ML20072J101
| ML20072J101 | |
| Person / Time | |
|---|---|
| Site: | Three Mile Island  |
| Issue date: | 02/03/1982 |
| From: | Roy D METROPOLITAN EDISON CO. |
| To: | |
| References | |
| TASK-*, TASK-GB NUDOCS 8306290926 | |
| Download: ML20072J101 (181) | |
Text
{{#Wiki_filter:. _ ~ . . . . . . . . . . .. . . -- - . . . . . . 4, 173 i O UNITED STATES DISTRICT COURT SOUTHERN DISTRICT OF NEW YORK
. dot
.1 -----------------------------------------x g GENERAL PUBLIC UTILITIES CORPORATION, s
! JERSEY CENTRAL POWER & LIGHT COMPANY, 1 METROPOLITAN EDISON COMPANY and :
4 t PENNSYLVANIA ELECTRIC COMPANY, Plaintiffs,
- 80 CIV. 1683
-against- (R.O.)
THE BABCOCK & WILCOX COMPANY and J. RAY McDERMOTT & CO., INC., : 1 Defendants. : '
-----------------------------------------x
!O continued deposition of The Babcock & Wilcox Company, by DONALD HENRY ROY, taken by plaintiffs pursuant to adjournment, at the offices of Kaye, Scholer, Fierman, Hays & Handler, Esgs., 425 Park Avenue, New York, New York, on Wednesday, February 3, 1982, at 9:53 o' clock in the forenoon, before Charles Shapiro, a Certified Shorthand Reporter and Notary Public within and for the State of i New York. 8306290926 820203 4 g PDR ADOCK 05000289 - DOYLE REPORTING. INC. T PDR _ CERTIFIED STENOTYPE REPORTERS
.369 LExlNGTON AVENUE
. WALTER SH APIRO. C.S.R. NEW Ycat<. N.Y 10017 CHARLES SHAPIRO, C.S.R. )
l, Tatsonont 212 - 867 6220 i LL l l l _ _ - .. ,---.~.-..--.. - . .~._.._ , , . . . . _ _ . . . . . - . . _ . . . ,- ..-~ ,,__ ,- _
jo . . - . _ . . i ! 1 } 172 i 1 j 2 Appe ara nce s : 3 KAYE, SCHOLER, FIERMAN, HAYS & HANDLER, ESQS. Attorneys for Plaintiffs i 4 , 425 Park Avenue l' New York, New York g 5 l' Er By: RICHARD C. SELTZER, ESQ., 6 of Counsel i I DAVIS POLK & WARDWELL, ESQS. } 8 . Attorneys for Defendants f' One Chase Manhattan Plaza 9 New York, New York 10 By: ROBERT B. FISKE, JR., ESQ.
-and-11 K. ANN MCDONALD, ESQ.,
, of Counsel l 12 i 13 Also Present: I 14 DAVID TAYLOR ' 15 JOHN ARBO l 16 JONATHAN QUINN 17 f 18 19 20 21 22 23 l 24
'25 i ,
I s
,I
1 173 7 ( 2 DONALD HENRY R O Y, having been 3 previously duly sworn, resumed and testified 4
- 4. further as follows:
5 EXAMINATION (continued) . 6 BY MR. SELTZER: 7 Q Dr. Roy, I am sure you are aware your 8 testimony today continues to be under oath. 9 A Yes, sir. 10 Q What are the occasions on which you have 11 previously given sworn testimony in connection with 12 the Three Mile Island accident? 13 A I was deposed by attorneys for the Kemeny
}~
14 Commission in Lynchburg and presented testimony in 15 washington before the President's Committee, Commission, 16 and deposed by attorneys for the Rogovin investigation. 17 I was present at a hearing before senator 18 McCormack. I am not sure that was sworn testimony at
~
19 that hearing. 20 Q You didn't say very much on.that occasion? 21 A I said very little. 22 And here. I 23 ' Q .Have you given ACRS testimony-regarding the 24 accident? l A ! ii
~
l 25 A Yes.- I made presentations to the ACRS.
._.s___-_____--
a 1 Roy 174 l I didn't include that. That wasn't a sworn t pe,of ( 2 3 meeting. [- ' 4 4 Q . How soon after the accident did yoy appear . , 5 before the ACRS? '
~
6 A I don't recall the specific time, but I 7 think it was within a couple of months anyhow, like
/
8 perhaps the June time frame comes to my mind. As I 9 r e c all', it wasn't a long time after the accident. ! ,,;
. ,4 10 Q Did you just testify once before the ACRS 11 regarding the Three Mile Island accident? 1,],
12 A It seems to'me that I was -- made - ( 13 presentations to a subcommittee of the ACRS and then ,, 14 later on, in summary form, to'the full committee. 15 If I haven't got my meetings mixed up. I think that
. /
16 occurred in that time frame, one, two, maybe three , 17 months after the accident. 18 Q What was the thrust of your testimony? 19 MR. FISKE: I don't think he can answer a 20 question like that. 21 Q , What was the principal subject of your 9 22 testimony? , 23 A As I recall, we.were responding to ACRS 24 recommendati,ons, a,f.d I am not sure what the source
' '1 25 of these recommendations were, but I recall one ' -
J.
, rm -
# c_ -_
f
~
1 Roy , 175 ,, T~ - , 2 presentation in which I was giving a B&W position on 3 certain ACRs recommendations that had evidently come -'
'4 somehow out of their own deliberations. '
5 Yes. S Q . w 6 A That is the only one that I -- the subject , l. 7 matter that I-recall right now. 8 Q What was the subject? What recommendation
~
m '\ L 9 were you addressing? ,
- i 10 A There were several recommendations.and I-l 11 nothing. crops to my. mind now. I'think there were 12 recommendations on additional instrumentation such , .
13 as a saturation. meter. There were quite a few." I ,
-1 s
14 would say there must have been six or seven differhnt. 15 recommendations which were addressed at that meeting. 16 Q What do the initials TPEP mean to you? 17 A The-initials that I recall for a program 18 which I called the technical evaluation program, 19 'I seam to recall TEP. I can't place the.TPEP right now.
' ns 20 Q Technical product evaluation project? .,
_t ', ! l - 21- A 'Right.' I recall' referring to it-in my C
;g i
22- mind.as the TEP program, but technical product' -
]
s 2 M% 23- . evaluation program. n s
- 24. Q That was a' program that was initiated in ,~'s\,
O.' J25 response'to.Something suggested by the president o,f I 1
. ~ ,.
V g
.. m.
"U ,
A. o+_
~~
* ' ~
+
2.,s - q ' : L; , .
)
- s,.
, >g .s s;
' \ ^ ~ \; . v 1 , s - 1 Roy 176
- . i A
lrh'sgj 2 B&W; right? Mr. Zipf? T % , 3 A I recall the purpose, but I don't recall
- 4. the impetus for that. That didn't ring a bell with me.
5 g You don't remember Mr. Zipf's comment on
, N 6 ' MacMillan's comprehensive business plan for 19807 7 A' As the incentive for the TPEP program?
's w.
8 Q Yes. . N ,- - 9 N ,. .' sI don't recall that.now. 10 Q? O.K. 11 What was your role in the initiation of the i 12 technical product evaluation project? 13 .A As I recall, first I assigned someone to
., 14 lead thaf, evaluation p'sogram.
~ ,
15 Q Quale?. 16 . i A N'Mr. Gary Quale.
- h. . \.
17 ', 'u Secondly, the charter that I had in mind a.
'18 fSr it was -- this program addressed issues associated s
19 with our standard plant, the ongoing design effort associateds with the NSS that was or would have been
~
20 21 offered to prosp6ctive customers for a nuclear steam e 22 s supply c,ystem, and as I recall, the kind of thing I 23- had in mind-is that, rom the Three Mile Island
- w. ,
s 24 ' accident issued a number of regulations and requirements f}-
'- ' 25 coming from the NRC that would affect the design of
-) -
-, - - L- ,
1 Roy 177 4 [aD 2 that plant. We also had, prior to the accident, an 3 ongoing standard plant design program detailing the 4 design that would be offered for a nuclear steam 5 supplied system, so what I wanted to do was integrate 9 6 all those efforts, pull together the best representation 7 that I could of the kind of design changes that we had 8 ongoing even before the accident, and particularly 9 the impact that various requirements and recommendations 10 would make on how that design should be nodified. 11 That was the broad charter of the evaluation program. 12 Q Did you recognize during the time that you 13 were chartering the program and guiding its early 14 course that there could be favorable spillover onto 15 existing plants? 16 A Yes. Well, that could happen. I don't ' 17 recall harboring that thought, but that is certainly -- 18 possibly could happen, and vice versa, many of the 19 recommendations that were being made and requirements 20 flowing.from bulletins and short-term and long-term 21 lesson plans and ultimately grew into the NRC's test 22 action plan, even though those requirements were 23 oriented to an operating plant, later to a plant under ' 24 construction, it was clear those requirements would (3
\! 25 have to be assessed for potential impact for any
- l. Roy 178
() 2 1 product we would offer in the marketplace. So it could 3 be both ways. We might come up through our design 4 effort on the standard plant with ideas that would be useful in the operating plants and backlog. Most of e 5 really looking at 6 the flow would be the other way 7 those requirements that were being promulgated for 8 operating plants, backlog plants, and then assessing 9 how they may impact the product which would be [ 10 offered in the marketplace as our standard plant. 11 Q At the time of the Three Mile Island 12 accident, had B&W already taken orders for some 205 I~ 13 fuel assembly plants? V) 14 A We had 205 fuel assembly plants under 15 construgtion, already under contract. There were no 16 orders that I recall pending domestically. I believe
~
17 we were in the process of bidding a 205 fuel assembly 18 plant to Korea at the time, but it was not under 19 contract. I believe we were working on a bid somewhere 20 in that time frame, and that is the only response to 21 request for bid that I recall in that time-frame. l
@ 22 Q Are there any 205 plans in the United 1 t
23 States in operation yet? 24 A No.
'~
25 Q LWill there be?
1 Roy 179 2 A Yes. 3 Q Which will be the first one? 4 MR. FISKE: Subject to the vagaries of 5 the NRC licensing process? 6 A As best we can project right now, it would 7 be the Bellefonte 1 unit on the Tennessee Valley 8 Authority networ%. 1 9 Q Was one of the WPPSS units that is being 10 cancelled a B&W unit? 11 A Yes. 12 Q Would that have been a 205 plant? 13 A Yes. 4 14 Q In your answer a moment ago you said that 15 in the T2EP you were seeking to incorporate items , 16 from a postaccident lesson plan.
, 17 Did you mean lessons learned?
18 A Yes, in the sense after the accident the 19 NRC generated or developed a couple of avenues, 20 really three or four avenues, and these differed 21 with time. The bulletin and orders task force that O 22 the NRC established was one source of requirements 23 that particularly affected B&W _ operating units.
- 24 ,Short-term lessons, learned was another
\)
25 intermediate step in the NRC's formulation of +
1 Roy 180 i requirements.
) 2 3' Q Let me interrupt. What I was focusing on 4 was you and your staff developed lessons learned 5 within B&W, didn't you? .
- '6 A Yes, we had an exercise -- well, it was i
7 both the technical review committee that was the 8 source of recommendations and many of-those had g captured some of the requirements and recommendations { 10- that were flowing from the NRC. And we did have an
. 11- exercise of identifying like the top ten lessons i ; 12 learned. I think that was in response to a request 13 from Mr. MacMillan. So all of these would be brought l 14 together, all of these various avenues, and to assess 15 what would be the appropriate design modifications i
16 for the standard plant, pull it all together. I 17 Q But as you use standard plant, you are 18 also meaning developed design modifications for 19 backlog plants and for existing plants that needed
- 20. to be modified in light of the Three Mile Island
- - 21 accident?
