TMI Unit 1 prior to the time that we suggest it's restart, 40 is to install a monitor and an alarm which compares the current system 21 pressure to the saturation pressure and which will alarm for the operator 22l if the pressure approaches the saturation pressure so that he understands 23l l
then that his top priority is to work on pressure control and if the pressurizer 24!
goes solid well so be it.
25i L
t
\\
29 l
lj FASANO:
He's got his safety valves there.
2!
i 3l KEATON:
That's right.
4i Si FASAM:
I have no further questions.
i 6i 7
MARSH:
I have one question...just of interest.
Throughout the course of gj these interviews there has been several comments about the lag time with g
the alarm printer sometimes running as high as 45 minutes behind in certain events.
Has this been addressed at all or is that being studied or looked 10 at or what's GPU's view on that?
12:
KEATON: Our analysis indicates that yes the printer did run substantially late.
It's not clear to what extent that influenced the operator's actions but it certainly is a situation that we consider unsatisfactory.
16; MARSH:
I'm not just addressing the incident on the 28th.
It appears to be something that has bothered the operators for some time that it ceased to 181 be a useful tool because of the lag in the backlogging of it.
I'm just wondering if it has been addressed and if anything were in the mill that might prove there is a corrective action for it?
22l KEATON:
Of course with respect to' Unit 2 we are a long ways from thinking 23!
i about corrective actions of that type but...
24!
25!
'h c u p,
l i
30 l
1l MARSH:
No, I'm just talking across the board.
2l l
3j KEATON:
Actors the board. Well, the other place that this arises for us i
4j is in Unit 1 and yes we are presently taking action to change that situation Si in Unit 1.
In fact we are taking more than one action.
One case is just i
Gj that they are putting faster printers in, but we are also taking steps to 7l substantially upgrade the entire computer system and make it much more i
gj responsive and alsc have it do a much better job of being able to communicate with the operator in a fashion that we will find useful.
g, 101 MARSH: Uh huh.
I have no other questions then if you...
y i
12l KEATON:
I'm finished.
, 31 A
14!
MARSH:
Thank you very much.
I would like to say thank you for your time Si Mr. Keaton and your comments the time being now 2:15 we'll terminate the 6i tape reading 770 on the meter.
Thanks again.
- 17l, ISt 19!
20' 21:
22l 231 i
24!
25i c t,9
~
r i
i