- t. -22 A No, the TPEP program was only oriented 23 at the standard plant. Its source was all of these l'
t- ! 24 other activities going on that were affecting operating i O 25 P lants. And other recommendations, sources, whatever, I a j e
1 Roy 181 2 we were trying to pull that together into a coherent 3 l picture, in a matrix that will say these are the kinds
, 4 of modifications that we ought to be considering for g 5 the product that we would offer in response to a 6 request for bid.
7 MR. SELTZER: I would like to mark as 8 GPU Exhibit 29 a memo from Dr. Roy to Mr. Quale, 9
Subject:
Comments on Technical Product 10 Evaluation Project Plan, October 8, 1979.
~11 (Memo from Dr. Roy to G. D. Quale, dated 12 October 8, 1979, was marked GPU Exhibit 449 13 for identification, as of this date.)
( } 14 Q Is GPU Exhibit 449 a copy of a memo which 15 you sent to Mr. Quale in or about early October 19797 16 A Yes, it is. 17 Q Is this a commentary which you had prepared 18 based on materials that Quale had submitted to you on 19 the technical product evaluation project? 20 A Yes, I believe so. I don't have a picture 21 of that material, but yes, this was deriven from a 22 plan that he had' submitted.for my review. 23 Q And you had reviewed it? 24 A Yes. s <. j 25 Q In the third paragraph, do you see where you i I
.-,...n... _ . . - < . , , , , . . ,. , . - . - -
1 1 Roy 382 2 began "The following criteria..."? 3 A Yes. 4 Q You stated, "The following criteria should 5 be taken as ' givens' at the outset and factored into 6 the project engineering work at an early stage." 7 Had you conferred with others before 8 deciding that these ten items would be taken as givens?
-9 A Well, these ten items were really 1
10 derivitive of not only perhaps conferring with others, 11 which I am sure I did and I don't recall any specific 12 instance, but more so it was deriven by the documentation (" 13 in which design features of the plant were either
\_)3 14 called into question or were going to need to be i
4 15 addressed because they had been called into, question 16 or concerns had been raised.about them through other 17 avenues. 18 one I can particularly think of that 19 was on my mind was the so-called report, the Tedesco 20 report. The NRC had after Threr Mile Island reviewed 21 some of the features of the B&W plant and had called O- 22 into question.or raised concerns, O.K., about various 23 , design features. That I recall was very much in my i 24 mind as a driving function that had -- these items that (3 25 I thought really had to be addressed with respectito
1 Roy 183 2 the ongoing product. That is one very specific item 3l that I had in mind as I was saying: Hey, make sure 4 these items are addressed. 5 Q So you had reviewed the criticisms of the g 6 B&W design articulated by the NRC and from that had 7 derived a list of givens for the TPEP project? 8 A That was one source of input that I recall 9 now that really sticks in my mind. Yes, the ICS was an 10 item which the NRC had addressed with us and raised 11 questions about. No. 8, for example, was not driven 12 by that but -- ("% 13 Q Are you saying " driven"?
'(_ -
14 A The incentive for identifying this 15 criterion was not driven by an NRC -- driven, yes, 16 was not impelled by an NRC defect or criticism, O.K.,
- 17. or what they -- what they considered to be a concern.
- 18 No. 8, for example, was a concern regarding better 19 performance of reactor coolant pump seals, which was.
20 an availability issue, so that was another source, 21 issues that had been raised during our evaluation of
.22 availability, means to improve availability.
23 MR. FISKE: What is the pending question?. 24 MR.. SELTZER: He is talking, Bob. 25 MR. FISKE: I know. I was wondering l
1 Roy 184 h 4 j 2 whether he was answering your question or just -- 3 MR. SELTZER: Don't be so formal. 4 A My only point was that the NRC material 5 is not the only source for these thoughts. 6 Q I think that responded to the question. 7 So, in other words, you were looking ahead F 8 at what should be included in a reappraisal of the 9 requirements for B&W's standard plant and you thought 10 that these ten points were the givens? 11 A Yes, sir. I wanted to list them up for 12 special consideration in the project review. 13 Q Item 1 in your words was: "The surge line C 14 configuration should be such as to permit pressurizer
- 15 level indication to be of maximum assistance to the 16 operator in determining reactor coolant system 17 inventory."
18 A Yes. j 19 Q Reactor coolant system inventory means the 20 total mass of water in the reactor coolant system; 21 right? O 22 A Yes, sir. 23 Q And when you say "to permit the pressurizer 8 24 level indication to be of maximum assistance," did
^O
J 25 you want the pressurizer level to function as a t
y -.-.%,- --- . . - . . - . - ~ . . . , . . - , . , , . - -
1 . Roy 185 2 manometer, in effect? 1 1 3 A In effect it would do that, a little more 4 complicated. What I really wanted to do here -- q 5 and this came about because of concerns expressed 6 regarding the surge line configurations that were -- 7 received a lot of attention after the accident -- 8 but what I wanted here was to look at other ) 9 configurations for the surge line which would maximize 10 the ability of the pressurizer to be a reasonably 11 accurate indication of inventory. 12 Q So that the operators in the control room 13 could look at pressurizer water level and if it were 14 in its normal range, they would have accurate assurance 15 that the reactor coolant system outside the pressurizer 16 was full? 17 A Well, obviously this again is to be of 18 maximum assistance. You don't -- wouldn't propose 19 and wouldn't suspect that they are going to identify 20 on one item. 21 Q I am just focusing on what you wrote here. 9 22 MR. FISKE: He is answering the question. 23 Go ahead, Doctor. 24 A This really responds to the questions
.('b '
25 that were raised, and since satisfactorily resolved,
+
- - - - w -,-*m p ,y9 y %,we-s y .c y ei
1 Roy 186 [O\ 2 as to whether the configuration of that surge line 3 at Three Mile Island contributed in any way to the 4 pressurizer level being high. That was since determined 5 to be an irrelevant concern, but at this particular 6 time that was a concern which was not resolved in 7 the NRC's mind, as to the role that the surge line 8 configuration would play in either permitting or not 9 permitting the pressurizer to drain, and I wanted 10 that considered because that was an item that had 11 received considerable notoriety. 12 Q I appreciate all of the extra information 13 on surge line configuration. 14 What I wanted to focus on was-your 15 statement that as a given, you wanted to permit 16 pressurizer level indication to be of maximum 17 assistance to the operator in determining RCS inventory. 18 And I am asking: In-what way did you envision 19 pressurizer level indication could be of assistance f 20 in determining RCS inventory? In other words, how 21 would an operator use pressurizer level indication O 22 to assist.him'in determining reactor coolant system 23 water volume?' '~
' 24' A As an element, one piece of information b' 25 available to him'in' helping him to diagnose what the
., . . _ _ , , , , , , , . . _ .m.,. . . . - . . . , - . . . - . , ,,% n, ,-w , e
1 R6y 187 r"%
.() 2 inventory might be, what I was driving at here was 3 consider surge line configurations, and this has to be 4 in relation to the various transients that the plants ggg 5 may be subjected to, which would permit that to be, 6 to the maximum, a useful means for verifying whatever 7 conclusion he drew with respect to inventory in the 8 system. You can't do'it all the time, but is there a 9 change in the surge line configuration which would 10 help that to be a more useful mechanism for verifying 11 whatever conclusions he draws on inventory. That is 12 exactly what I had in mind. ~s s 13 Q At the time you wrote this, was it your
, [)h x 14 understanding that the operators at Three Mile Island i 15 had relied on pressurizer level indication as an 16 indication of reactor coolant system water inventory? 17 A Yes, at this time I had been quite well 18 aware of that. 19 Q Were you also aware at the time you wrote 20 GPU Exhibit 449 that the Davis-Besse operators had 21 relied on pressurizer water level as an indication 9 22 of reactor coola'nt system water inventory when they 23 manually terminated high pressure injection? l 24 A Well, as I testified after the accident -- 25 -Q Yes.
i
- s _
l . 1 Roy 188 1
) 2 A --
and in discussions with counsel and 3 the whole Davis-Besse issue came back up again, yes, 4 that was refreshed in my mind through the discussion i with counsel.
- g. 5 .
6 Q so'you knew when you wrote this on October 1
- 7. 8th that there had been.at least two instances when 8 operators had terminated high pressure injection 9- relying on pressurizer water level as an indication
- ' 10 of reactor coolant system water inventory; is that i
11 right? 12 A Yes, I was aware of that information at L
-O 13 this time.
- Y)-
14 MR. FISKE: Excuse me,_let me just say. 15 something to Dr. Roy.
- 16. (Counsel conferring with the witness off 1-17 the record.)
18 Q TMI-2 is a 177 fuel assembly design -right? j_ 19 A Yes. 7 20 Q In that design there are conditions under 21 which the surge line leading into the pressurizer
'l h 22 acts as a loop seal; isn't that right?
? 23 A Yes. It could -- because it is looped,
; 24 it could present what is called a loop seals right.
! 25 Q ... And'because it can operate as a loop seal, 4.- 9 E I' .
,- , ,- . . -n..,+.-n - .. + , . , . , - , , . --~,,.,,,e,., ,,- ,-, ~ -,e-n- -,,n , , e-u
1 Roy 189 ( 2 it can end up holding the water in the pressurizer 3 even when the level in the reactor vessel has fallen 4 below the level of water in the pressurizer; right?
, 5 A- Well, you are going to get me into an 6 analysis here that I don't know the details on, but 7 we have done considerable analysis of the role that 8, that configuration might play in other than just a 9 static manometer type situation.
10 MR. FISKE: Just stop there for a second 11 and let me go off the record. 12 (Discussion off the record.) q 13 MR. FISKE: Will you please repeat the 14 question and the beginning of- the answer. 15 (Record read.) 16 A (continuing) And these analyses have 17 confirmed that.that configuration is of no consequence 18 with respect to the response of the level during 19 ~ dynamic situations and I believe the NRC's review 20 concluded the same thing, that we see no compelling 21 reason to change that configuration. 9 22 Q All right. I wasn't asking you for a i 23 , defense of the present design. . I would like it if I 24 you could answer my question, which was whether the , ("T 1- 25 surge line would act as a loop seal and retain water
-,,,m ,, . , . _ , _ , . ~ , _ . . . _ . . . . . _ -
. __7 . . - - - - ,
I Roy 190
)
2 in the pressurizer even when the water level in the , 3 reactor vessel had fallen below the level of the 4 pressurizer? g 5 MR. FISKE: You mean under all circumstances 6 or could it? 7 MR. SELTZER: I think I.said in my question, 8 are there circumstances under which the surge 9 line acts as a loop seal and would hold water 10 in the pressurizer even though the level in the 11 reactor vessel had fallen below the level of 12 the pressurizer. , '( 13 A My answer to that is that under a static , 14 manometer type of situation that could happen, but 15 to the best of my recollection, our analysis shows ) 16 that even if you are voiding, you know, you have ~ 17 saturated the hot leg, t' hat you can drain the 18 pressurizer, you can have two-way flow in the surge 19 line and drain the pressurizer. 20 Q During the Three Mile Island accident 21 isn't it a fact that the configuration of the surge l 22 line prevented the water from flowing out of the 23 pressurizer and into the reactor vessel? 24 MR. FISKE: Is that the end of the question? O- 25 MR. SELTZER: Yes. l
, . _ _ . _ . . _ - . _ ~ , . . . . _ _ _ . . - . - - . . , _ - . - - . . , , ,.. . - .. .. . .
1 ' Roy 191 , 2 ~MR. FISKE: I think, Mr. Seltzer, you are 3 now asking a question which requires expect 4 testimony and analysis and I don't think Dr. 5 Roy has to answer that. question. He is not here 6 to express expert testimony on the accident 7 sequence. 8 MR. SELTZER: If he knows the answer, he 9 is required to testify to it. That is the 10 silliest thing I have heard in several weeks, 11 Here is the man who is the head of the engineering 12 department at B&W, who got an enormous amount 13 of data on the accident, and I am asking him 14 a factual question. If he knows the factual 15 answer, I don't see how it is improper. I am 16 not asking him to do an analysis in front of my 17 eyes. 18 MR. FISKE: In other words, your question 19 is, has he analyzed that question prior to coming 20 here today and reached a conclusion? 21 MR. SELTZER: Or has he seen or heard 9 22 an analysis that gave him an answer. 23 , MR. FISKE: You can answer that. 24 A I know that analysis was performed, but
-Ql
~ 25 - I am not -- I don't recall the answer to that,_except I
i
- - - -- - ,-e v +
1 Roy 192 . I
) 2 from this standpoint. When the PORV is open, when 3
there is an opening at the top of the pressurizer and 4 the system is pressurized, obviously all the water is 5 g ing to try to make its way.out the hole, so the 6 configuration plays no role at all. When that path 7 is secured and there are voids in the primary loop, 8 the response of the pressurizer in that case, I don't 9 know the result of that. I don't know what that 4 10 analysis showed or would show. 11 Q Has'that analysis been done, to your 12 knowledge? () 13 A I believe so. 14 Q By whom? 15 A And I don't know. But I know questions 16 with respect to draining of the pressurizer have been 17 analyzed to some extent at B&W and at the NRC, but I 18 am not familiar with those analyses. 19 Q How would you get a copy of those analyses 20 if you wanted to get them? 21 A How would I get a copy? O 22 Q Yes. 23 A I would probably go back and ask the plant l l 24 design manager if he is aware of that analysis. l-
~
25 Today I would probably ask Byron Nelson,
I Roy 193
) 2 has he run across that analysis, legal counsel at NPGD, 3 since he has the most voluminous-accumulation of 4 information. Mr. Carlton would be a possibility, 5 to have an awareness of that-analysis. That is where 6 I would start.
7 Q Let me show you a memo that you sent to 8 Al Womack on April 5, 1979, shortly after the Three 9 Mile Island accident. It has been previously marked 10 GPU Exhibit 40 and has as its last two pages a
~11 handwritten attachment.
12 Have you taken a look at the handwritten
-13 attachments?
14 A Yes, I have. 15 Q Is GPU Exhibit 40, together with its 16 attachments, a memorandum which you sent to Dr. Womack 17 on or about April 5, 19797 18 A The typed memo I recognize but -- 19 Q Take a look, before you finish, do you 20 see where under *DHR:dmd" it says." attachment"? 21 A Right.
@ 22 I do'n't recall the attachment at the time 23- that I prepared my written material to the --
to E 24 Allen, but I do recall very specifically my thoughts 25 as I was raising these questions after the accident. t
- - - - ~ A e,,- -- - e-
1 Roy 194 T ) 'l Q Does the format in which you and your v. 3 office worked at this time give you reasonable 4 assurance that when you said " attachment," it meant that 5 there was something attached? gg 6 A Yes, and I am not saying it didn't go 7 out from my office with the attachment. I don't l 8 recall the attachment when I was, you know, posing 9 my questions to Dr. Womack. 10 Q Do you know Henry Bailey? 11 A Yes. 12 Q He is in Taylor's licensing outfit, or 13 was at the time?
- f~J}
14 A He was at the time. 15 Q Do you see the statement at the bottom 16 of the first page of your attachment? 17 A Yes. 18 Q "The pressurizer did not drain because of
/
19 the loop seal which exists"? O 20 A Yes. 21 Q Did you ever have any discussions with 9 22 anybody about the fact that the pressurizer did not 23 drain during the Three Mile Island accident because 1
,. 24 of the loop seal which existed? 1 \'# '
l 25 A' I don't recall any specific conversations,
1 Roy 195
-( 2 but that was an issue that the NRC had raised shortly 3l after the accident and, as I say, was discussed in 4 this Tedesco report that they had prepared, so it 5 was an issue to be addressed.at that time, a major 6 issue to be addressed at that time. And I don't 7 recall any specific conversations I had with people 8 on the issue.
9 Q At the end of your attachment, the last 10 sentence says, "If the pressurizer had drained, the 11 high pressure injection pumps would have probably been 12 left on." 13 A I see the statement. 14 Q Did you ever take issue with that statement? 15 , MR. FISKE: Do you mean did he ever say 16 to anybody that he thought that was wrong? 17 MR. SELTZER: Right. 18 A I didn't pursue this handwritten attachment 19 with anyone that I can recall. I directed this 20 question as outlined on the first page to Dr. Womack. 21 Q Did you ever discuss with anybody the l 22 subject matter of the last sentence of your attachment? 23 . MR. FISKE: You mean the attachment as'part 24 of Exhibit 407
.O k_)- 25 A I have no recollection of any specific
1 Roy 196
~s
) 2 conversations I have had on that subject, and again, 3 I didn't address this attachment, to my recollection, 4 with anyone.
5 Q What did you understand was Henry Bailey's 6 area of expertise? 7 A Mr. Bailey had -- well, had previously 8 worked, I believe, in the ECCS analysis so he was -- 9 and frequently handled licensing issues associated 10 with ECCS analyses. That would be an area of his 11 special knowledge. 12 Q Have you ever seen anything that refutes ['/
\_
T 13 Henry Bailey's statement that the pressurizer during 14 the-Three Mile Island accident did not drain because 4. 15 of the loop seal which exists? 16 A That is what I was referring to, the 17 detailed analyses, becau'se this entire issue of 18 pressurizer response during TMI-2, and I think in
;19 considerations of other kinds of transients too, was 20 part of a. detailed analysis, and I don't recall the 1
21 . analysis, wasn't involved in it mycelf, but the two
@ 22 things that stick out in my mind were that configuration 23 was irrelevant with respect to the response of the 24 pressurizer at TMI-2,.and the analysis did not also 25 form a basis.that would compel change in that
1 Roy 197
/ f 2 configuration in our standard plant, I recall those
, 3 bottom line results, but I am not familiar with the 4 detailed analysis. And I believe that is corroborated 5 by independent NRC analysis of that configuration. g 6 Q It doesn't sound irrelevant to me if 7 the pressurizer would have drained and the high 8 pressure injection pumps would have been left on. 9 MR. FISKE: I think the conclusion was 10 that the' statement is not accurate. 11 MR. SELTZER: I haven't. heard any testimony 12 to that effect. 13 MR. FISKE: That would -- 14 Q Do you know who wrote the report that you 15 are referring to? 16 A No, I don't. 17 Q Do you know what the title of the report 18 was? 19 Q I don't even know if there was a report. 20 I know there was an analycis and so no, I don't. 21 Q Do you know who did the analysis? O 22 A' No. 23 Q Do you know when the analysis was done? l 24 A I can't pinpoint it very accurately. I
'- 25 would say within several months of the TMI-2 accident.
l
1 Roy 198 . . gg t j 2 O Do you know how the analysis was done?
%/
3 A You mean the detailed code? 4 Q Anything.
- 5 A No.
9 Do you know what unit performed the analysis? 6 Q 7 A To the best of my recollection, that would 8 be the control systems or what is now called the power 9 systems.and controls unit. ,. 10 Q Did you ever attend any presentstion of 11 the results of this purported analysis? 12 A Not that I recall. 13 MR. FISKE: I object to the form of that 14 question. 15 Q How do you even know that such an a nalysis 4 16 exists if you never attended a presentation and don't 17 even know whether a report was done? 4 2 18 A My recollection is that an analysis was 19 done on this issue and the results that I have just 20 testified to are in my head, those bottom line 21 ' conclusions. I can't walk you from how they got there. O 22 Q Who did you hear about it from? , l 23 ; A That I don't recall. I s ,, 24 Q Returning to GPU Exhibit 449, which is your t V () L 25 memo on TPEP, item.2 states, "The pressurizer and l
s 1 Roy 199
- () 2 pressure control system should be such as to assure 3 that level indication is not lost for anticipatcd
. 4 transients with no credit taken for letdown isolation 5 or makeup flow." 6 MR. FISK: I am sorry, which paragraph is 7 that? 8 MR. SELTZER: 2. Paragraph numbered 2, 449. 9 MR. FISKE: Oh, I am sorry. 10 Q Is item 2 concerned with a problem of 11 losing pressurizer level indication low? 12 A Yes. (k 13 Q It is a fact that the NRC was displeased 14 with the prospect of the pressurizer draining to the 15- point where loss of level indicatfon occurred; isn't 16 that right? 17 Q That was a concern called out in the 18 Te desco report. I am not sure about displeased or 19 what, but that was an issue that had come up in the 20 Tedesco report. - 21 Q By displeased, I just meant that they 22 indicated a concern. 23 A A concern, right. l 24 Q You were aware that the NRC had indicated
'- 25 a concern about the loss of pressurizer level i
n .r 4 ., ,- , - , - . - , n ,. -, - -.._.,w -- r, , , , . . -
- - - , a -
r , -.
1 Roy 200 I 1 i l 1 , 2 indication following a November 1977 event at
- 3 Davis-Besse?
4 A Yes. 5 g And you knew about that before the Three 9 6 Mile Island accident? 7 A Yes,'right. 8 Q Had anything been done to correct.that 9 problem or respond to that concern before'the Three 10 Mile Island accident? 11 A That is a concern that is primarily 12 associated with overcooling events, and also for the
- 13 ' propensity of the operator to initiate high pressure s-14 injection to keep that level, level up.
15 so the corrective actions would beJdirected 16 at, for example, at-the SMUD which was an overcooling
. 17. transient. One action I think-of right offhand is 18 the review to correct that circuitry so you wouldn't 19 have, you know, so many instances of overcooling.
20 That is one. 21 I know of no generic actions taken by 4 22 our customers to address that either from the
. 23 standpoint of mitigating-the number of overcooling
,. 24 transients or other actions that wuold help keep that
-(-
\> 25 pressurizer level ~on scale,fother than that specific d'
t lD r.
}.
1 Roy 201 ^ review that took place at SMUD. ( 2 3 Q The other domestic PWR manufacturers have 4 larger volume pressurizers, don't they? f 5 A I have seen, recall comparisons of 6 pressurizer volume relative to RCS volume which is 7 the key parameter, and as I recall the perception 8 as it existed before, that it was looked at a little 9 more closely, that the B&W unit had some sort of'a 10 uniquely small pressurizer was disspelled and that is 11 not the case. As a matter of fact, in terms of -- 12 again I am recalling here a table which was looked at 13 and that pressurizer volume-relative to RCS volume, 14 that B&W was not all-that different and in some cases 15 better than other PWR units. 16 Q ' Bigger than some and smaller than some? 17 A That ratio which would be the important 18 determinant of volume change as a function of pressure. 19 It is not unique in that respect. That was an earlier 20 perception which was disspelled. '. 21 Q. Did you ever -- 22 MR. FISKE: Off the record. 23 (Discussion off the record.) 24 Q Dr. Roy, are you familiar with'any (~h
-\-) 25 comparison of pressurizer volume to hot leg volume?
l i s W -.
3 J J- 1 Roy 202 7 2 A That may have been one of the parameters l 1 3 compared, but I don't recall whether that was in that l l 4- table or not, that-I see an image in my mind, so I 5 have to say I don't know. 9 -
- 6. Q So you don't know whether B&W came out 7 uniquely small on that ratio as constrasted with other-
- 8 FWR manufacturers, do you?
9 A I don't recall that parameter, no, sir. 10 Q Did you compare the volume of water in 11 ycur once-through steam generator with the volume of f 12 water in competing PWR manufacturers' steam generators? 13
. A That issue has come up and yes, we-looked 14 at some comparisons of those. Those comparisons are 15 like comparing apples and oranges. I mean you've 16 got two different --
two different design bases for 17 the two units.
' 18 MR. FISKE: May I just get this' straight.
19 F I assume-that all these questions that you are 1 20 asking him about, comparisons that they made with 21 the equipment or design of other manut'acturers, h^ 22 refers to a time period after the_ accident. 23- MR. SELTZER: I wasn't limiting it to'that, 1 l
.O 24 and if Dr. Roy wants to indicate whenLwe get to
\ l~ 25- a-point,.where he knows that a comparison was W
a t -
'N 4
g - - , - .,w, v.w,., ,-m,. -r,-y- -- v,-- ,>...----,-.3-- y3 ,y y--w-,- +,y,,,-,tww,y,,y
-,,w,, ~,vec,,e , - -
..,,,,.c
*re-m,, ,-
1 Roy 203 2 made, whether it was done before or after the n_ s i accident -- 3l 4 MR. FISKE: I think it would be -- 5 assuming that this whole line of questioning is 9 6 ever going to see the light of day in a 7 courtroom, in light of the positions taken by 8 GPU and Het Ed in the TMI-1 hearings, I think it 9 would be useful to have some differentiation by 10 Dr. Roy when he is asked about these analyses 11 as to whether he is referring to analyses done 12 before the Three Mile Island 2 accident or
~
7 ) 13 after, and maybe we could --
'LJ 14 MR. SELTZER: All right, let's go back, 15 because I think that is what you are going to 16 suggest.
17 g You said that you recall there was a 18 comparison of pressurizer volume to RCS inventory 19 performed and that it ranked B&W's plant and competing 20 vendors' plants. Was that done before or after the 21 Three Mile Island accident? 22 A After. 23 Q Do you recall a comparison of steam i I 24 generator volume? s 25 A Yes, a very cursory examination. I remember 1 h
\
\
s _.
- l. A 1 Roy 204 p- . < N.
's<,
(~ 2 it was prepared for Mr. MacMillan's use. I am not
.( j
.. 13. l ="re what he used it for. That was well after the
'! accident, I am' talking about maybe a year, two years.
, 4_
s
, ;T 5 Q Is it correct that your steam generator u 6 water volume is less than any competing PWR vendor's
'- '1-7 water valume?
( , 8 A Part.of that analysis determines -- is based 9 on what inventory you are looking at. For example, if 10 you look at just that which is contained within the . 11 boundary called the steam generator, yes, it is l [ 12 smaller because it is an entirely different kind of 13 design. 14 If you look at the inventory, which we did, 15 which comprises that in the generator at a given time
' 16 plus that available through coastdown of the main s
17 feedwater available from the condensate storage 18 tanks, it is very nearly equal. 19 Q In terms of the sensitivity of the steam 20 generator to temperature change, isn't it the volume b
, 21 of water within the generator that is most relevant?
22 A Do you mean temperature change? 23 Q Yes. e e 24 A Temperature of what? C.~ ' , 25 Q Temperature of the water. The ability of
- gs 3 .
v
.h
$b Q .. ' '. -
1 Roy 205 () 2 the generator to be a heat sink. 3 A That one I haven't thought through. 4 Let me cast it in a little different -- 5 MR. FISKE: Just a second, Dr. Roy. If g 6 you haven't thought that through before, I 7 don't think you have to today do an analysis 8 for Mr. Seltzer, particularly since, as I g indicated be fore ,my understanding is that this 10 whole issue of the B&W steam generator as opposed 11 to the steam generators of other utilities was 12 specifically raised as an issue in the TMI-1 13 restart hearings and GPU and Met Ed have taken l 14 a position that there is no safety problem 15 presented by the B&W design. 16 MR. SELTZER: Well, I think once you know 17 the deficiencies in the design, you can adjust 18 for them. i 19 MR. FISKE: I think the position that your 20 client has taken, if I understand it correctly, ; 21 is even after the Three Mile Island accident, O 22 .looking back on that, they are stating under 23 oath to a regulatory commission that they don't 24 believe there is any safety problem presented
.A)
I
\' 25 by.the present design. If that is their position
1 Roy 206 2 there, I don't see how it can be much of an 3 issue in this case. 4 MR. SELTZER: I think it is like comparing g 5 a stool to a chair; as'long as you know there is 6 no back on it, you don't lean back and the stool 7 isn't inherently unsafe. 0 MR. FISKE: Obviously I am not cutting off 9 this line of questioning. 10 Q Item 3 in GPU Exhibit 449 says, in your 11 words, "High pressure injection flow should be such 12 that the full range of small breaks can be accommodated Q 13 regard 3c:2s of reactor coolant pump status at any time 14 after the break." 15 What was the genesis of that given? 16 A After the accident when we were preparing 17 guidelines for management of small. breaks, the so-called 18 blue books, somewhere in that time frame I became 19 aware of the issue of system response when trips were -- 20 when pumps were tripped early versus.when they were 21 left running continuously or may have tripped at some 9 22 intermediate time relative to ESFAS initiation, and 23 I wanted that to be a consideration in our TPEP
- q. 24 program from this standpoint, that it wouldn't make b' 25 any difference, you know, when the pumps tripped.
1 1 Roy 207 2 It would be sufficient ECCS flow,it wouldn't make 3 any difference when the pumps tripped as far as the 4 peak allowable linear heat rate in the core, so that 5 issue had come up somewhere in that time span and I ggg 6 wanted that to be addressed. 7 one of the things I wanted to avoid in 8 this was the requirement to pump trip -- to trip pumps 9 early, which had been imposed, I believe, by this 10 tims, or was about to be imposed, a requirement that 11 said when ESFAS trips there may be another requirement. 12 But anyhow the reactor coolant pumps tripped and the 13 NRC required that operators trip that pump, and I\_)\ 14 therefore the plant was in natural circulation. I 15 wanted to eliminate that requirement. 16 Q You felt that it was desirable from a 17 core cooling standpoint to keep the-primary coolant 18 pumps running? 19 A Not so much from a core cooling, but in 20 . terms of going from forced convection to natural 21 circulation when you didn't have to. O 22 Q Natural. circulation is.one way of. cooling 23 _ the' core? l 24 A Yes. b)' L' 25' Q Forced. convection is another way?
1 Roy 208 4W 2 A Right. 3! Q other things being equal, you prefer to 4 rely on forced convection; is that right? 5 A Yes, you can keep your pumps running, 1 ggg 6 keep them running and looking ahead at our standard 7 plant, I thought we could achieve that if we take a g look at our ECCS full requirement. 9 Q When did you become aware that ac long as 10 the reactor coolant pumps stayed on, the core could it be effectively cooled even with very high void 12 fractions? MR. FISKE: I object to the form of the
) 13 14 question.
15 A well, that issue came up after TMI-2, at 16 least in my mind, with the questions regarding the 17 tripping of the pumps thit occurred a couple of hours 18 or so, or a little less than a couple of hours, after 19 the transient. That is the first time it was raised 20 in my mind, came to my attention, that I recall. 21 Q Did any engineers in your department give 9 22 Y,u any briefing on the ability of the core to remain 23 -._ effectively cooled with high void fractions so long 1 24 as the main coolant pumps' stay on?
-(3). A No, none'that I recall.
25
1 Roy 209 2' Q How did it come to your attention that 3 the core could remain effectively cooled with high
- 4 void fractions so long as the reactor coolant pumps i
5 stayed o.n? - 6 A I am not sure I even arrived at that result. 7 The question came up whether it could, and I don't know 8 how it has been resolved. I haveen't had any discussions 9 myself or reviews of that in any conversations that 10 I can recall. 1 11 -Q so you have never seen any analysis of that? i ! 12 A No. The question came up, but I don't 4
- ("N 13 know what's been done at B&W or anywhere else, you 14 know, to say it can or can't, or what the response l
15 of the pumps would be. No definitive analysis that 16 I can recall.
. 17 Q In what context did it come up?
i 18 A- Just could it or what -- you know, to what. 4 4-19 degree of cooling would you get, what impact would 20 there be on the pumps; this kind of thing. I don't
. 21' know what -- I don'.t recall a specific' conversation.
22 They were the questions that I know were raised, but n . I I don't know of any process or any work to address 24 that definitively and come-to some conclusion, as I O
~
.25 recall.
i
+ - ,en--e vn- ,,.r,-*
1 Roy 210 2 Q Do you know that forced convection of 3 steam can provide core cooling? 4 A It can -- if I can move steam through the 5 core, you can get cooling, but I don't know how much. 6 I haven't seen an analysis and haven't done one, so 7 in terms of, you know, you get enough or what. 8 Q Would you have to be an experienced 9 . thermohydraulic dynamicist to evaluate that? 10 A Probably so, yes. You are going to get 11 - some cooling, o.K.; I just can't produce for you some 12 quantitative assessment that says it is going to be (} 13 enough or it is not going to be enough on it. But 14 obviously forced convection steam cooling, depending 15 on the properties of the steam, can remove substantial 16 heat. 17 MR. FISKE: Is this a good time for an 18 11 o' clock break? 19 MR. SELTZER: Yes, sure. 20 (Recess taken.) 21 MR. SELTZER: What was the last question,
@- 22 please?
23 (Record read.) i 7, 24 BY MR. SELTZER: . (
'~'
25 Q Let me show you GPU Exhibit'262, which was I 5 e - , - - ,- ,.s -
-,-.--y~-
l 1 l 1 Roy 211 s, 2l a preliminary report on safety concern PSC 16-79. 3 Do you see that you are marked for a copy 4 I on the first page? g 5 A Yes. . 6 Q This PSC deals with the pumps running - 7 pumps off question. 8 I just want to ask you, do you recall from 9 the period when that dispute was ongoing that one of 10 the reasons that was advanced for wanting to keep 11 pumps running was that there was core cooling with 12 voiding as long as the pumps stayed on? () 13 A When you said "this issue," this issue or 14 the one we were discussing in item 3 on GPU Exhibit 449? 15 I an a little mixed up right now. 16 Q What I am' talking about right now is the 17 issue that arose over whether the reactor coolant 18 pumps should remain on or should be tripped in the 19 -early stages of a transients that is the issue. 20 Do you recall that there was discussion 21 of that issue immediately following the Three Mile 22 Island accident? 23 A Not to my recollection. The first time
,3_ '24 that I can recall this coming up in my mind and -V aware of these analyses I associated with the 25
1 Roy 212
.( ) 2 l
preparation of those blue volumes where we were doing 3l additional small break analyses. 4 Q That is pretty soon after the accident, isn't it? gg 5 . 6 A Pretty soon. 7 Q That is the period I was referring to. 8 A O.K. 9 Q And you recall that during that period 10 pretty soon after the accident there was a discussion 11 about whether the reactor coolant pumps should. remain 12 on during a transient or should be shut off early 13 in a transient? 14 A I don't recall any specific conversation 15 that I was involved in. This issue in my mind, even 16 today, dealt with the identification of worst case 17 ECCS accident in terms of complying with 10 CFR 50.46. 18 Previous to this issue being addressed 19 after Three Mile Island, the assumption we -- we 20 looked at two cases: we assumed, one, the pumps 21 tripped early, or they continued to run. This issue 9 22 then came up that there may be a worse case in between. 23 Q I understand that. I am not really asking l i 24 for the whole scenario. l
- O 25 The' issue had come up before the Three h -m- p q --- w J-,g ,-- ,rs-w,- --,,w.m % v 7me -- , y e- --n-+
1 Roy 213
)- 2 Mile Island accident, right, in the context of 3 l shutting them off or leaving them on throughout the i
4 whole transient? ggg 5 A That I didn't have any recollection of. 6 Q Take a look at Attachment 1 which refers to 7 plant design's progress report for January 1979. ~. 8 Is it your understanding that the reference 9 to plant design's progress report for January 1979 10 would be the monthly activities report that the
.11 plant design section would submit to the' manager of the 12 engineering department?
13 A Yes. ! 14 Q Have you read the indented and single-spaced 1 15 paragraph that is excerpted from the January 1979 16 progress report? 17 MR. FISKE: You mean did he just read it 18 this minute? 19 MR. SELTZER: Yes, j 20 A Yes. 21 Q Were you aware, before the Three Mile 9 22 Island accident, that as between pumps on throughout 23 'a loss-of-coolant accident and pumps off throughout 24 a loss-of-coolant accident, it had been determined 25 that pumps off or pumps ~ tripped was a worse situation
1 Roy 214
) 2 situation for small loss-of-coolant evaluations?
3 THE WITNESS: Excuse me, would you repeat 4 that question, please. 5 (Record read.) 6 Q And please note I was reading in part from 7 this indented portion when I referred to pumps tripped 8 case remains a worse situation for small loss-of-9 coolant accident evaluations. 10 A Prior to the accident I knew that between 11 pumps tripped early and pumps running, the pumps 12 tripped case was the worst case. 13 Q It was the worser of the two; right? 14 A Yes. 15 Q The reason pumps tripped was the worser of I 16 the two is because with pumps running there is a l l
)
17 forced flow steam coolin'g situation which will exist 18 in the core and which will result in core cooling; 19 right? 20 MR. FISKE: I am not sure, Mr. Seltzer, 21 whether this question that you are asking now e 22 ' assumes HPI flow or not. 23 MR. SELTZER: I don't know whether it does
- 24 or'it doesn't. I think the question'is that
.p-G ' 25 even though.there is.less liquid, inventory with
=______________________..__:_____ - . _ _ _ _ _ _ _
l l 1 Roy . 215 f 2 pumps on, in other words, greater voiding with 3 pumps on because the pumps are forcing coolant l 4 out a break, nevertheless the forced flow of , I 5 steam and water results in better cooling 6 than no forced flow at all. 1 7 MR. FISKE: You mean in those conditions
.8 where there is steam?
9 MR. SELTZER: Right. 10 MR. FISKE: Going across the core? 11 MR. SELTZER: Yes. 12 A All of these analyses assume that one, t j 13 HPI train is on, functioning, and delivering flow. 14 That is a basic premise of the ECCS analysis. Although 15 there are two trains available, analysis requires 16 that you assume one is failed. 17 All of the analyses and conclusions with 18 respect to what is the worst case, pumps running, 19 pumps tripped, assume one HPI train flowing. 20 Q All right. So with that assumption, 21 did you understand before the accident that pumps on O' 22 was a better. situation than pumps off because with L l. 23 pumps on you had forced flow of steam that created l . 24 effective core cooling?
"'" . M;R . FISKE:
l 25 May I hear the question again, 1.
-_ , ~ . _ . ._
1 Roy 216 2 please. 3 (Record read.) 1 4 A We had better than forced flow of steam 5 because you had the HPI train which is provided and 6 making up a portion of the lost inventory, so you 7 have a constant supply of water to the system so 8 forced convection plus continuous supply of cooling 9 fluid from the HPI is a better case, and that was my 10 understanding. It is -- you produce less of a 11 cladding excursion, if any at all, than you would be 12 for a pumps tripped case under the same assumptions. Have you ever learned outside of the
) 13 Q 14 presence of counsel that B&W had informed Met Ed or 15 GPU of the B&W conclusions about whether pumps on 16 was better than pumps tripped?
17 A No, I would have no knowledge of that'. 18 Q Do you have any information about whether 19 B&W had advised any of its operating utilities about 20 the advantange of pumps on prior to the Three Mile 21 Island accident? O -22 A I had no knowledge that we did or didn't.
- 23. MR. SELTZER: I would like to have marked s 24 as GPU Exhibit 450 a memorandum from Dr. Roy-
~
^(1). 25 to Distribution "c,"
Subject:
Technical i
< . - .. . - , . . , - - - - . , , , , - - ,.e, c -- , , - - , -
1 Roy 217 (3 Product Evaluation Project, August 10, 1979. () 2 3 (Memorandum from D. H. Roy to Distribution 4 "C," dated August 10, 1979, was marked GPU 5 Exhibit 450 for identification, as of this g 6 date.) 7 Q Is GPU Exhibit 450 a copy of a memorandum i 8 which you circulated on or about August 10, 19797 9 A Yes, it is. 10 Q Does this refresh your recollection that 11 TPEP was a project instigated in part by President 12 Zipf? (o; 13 A Yes, it does. 14 Q You said in the last sentence of the
'15 second paragraph, "The purpose of the evaluation is to 16 recommend those modifications, if any, which are 17 required to be responsive to the concerns raised as 18 a consequence of the TMI accident and to assure that 19 the product is configured so as to conservatively
- 20 comply with all safety and operational criteria."
! 21 was that a concise statement of what'you 22 believed the pur' pose of TPEP was at the time you wrote l l 23 this? 24 A Yes. V 25 MR. SELTZER: I would like to have marked
.. = . . . . ~ . -
l l 1 Roy 218
) 2 a memo from George Zipf to John MacMillan, 3 ,
may 23, 1979,
Subject:
Fiscal Year 1980 4 Comprehensive Business Plan and Financial gg) 5 Commitment, as GPU Exhibit 451. 6 (Memo from George G. Zipf to J. H. 7 MacMillan, dated May 23, 1979, was marked 8 GPU Exhibit 451 for identification, as of this 9 date.) 10 Q You will note that the sentence which you 11 quoted in GPU Exhibit 45 is underlined in the second 12 paragraph of GPU Exhibit 451. 13 Is GPU Exhibit 451 a copy of a memorandum ( 14 which you received either directly or indirectly in 15 the regular course of business? 16 A Yes. I believe I recollect receiving this 17 indirectly, either by routing or a copy mailed to me 18 from Mr. MacMillan. 19 Q Who is the president of B&W now? 20 A Mr. Cunningham is the president of 21 McDermott, of which asW is a wholly owned subsidiary, 22 and Mr. Vannoy, Walter vannoy, is our chief operating 23 , officer of the B&W operating unit.
,- 24 l Q Where is George Zipf?
25 A He has retired and I don't know where he is
l 1 Roy 219 f% Q 2 I now. 3 Q Have you ever had any meetings with 4 McDermott personnel to discuss any implications of Jgg 5 the Three Mile Island accident? , 6 A No, I have not. 7 Q Have you ever been asked to gather any 8 materials or prepare anything for anyone in McDermott 9 relating to the Three Mile Island accident? 10 A None that I recall. 11 Q Do you still have in front of you GPU 12 Exhibit 40 which was your memo to Dr. Womack?
.() 13 A Yes.
14 Q You were asking Dr. Womack to research a 15 number of questions; right? 16 A Yes. 4 17 ' Q These were questions that arose out of the 18 Three Mile Island accident; right? 19 A Yes. This principally surfaced on my 20 mind as I reflected on the TMI-2 accident. 21 Q one of the principal issues that you were 22 raising with Allen Womack was whether the eight-minute 23 interval before emergency feedwater was initiated
- 24. I made any difference in the outcome of the Three Mile
!, -)
\_/
25 Island accident; isn't that right?
1 Roy 220
<~s h 2 A Yes.
3 Q And specifically to set the stage, at the 4 start of the Three Mile Island accident there was a jg 5 loss of. main feedwaters right? 6 A Yes. 7 Q And because reportedly some valves were 8 shut, there was no automatic initiation of auxiliary 9 or emergency feedwater until eight minutes into the 10 transient; is that right? , 11 A Yes. Yes, this particular -- I had two 12 issues in mind, only one of which could be addressed () 13 quantitatively, and that is whether that delay had 14 any impact on the system's dynamic performance, and 15 that could be analyzed analytically. 16 Q By doing water balance analysis; right? 17 A Yes, just lo'oking at it with and without 18 the feed flow for eight minutes on it. 19 Q Let me show you GPU Exhibit 288, which is 20 Danny LaBelle's response to your questions dated 21 April 30, 1979. 22 Do you recognize GPU Exhibit 288 as being 23 ,. a response which you got from Mr. LaBelle at or about 24 the end of April 1979, responding to the questions
\.j 25 you had raised in GPU 407 1
. = _- -. . - ._ . . _
l t. 1 Roy 221 l-() 2 A Yes. ! 3 Q Is it correct that the answer to the answer 4 to the quantitative questions was that the loss of 5 auxiliary feedwater for eight minutes did not ggg 6 aggravate any of the damage to the plant? 7 MR. FISKE: Could I hear that question 8 again? , 9 Q The question is, did you learn from the 10 response to your questions that the eight-minute delay 11 in initiation of auxiliary feedwater was not 4 12 quantitatively responsible for the Three Mile Island 13 accident or for aggravating the accident? 14 A The only thing that I had learned.from ' 15 that analysis that I asked'him to perform is whether 16 given a sequence that we would model and analyze 17 with auxiliary feedwater flow coming on as designed 18 versus delayed, whether there would be any essential 19 predicted difference in pressures and flows and 20 temperatures and this kind of thing. 21 I had another issue --
@ 22 Q Answer that one first.
1 That is.my question. 23 A so the analytical analysis would say that
- 24 the analytical prediction would not greatly' differ, s/
25 be different but not greatly different with the block g _
l l l 1 .Roy 222 2 valves. You would have a faster pressure recovery 3 after reactor trip, the level would go up faster 4 because obviously it is going to go up if you have an I 5 open PORV and everything is going to try to go to the ggg 6 hole, but there were relatively minor differences 7 in terms of how the parameters would look if you 8 plotted them on a graph and overlayed those two cases. 9 There would be differences but not measurable. 10 Q so you are saying, looking at it 11 quantitatively, in saying what difference would eight
- 12 minutes of water flow have made, you couldn't say
( 13 that the accident would have transpired any differently 14 had AFW been on -- 15 MR. SELTZER: I am sorry, Mr. Fiske, do 16 you want to say something in the middle of my 17 question? 18 MR. FISKE: I thought you concluded. I 19 do want to say something when you have finished 20 it. 21 Q I am just asking you about what you have
.9 22 called the quantitative aspect. I know you'want to 23 volunteer something about a nonquantitative aspect 24 and whether operators were confused by not having
,O auxiliary feedwater on.
25 Since you weren't in the i I l 1
l' Roy 223 p () 2 control room, I am not really going to ask you , 3 questions about whether they were confused about i .t . 4 I do know that you called for quantitative analysis, ggg 5 you got the answers to the quantitative analysis, and 6 I want your understanding of the implications of those 7 answers. The question is: Was it your conclusion 8 that just dealing with the difference between having 9 auxiliary feedwater on as it was intended to be on, U 10 versus not having it on for eight minutes, was it 11 your conclusion that that fact alone did not change 12 the outcome of events on March 28, 1979? () 13 MR. FISKE: Well, Mr. Seltzer, I don't see 14 how Dr. Roy can possibly answer that ouestion 9 15 in the very narrow way that you constructed it. 16 He has a memo here from Mr. LaBelle which gives 17 him specific answers to questions that he asked, 18 and you put that in front of him and I thought 19 that your questions were going to relate.to the 20 answers that Mr. LaBelle gave. 21 MR. SELTZER: They do. 22 MR. TISKE : For example -- l 23 MR. SELTZER: Wait, wait. l "T 24 .MR. FISKE: But then you go on to this 25 broad question'about would it have made any
1 Roy 224 ( 2 difference in the accident sequence if Met Ed 3 hadn't left the block valves closed for the 4 emergency feedwater for eight minutes, and that lg 5 is a much broader question. I adopt fir . 6 LaBelle's answer -- 7 MR. SELTZER: Mr. Fiske, please, there is 8' a way to make an objection that states it without 9 the reporter taking down two pages of your 10 colloquy. 11 I am trying to narrow it, and maybe it was 12 because you were on the telephone when we got
/^)
kJ 13 into these documents and came out of the 14 telephone booth and don't know we have laid a 15 foundation for what I am asking. 16 , Let me, without being unkind, and I 17 understand you had something else important -- 18 MR. FISKE: It was a call.from Chief Judge 19 , . weinberg. 20 MR. SELTZER: I respect that. 21 MS. MCDONALD: Most of the conversation 22 that went on while Mr. Fiske was on the phone 23 had to do with a joke told by Mr. Seltzer off
' 24 the record.
.f~)
~
25 MR. SELTZER: All right, but I am not l
I 1 Roy 225 () 2 asking the reporter to mark when somebody getc 3 up to take a phone call and I don't want to get 4 into that. 5
) Dr. Roy had said'there were two aspects 6
to his consideration of the loss of auxiliary 7 feedwater. One was a quantitative analysis and 8 another was some other consideration. So I 9 have been directing my questions at the
.10 quantitative analysis, and I don't know whether 11 you heard or didn't hear him' describe what the 12 quantitative analysis was and what was considered
} 13 during the quantitative analysis, and I have made 14 it quite clear that my present question is just 15 directed at the quantitative analysis.
16 Now may I proceed just to find out what 17 the answer is on the quantitative analysis? 18 MR. FISKE: Yes, but I don't even think. 19 on what you have referred to as a quantitative 20 analysis that you can isolate the questicn the
; 21 way you purport to be doing it.
-22 MR. SELTZER: Maybe Dr. Roy can shed some 23 . light'on it. I don't think that you are an 24 experienced thermohydraulic dynamicist, and you
~'
25 ! may have to be one to make a statement like the-
'l' i
i
1 Roy 226 2 one you have just made, but I think what is 3 most pernicious about what you are doing, 4 you are suggesting to the witness that he shouldn't 5 give me an answer. ll 6 MR. FISKE: I am not suggesting anything 7 to Dr. Roy. I am making an objection to the 8 question -- 9 MR. SELTZER: Let me finish. 10 MR. FISKE: -- which I am entitled to do. 11 MR. SELTZER: Yes, but Dr. Roy -- 12 MR. FISKE: I want you to be on notice (} 13 that I think your question is defective. 14 MR. SELTZER: You do a good job of putting 15 me on notice and you can do it a lot more 16 succinctly than you have been doing. 17 My problem is, you end up giving 18 a signal, whether intentional or unintentional, 19 to a witness who is one of the brightest 20 witnesses I.have encountered in ten years or 21 more of taking depositions, and I think that he 22 is very sensitive to what counsel may be 23 suggesting, even if it is not your-intention j_s 24 to be suggesting something, so it interferes s 25 with my examination a lot for you, by lengthy < j P 1
1 Roy 227 (_ 2 objections, to be. consciously or unconsciously 3 i suggesting answers to Dr. Roy. I 4 MR. FISKE: Do this for me, will you? Y lll 5 Whether this was done while I was on the phone 6 or not, will you tell me what you mean by 7 quantitative analysis? 8 BY MR. SELTZER: 9 Q Why don't you repeat, Dr. Roy, for the 10 benefit of counsel what you meant when you said that 11 what you were asking Dr. Womack in GPU Exhibit 40 12 was directed at a quantitative analysis of the loss 13 of auxiliary feedwater. ( 14 A The purpose of that question was to compare 15 a computer code prediction, assuming that everything 16 else in the sequence of events stays the same except 17 auxiliary feedwater flow is initiated when it is 18 designed to versus a delay of eight seconds and how -- 19 Q Minutes. t 20 A Excuse me, minutes,'and how would the 21 principal parameters versus time compare, what would 22 be.their differences. That assumes nothing about that 23 the sequence of events could'have been different, i. l 24 whether-it actuated as it was or versus the operators 3 (O~ 25 discovering t' hat it did not perform as' designed,
t 1 Roy 228 OAJ 2 the block valves were closed. 3 Q In other words, you assume high pressure 4 injection goes off at the same time that the block lh 5 valve is closed at the same time that main cooling 6 pumps are shut off at the same time? 7 A It didn't go that far in time. I was 8 looking for the earlier response of the transient.
, 9 It is comparing the output of a computer code with 10 two different assumptions, everything equal; one 11 computer code delays the auxiliary feedwater eight l 12 minutes, the other does not. How do the plots of 13 principal system variables compare.
14 MR. FISKE: I think that clarifies it. 15 Q So you ran the computer codes twice, 16 once with an eight-minute delay and ones with no dealy; 17 right? , 1 18 A Yes. 1 19 Q And was one of the conclusions, just . I 20 reviewing the data from those two runs, that' you 21 could not see any significant difference in terms of
.9 22 impact on core cooling from the eight-minute delay?
23 A During the early time period -- and I (~N 24 - don't know how far we went.on'the runs, about. ten
\_)
25 minutes or so is my recollection -- the parameters
1 Roy 229 () 2 are different but the conclusions is, one, there were 3 no surprises with the dealy, one would have expected I 4 pressure to turn around faster, level to increase lll 5 faster; and, two -- 6 Q You mean but it didn't? 7 A It did. The difference between the two 8 cases, one would expect in comparing a case with 9 auxiliary feedwater as designed versus with the 10 feedwater delayed eight minutes, you would expect 11 that the pressure drop after trip due to -- contraction 12 would be arrested earlier and that the rate of increase () 13 of pressure after the minimum was reached would be -- 14 the slope would be higher, pressurizer level would 15 increase -- 16 Q The slope would be greater? 17 A Greater. Pressurizer level increase 18 would be greater, so there were no surprises, what 19 one would expect just intuitively thinking through 20- in your mind. 4 21 And secondly, as I recall, the displacements 22 and slopes were'not greatly-different; O.K.? I can't 23 l quantify that,.but that is my conclusions. I
~
,_ 24 Q The displacements in slopes were not.
y , \ - 25 greatly.different-between aux feedwater on and aux
-%- w e- --
9 >.- w .F,w- , -m e y 3g .
,r e
/
,a 1 Roy 2'10; e
2 feedwater off for eight minutes; is that what you 3 are saying?, / c 4 A That is ray recollection, yes. , (
/
( g' 5 Q And you carried-the analysis out to a 6 point after auxiliary feedwater came on; is that right? 7: A Y e s..
~
l 8 9 Auxiliary feedwater came on at eight 9 ; minutes arid you- had carried your side-by-side. comparisons 10 out beyond eight minutes; right? Il ' 11 j: A Yes. 12 , Q And at the point at which you stopped 3
\ 13 the side-by-side comparisons, both the system that 14 had had auxiliary feedwater off for eight minutes and 15' the system that had' auxiliary fepdwater on throughout 16 were at about the same position?
f 17 ,. A With respect to primary variable's , yes. '
. -18, .i t goes without saying, and I didn t say it, that 19 -obviously secondary side parameters varied greatly ,
20 because sin one case the steam generator dried out, ! l 21 .the steam prassure' dropped to very low and levels 22 dropped gery low. In terms of primary sysfem response 23 they would come together at' pretty much the same point' l 24 after eight minutes. ' 25 j Q: Did you understand from<the dataIthat you A ' n , ,
. s
_ . - _ _ - - - _ _ - r*
b 1 Roy 231 ( 2 received that if the comparison had been carried out 3 I further in time, that it would have showed no 4 significant divergence? lg 5 A I really didn't analyze that question or 6 arrive at any conclusion on it, based on the results 7 of the analysis that we performed, i 8 MR. SELTZER: Off the record. I 9 (Discussion off the record.) 10 (Recess taken.) [ ; 4 , 11 MR. SELTZER: Why don't we take an early ' 12 lunch today. 13 MR. FISKE: All right. 14 (Whereupon, at 12:15 p.m. a lunch recess
- 15 was taken.) -
16 17 18 i 19 20
.<gp 22 i 23 ,
i n 24
%s
' 2a.
,t
- e. - _ _ _ _ . _ _
j' . s l
i 7 1
232 2 AFTERNOON SESSION 3l 2:05 p.m. 4 DONALD HENRY R O Y, resumed. lll 5 s MR. SELTZER: I would like to mark as 6 GPU Exhibit 452 a. memo drom D. H. Roy to L. J. 7 Stanek, subject: Valve Availability Improvement 8 Program, September 11, 1978. 9 (Memo from D. H. Roy to L. J. Stanck,
't 10 dated September 11, 1978, was marked GPU Exhibit
[ 11 452 for identification, as of this date.) 12 EXAMINATION (continued) h .13 BY MR. SELTZER:
\ 14 Is GPU Exhibit 452 a copy of a memo that Q
v 15 you sent to Manager Stanek in September 19787 i 16 i 'A Yes, it is. 17 Q Is Stanek th'e manager whose section involves
- s- 18 responsibility for valve problems?
19 A He wAs'the manager of that section at that,
\ '
^ ' '
20 time'.
' 4
, s 21 -Q Of the.section that had responsibility 22 for valve problems?
i .. .
. .w;
- 23 A Yes, sir!"
l 4
- 24 Q Am I co[ rect that by September.1978 you
* .25 had formed a belief that valves were causing more A 1 x
---4,-
~r _ __
,, , _ , _., , _ _ _ _ _ , , _ __ , ,.._.,w,, ,
1 Roy 233
'f, 3) 2 outages on B&W plants than you would like to see?
i 3 A Through the review of the equipment 4 outage factor reports, we were finding that valves lll 5 were a significant contributor to lost capacity days 6 on B&W plants and we hoped we could put the program 7 together to improve that record. So, yes, more than 8 we would like to see. 9 Q You said that you had a review of a valve
- 10 a rallability improvement program on September 7.
11 Do you remember who participated in that? 12 I .sssume Stanek did. 13 A Yes. I don't recall the meeting ( 14 specifically. But no, I don't recall who participated 15 in that. 16 Q I take it from this memo that you wanted 17 to initiate programs to study what was causing valve 18 failures and improve reporting and follow-up. 19 A Yes, this was part'of an approach, overall 20 availability improvement. One of the elements was 21 valve problems that might be failure or leakage, 22 'that would bring the plant down or require excessive 23 maintenance, perhaps extended downtime, a wide 24 variety of valves, both in the NSS.and the balance of
'~') 25 the plant, that contributed to these lost capacity
1 Roy 234 7 2 days'. We wanted to try to highlight now and see what 1 3l kind of program we could propose to work with our l e 4' customers _to get those lost capacity days minimized. llk '5 Q Among the valves that your program was , 6 to be directed at would be pilot operated relief 7 valves; right? 8 A I don't specifically recall the listing. 9 one of the intents of this program was to get a 10 - listing of.all of the valves that were making some 11 . significant contribution to lost capacity days, 12 and I' don't recall whether the PORV was in that } () 13 matrix' explicitly or not. We were dealing with a 14 global issue of many vendors and many kinds of valves. 15 Q When you say " global," you don't mean to j 16 limit it to globe valves, do you? 17 A No, in a broad sense. 18 Q That.was a joke. 19 A Yes.
- ~ 20 Q There is nothing.in'this~ description.
21 of the-program in GPU_ Exhibit 452 that in any way 22 excludes pilot. operated relief valves'from the valve
- - 23 improvement program? .
24 A- That'is true. -I didn't limit the program
&em 25 to any specific class of1 valves, just identify those
=
W .,r -- g -p,-+y.+--,.--.ec- , ,y egr- y w -+y r
,y-9y, g- *-yw -
31ip g g w se --cy-9e w,-9-mwwy- +
1 Roy- 235 e w ( 2 . kinds-of valves and kinds of problems that were l 3 Produc.ing large lost capacity days. 4 Q When you refer in item 1 to " improve our gg) 5 knowledge cf the failurec which are' occurring in the 6 field," you meant failures of valves to operate as they 7 were intended to operate; 's i that right? 8 A It could include that, but not be limited 9 to that. Others are leaking valves that can either 10 bring the plant down or, if.it is down, require 11 excessive maintenance. One of the-elements was 12 occupational exposure on maintaining valves, so it () 13 woul,d include that but not be limited to that. Each 14 of these issues has to be examined for what kind of 15 . service that valve is performing, the environment it is - 16 immersed in, the nature of the problem.that could be 17 causing lost capacity days or excessive maintenance, i
-18 and the intent of the program was.to begin to see.how 19 we could work with our customers, get the information 20 we needed and help them reduce that contribution.
l 21 Q What sparked your interest in the problem l 22 of valve failure's? 23 Q I mentioned our equipment outage factor w 24 report called the EOF report. We received data which
)
25 is computerized' and classifications or categories
4 1 1 Roy 236
)' .2 of items which are contributing to lost capacity days 3' are displayed in the output, which is this EOR report.
4 It will identify various peieces of equipment or l 5 other categories that can be. causing lost capacity . 6 days: human error, NRC, pumps, control rod drives, 7 stators, valves. Sort of a pie, if you can imagine 8_ a graph representation using the shape of a pie. 9 valves was contributing a significant portion of _10' lost capacity days and that looked like an opportunity 11 to try to improve the overall performance of the
-12 = plant by going after that class of equipment problems 13 that were causing lost capacity days. We hoped that 14 would be a service that we would join with our
. 15 customers and they would work with us and improve the 1 16 overall performance of the plant. , 17 And it would make the B&W NSS -- I.would 18 like to be able to point to the NSS, B&W NSS, the 19 l plant assembly, and get the availability as high as 20 possible. 21 Q You knew it was a smudge on B&W's 22 escutcheon :to have excessive valve failures? . 23 A The overall availability record. I would i 24 like to get that performance up as-high as possible
- (A'j 25 on it'. And primarily'it's being driven by being able y -ar . p ,,,v.,. r v y - + g- -*.p.-,...wew.,y..yev.-- _ . , - - .,e w. - , sr,, ,yw..4-,- ,g we 9
l l l l 1 Roy 237 l 1 (g ) 2 to offer a service of ours to our operating plant Am/ l customers. Those were pretty much in a service 3l 4 mode with the operating plant customers. lll 5 Q Y u wanted to initiate action and one of 6 the things you wanted to get in item 1 was improved 7 reporting. g What type of improved reporting did you want 9 to initiate? 10 A I don't exactly recall my thinking at that
'l l time, but I think I can recreate it today but --
1 12 MR. FISKE: Well -- 13 Q Let me suggest, I would like you to use 14 this document as a refresher of your recollection. 15 If this triggers a recollection of what improved i 16 reporting you were looking for at that time, that is i
! 17 what I am asking you for.
18 MR. FISKE: That's fine. We want a 19 recollection, not ~ a reconstruction. 20 THE WITNESS: O.K. 21 A' This item would refer to getting a better 1 22 communications path with the customers so that we could 23 get the data on specific valves, the service 24 environment, maintenance records, and so forth, which
-~'
25 is required;before you can implement a useful program
._ . _ - _ _ , . . . _ . _ . __ -- - ~ . . . . . - . -
1 Roy 238 i q to come up with improvements, so this would be j 2 3 primarily the reason I mentioned to Dick Kosiba 4 couplings made through the service department with the g 5 customers. 6 Q What kind of improved follow-up methods 7 were you looking for? You said you wanted improved 8 reporting an,d follow-up methods. 9 A Yes. Based on, and our surmise here -- 10 MR. FISKE: I don't think Mr. Seltzer 11 wants your surmises any more than any of the 12 rest of us do.
/~h 13 A I believe this is,what I had in mind, V
14 that once we had identified the nature of the problem 15 and we had mechanisms or came up with ideas for 16 improving it, that we get that back out to the 17 customer. That is a follow-up method. 18 C And you wanted to see improvement on the 19 follow-up at B&W? 20 A Yes, coupling both ways with the customer, 21- once we get the information in, we have to have that 9 22 before you can generate concepts for improvement, and 23 . then once we get them back to the customer and get 24 ' some determination of whether he is interested in
'~
25 working with us on such a program. p .c ,- -- , , ,.- -r--,, -- ~ ,e r -
.-w, r , , , - - ,,-y , - - - .
1 Roy 239 l r ' (j 2 Q You were the head of the engineering 3 department at the time you wrote-this? 4 A Yes, sir. You said that you wanted to initiate l 5 Q 6 action at the interdepartmental level, not the 7 intradepartment level. 8 What were the departments that you wanted 9 to see involved in this improved knowledge of failures 10 and improved reporting and follow-up methods? 11 A The only department that would come to mind 12 that would be involved to any great extent in this
/~) 13 would be the customer service department, so --
\_/
14 Q Where did R&D reside? 15 A At this time, in engineering. 16 Q so you wanted to get the customer service 17 department and your engineering department to work 18 together on improving availability on valves? 19 A Yes, That was actually a thrust of both -- 20 really both Dick and I, this-whole availability 21 improvement thing, and we would have to work together 22 as a department to pursue these various items. 23 Q Yes. l 24 A. Anyway, it would be principally the
'~'
25 customer service department. t
=
l
1 l 1 Roy 240 (f 2 Q Can you describe briefly what the " valve 3 categorization and failure matrix" was to which you 4 were referring in item 27 lll 5 A Yes. What I had in mind there was that 6 we generate an overall picture of this valve issue, 7 that we list plant by customer, valves that were in 8 service, the contribution of those valves to lost 9 capacity days, as best we could forret that out. 10 Q Ferret, is that.what you said? 11 A Ferret. 12 Q I thought he might have written.down (s 13 " ferry." 14 A That we would list the type of valve, 15 the service to which the valve -- in which the valve 16 was employed. These would be the kinds of items 17 that you would build to get a picture of the valve 18 situation. From that you can then focus on those 19 particular ones that were major contributors to lost 20 capacity days or occupational exposure due to 21 maintenance,.whatever. You had to have that kind of 22 information because there-are a large number of valves t 23 -and obviously a large number of vendors.and a large 24 number of services to which' valves are' assigned. 73
^)
( . That is what.I had in mind. 25
1 Roy 241 2 (_/ Q Was there a place in the matrix for a 3 description of what had caused the valve failure or 4 what B&W believed had caused the valve failure? lll 5 A I don't recall. I just can't picture it 0 right now in the matrix. I just can't recall -- I I can't picture that item right now in the matrix. O Q Do you know whether your action plan in O
, GPU Exhibit 452 ever was acted upon?
A In generating the matrix? 11 Q No, in doing 1, 2 and 3. This sounds like 1~0 a good idea. 13 A Yes. 14 Q I am asking, was anything done to implement 15 it? ' 16 Q We did generate a failure matrix. This 17 issue of availability, including a focus on valves, 18 was discussed with customers'-- I don't recall
- 10 specifics on it --
along with other suggestions for 20 improving availability. Not much response on this 21 valve issue. Very little. Those were done, as best I can recall. 2 1 MR. FISKE: A response from whom? I 24
' %. THE WITNESS: Customers, on the valve. issue. %/
25 Q Did B&W do anything to try to improve its
I l . l
-1 Roy 242 I l
1 (~)j 2 follow-up on the valve failure information that it ' 1 i 3 already had in its control? 4 A Part of that would be the matrix. Taking 4 lll 5 that data and casting it in a form that would give 6 us better insights where we could suggest focusing 7 attention so you get the maximum improvement is one, 8 sorting with the data that we had. . 9 Q The matrix is just a first step though. 10 A That's right.
-- 11 Q once the matrix was developed, what, if 12 anything, did B&W do to improve the availability of 13 its valves?
v j 14 A I can't think of anything, and I don't 15 know really what you could do. I mean the valves are d 16 something that weren't deficient. The whole thing of 17 coupling to get up-to-date information and complete 18 -information was the important thrust of it. 19 Q Did you'ever follow up on this to try to 20 encourage the implementation of the valve availability 1 21- improvement program? 22 A Personally? I had reviewed the matrix, 23 these kinds of items-that we have put together in 7_s 24 engineering, but I don't recall being involved where t N 'r) 25 I was making a presentation to the customer of this
-. . - , . - - , ,, - - . ~ , _ ,r-.- ,, - , . -,,,,-c.-- , - , - , - - - ,
1 Roy 243 im idea or concept, but I have a clear recollection that ( ) 2 i 3l it has been done through service managers, service 4 personnel, owners group, users groups where we
~
jgg 5 presented not just the valve but other elements of 6 the overall availability improvement program to get 7 .the customer interested. I do know that very little 8 interest in pursuing it, a valve program on the part 9 of the customers -- 10 Q I heard you say that. 11 A Yes. You can't make a move in these things 12 without a customer being a key actor, a key role.
/~ 13 Q The valves that you were referring to C}
14 in GPU Exhibit 452 were valves within B&W's scope 15 of supply; right? 16 A Not necessarily. 17 Q You certainly were including the valves 18 that were within B&W's scope of supply? 19 A Yes, we weren't excluding any of the valves 20 in service that were contributing to lost capacity 21 days. 22 Q so your matrix would include many steam 23 , isolation' valves? i 24 A Yes. 25 MR. SELTZER: I would like to mark a
, , . . - _ . . ,, -- . g --, ,
1 Roy' '244
-\ O ,/ 2 memorandum from D. H. Roy to L. P. King, 3 September 11, 1978, regarding valve problems.
4 (Memorandum of September 11, 1978, from lll 5 D. H. Roy to L. P. King, with attachments, was 6 marked GPU Exhibit 453 for identification, as of 7 this date.) 8 MR. FISKE: Do you want to read it? 9 (Pause.) 10 Q Is GPU Exhibit 453 a copy of a memorandum 11 which you sent to L. P. King on or about September 12 11, 19787
) 13 A Yes, it is.
14 Q. King was your manager of fluid systems? , 15 A Yes, he was a unit manager -- I don't 16 recall the exact title -- of a unit in the fluid'
- 17 systems and mechanical a'nalysis, mechanical section, 18 equipment section.
19 Q He reported to Stanek? 4 20 A Yes. 21 Q P. E. Perrone was a special assistant to 22 the engineering department manager; right? 23 A Yes, he was. g 24 Q Had he been special assistant to.the
,f l'
( ! w/ 25 engineering department manager before you became --
. . - - - . . . . -. - - . . _ ~ . - _ - . . . . - , .- . . - - - - - . .-
1 Roy 245 2 A Yes.
! Q -- manager of the engineering department?
3 4 A Yes, he was. lll 5 Q Had that been his role for a year or more 6 before you came in as engineering department manager? 7 A I would say at least a year before. 8 Q The SPR's and SPR comment forms that are 9 attached to your memo by you refer to valve problems;
- 10 right?
11 A valve and operator -- valve operator 12 problems. 13 Q What is a valve operator? 14 A The motor, either electric or pneumatic, 15 which operates the valve. 16 Q Drives the valve? 17 A Drives the valve. 18 Q Your memo to King is dated the same as your 19 memo to stanek,'GPU Exhibit 453. 20 Are the valve and valve operator problems 21 that you are writing about in 453 items which had 9 22- come to your attention in the review of the proposed 23 . valve availability improvement program that you ' 24 described in 452? [
\~# A 25 I don't think so. The best recollection i
1 Roy 246 (_, 2 I have is this meno here dealt more with closecut 3! of.the SPR, that these had been pending, a lot of them 4 already had action taken with respect to them, but it 5 was getting them closed out in terms of signed off and l
.6 submitted back to customer service engineering , 7 as resolved. It is closing them out in a formal way.
8 Q Finishing up whatever action needed to be 9 i taken on them? 9 10 A closing them out. 11 Q Is that right? 12 A Yes, that is what I think -- to the best 13 of~my recollection, that's what this particular note [) x_- 14 was driving at, was getting the SPR closed out. 15 Q So you are saying that this is a package
- 16 of SPR's and SPR comment forms all relating to valves, 17 and you think that these came to your attention 18 independent of the valve program that you were 19 discussing with Stanek in GPU Exhibit 452?
20 A Yes, my best recollection is that they 21 were brought to my attention by-Mr. Perrone. 22 Q How is it that Perrone had been functioning 23 on these valve problems to.the extent of writing 24 SPR comment forms on the valve problems that are 25 included in 4537 i i
1 Roy 247
? ~ .(. 2 A Mr. Perrone operated both for myself and 3 for Mr. Deddens, technical staff capacity, quite 4 knowledgeable in equipment-related issues, and this would have fallen under his purview in terms of --
lll 5 6 he may have on his own, and my best recollection is 7 he did, identify that these were still outstanding 8 and suggested that I send the note to the unit manager 9 and encourage -- get those closed out. That would 10 be my best recollection of how this came through. 11 And he would work in that area as a technical 4 12 consultant looking into special issues in the equipment 4 13 area. He was quite good in the equipment area. 14 Q Did he have a subspecialty in valves? 15 A No, I don't think there wuold be a 16 subspecialty. 17 Q He brought this package of valve problems 18 to you that had not been closed out. 19 Do you have any understanding why he 20 picked out valves as an equipment area to bring to 21 your attention? 22 A No,'I don't know why he would do that, 23 other than.this was a collection of outstanding SPR's
~g 24 that he was encouraging a closeout on, and I don't i I 25 know why they ended up to be dealing with valve and
1 Roy 248 l l b). ( 2 valve operating problems. 3 Q By an outstanding collection, you mean it 4 was an excellent collection? ggg 5 A No. The closure was still outstanding. 6 Q Did you know that in the package that Paul 7 Perrone sent to you he had included the failure of 8 the pilot operated relief valve at the Davis-Besse 9 plant in the fall of 19777 10 A I see that that is in the package now. 11 Q Do you see that in the last page of the 12 document that you put together there is a description J 13 of testing that was conducted on the pilot operated
, 14 relief valve at the Davis-Besse plant shortly after i ;- 15 the September 24, 1977 incident? ,
16 A I see that description attached here. i 17 Q' And that is a description of testing that 18 was written up by B&W; right? 19 A Yes. I would assume this was prepared 20 by the originator of the SPR. 21 Q Who is that? O - 22 A I don't know that-for a' personal fact. 23 MR. FISKE: I will object to that. I I j3 24 think Dr. Roy'doesn't know where this comes from. tx.s)
- 25 Q Do you recognize your company's logo in I 9 + w wr v-r -vv,., , - - , v rm < =w. -, w v-wr y w
1
- 1 Roy 249 !
() 2 the upper left-hand corner? 3 i A Yes. 4 A And J. E. Anderson is somebody you jg 5 recognize as an employee at a&W7 6 A Yes. 7 Q And the site problem report form which is 8 the next to the last page is a B&W form; right? I 9 A Yes, it is. 10 Q And a B&W employee would fill out the 11 boxes at the bottom; right? 12 A Yes. 13 Q So there is field change request and there 14 is an opportunity to check off yes or no, and that is 15 a box that would be filled in by B&W? 16 A Yes. To my knowledge, yes. 17 Q And where it says "F.C.A. No," what does 18 F.C.A. stand for? 19 A This is a field change authorization. 20 Q And somebody has written in " N/A." Do ' 21 you understand that means not applicable?
@ 22 A Yes.
23 Q And do you see the other box in that line l 24 I where it says significant' deficiency"? 25 A Yes. .
,.-- <g.. ,+- 1,- 4env_-,--ee ay---- mm ,,,.4y.oy.,4 -- _ ,4y ay--,n.,y sp_e-- _-
3 p w _ - -g.yy
1 Roy 250 i ( 2 Q And under that someone has checked off 3 "No." 4 A I see that. 4 gg 5 Q Significant deficiency is a reference to
- - 6 a significant safety deficiency, right, in the 7 terminology used in your preliminary safety concern j 8 classification; right?
9 A Yes, with respect to the PCR -- excuse me, 10 the PSC procedure, that is a determination that is 11 .made through evaluation of the PSC. What this 12 determination here is relative to the SPR and the 13 procedure at this time, I don't know. 14 Q Are you saying you don't know what the 15 yes and no signify on a site problem report form? , 16 A The term "significant deficiency" as 17 related to this SPR and 'its relationship to part 21 -- 1, 18 Q Yes. 19 A -- I don't know. 20 Q Let me ask you a different question. 21 Do you understand that as site problem 9 22 reports were being used in this form, which is part 23 of GPU Exhibit 453, the notation "significant deficiency" i j_ 24 referred to - significant safety deficiency? 4 25 A That is.what I don't recall, within the
- + - ,,,. , - , -,- .,, .-v.- , - - , . - - - - , , , , , , - , - - -
1 Roy 251
); 2 context of the SPR, the SPR procedure, what that 3 block designation means, what its interpretation 4 as contained on the SPR. That is what I don't know.
5 I can't recollect exactly how that is referred to in ggg 6 the SPR and what the interpretation of it is in the 7 context of an SPR. S Q I take it you were attaching the item 9 relating to the Davis-Besse valve failure because of 10 Paul Perrone's comment on the SPR comment form which 11 is the third from the last page; right? 12 A What is the question again? 13 Q You were attaching the material about the 1-4 Davis-Besse valve failure because of Paul Perrone's 15 comments on the third from the last page? 16 A No, I don't recollect doing that. 17 My recollection of this is a group or i 18 package of SPR's. I have no recollection of reviewing, 19 you know, the individual SPR package, but attaching
- 20 a note which encouraged the closeout of them. I 21 don't recall reviewing, you know, these SPR's singly 9 22 at that time.
23 l Q consistent with your instructions that you l l 24 wanted these SPR's and SPR comment forms closed out, f'- 25 what action did you expect would be taken on the I r I
1 Roy 252 l
) 2 Davis-Besse valve modifications referred to in Paul 3 Perrone's SPR comment form?
4 A I don't recall building any specific expectation, you know, with respect to individaully, gg 5 6 or technically, with respect to collectively the 7 package. The key thrust, as I recall, was to get these 8 closed out and off the books as open SPR's. Just in 9 that context. 10 Q Perrone asked two questions in the third 11 from the last page of Exhibit 453. His first is, 12 "Why_ don't the modifications to the valve made in the
/~'\ 13 field require a field change?" His second question
(_) 14 is, "Should these modifications be incorporated on 15 other valves to prevent similar failure?"
- 16 In suggesting that;there should be 17 closeout, weren't you suggesting that these questions 18 should be answered?
19 A I encompassed that entire concept-in 20 my letter by saying." closed out as soon as possible 4 21 with Paul Perrone." I delegated to Paul to oversee 22 the closeout of these things and, you know, that was 23 up to him. He raised questions and asked questions, I _ 24 and review it and see that the things were closed
- 25 out as quickly as possible.
,- ,- , , , ~ , . , , y. _ ,-- _ , . , . . . ., f m.,, -,y,_ .- ~ . 7
I 1 Roy 253 2 Q Did you ever follow up to see whether any 3 subsequent action was taken on any of the items that 4 you' attached to GPU Exhibit 4537 5 A Not that I recall.
.6- Q Do you have any knowledge whether any of 4
7 the modifications to the Davis-Besse valve were 8 incorporated on other pilot operated relief valves 9- to prevent similar failure? 10 A At this time no, or as a consequence of 11 this -- of this item. 4 12 Q You got this memo from Paul Perrone, 4 13 you got -- and by "this meno," I mean the last three-14 pages of GPU Exhibit 453 -- you got the matrix that 15 was developed in response to GPU Exhibit 452. 16 Did you take any action that you can recall 17 to get B&W to do anything to improve pilot operated 18 relief valves before the Three Mile Island accident? 4 19 A Not specifically related to that valve. 20 That would be covered. I didn't exclude it from review 21 as part of the overall availability improvement 22 program. It wouldn't be included -- I don't have any 23 recollection of directing it be included or not be - l 24 included. The whole idea of this program here.was to O
\/ 25 determine where best to put efforts to improve overall I
1 Roy 254 I 2 lost capacity days, so I wouldn't have excluded it 3 or I have no recollection of directing its inclusion. 4 .It was part of the overall review. lll 5 Q Do you understand from Perrone's report 6 and the other pages attached at the back of GPU Exhibit 7 453 that during a test of the Davis-Besse pilot i 8 operated relief valve the valve stuck open? s 9 MR. FISKE: Are you asking him, did he 10 understand that back in September 1978 when he 11 got this package? 12 MR. SELTZER: Yes. () 13 A I recall no -- or have no recollection of 14 reviewing any of these SPR's for the detailed content 15 at all. 16 Q After B&W apparently found out that the 17 Davis-Besse pilot operated relief valve had failed 18 open on September 24th, and then failed open in a 19 test afterwards, do you know of anything that B&W , 20 did to advise other owners of B&W plants to check
- 21 their pilot operated relief valves to look for 22 sticking open problems?'
23 ! A I have no knowledge of whether we made i t 24 l any communication - or not with respect to that incident. i
.x'^'
O> g 25 You don't know whether they gave any
Y 1 Roy 255 i- { f 2 warning to owners of other B&W plants that their 3 valve might bind in the open position? l 4 A I have no knowledge or recollection that i 5 they did or didn't. - 6 MR. SELTZER: I would like to mark as 7 GPU Exhibit 454 a memo that Dr, Roy wrote to 8 , Taylor, Phinney, Spangler, LaBelle and Vosburgh 4 9 on May 30, 1978,
Subject:
Procedure Review 10 for Safety Implications of Site Problems 11 and Operational Upsets. 12 (Memo dated May 30, 1978, from D. H. Roy 13 to Distribution, was marked GPU Exhibit 454 for 14 identification, as of this date.) i 15 Q Is GPU Exhibit 454 a copy of a memo which 16 you sent on or.about May 30, 1978? 17 A Yes, it is. 18 Q You were telling people you wanted them 19 to attend a meeting on June 5, 1978; right? 20 A Yes. I-21 Q Snd you were convening the meating in 3 22 order to see whether procedures needed to be developed 23 . or modified to assure adequate review of safety 24 implications of site problems and operational upsets;
.\ O) 25 right?
t
r 1 Roy 256 2 A Right. 3 Q What triggered your feeling there was a p 4 need for this kind of review? 5 A I have no recollection of this meeting at
-llh 6 all right now.
7 I don't recall what would cause me to 8 convene that meeting today. 9 Q Let me show you again GPU Exhibit 256, 10 which was'your memo to Taylor and Phinney attaching 11 Vosburgh's meno. That is dated May 16, 1978. 12 Let me also show you GPU Exhibit 455, which () 13 we will now mark, Roy to Distribution,
Subject:
Site 14 Problem Review for Sa5ety Analysis Impact, June 8, 15 1978. These are your notes of the June 5 meeting. 16 (Memo from D. H. Roy to Distribution, 17 dated June 8, 1978, with attachment, was marked 18 GPU Exhibit 455 for identification,.as of this 19 date.) 20 Q First.let me ask, is GPU Exhibit 455 a 21 copy of a memo that you sent to other managers at 22 B&W in early June of 19787 23 , A Yes, it is.
- 24 Q Have you had a chance to review GPU x/
25 Exhibit 4557
r + 1~ Roy 257 4 W 2 JL Yes, I have. t 3 Q Does any of the material that I have put 4 4 in front of you refresh your recollection as to what jlg 5 triggered your convening the. June 5 meeting? 6 A I have no recollection of this meeting at 7 all. It is obvious from this memo that it was 8 following up on this memo from Bob Vosburgh and it i
.. _ , 9~ seemed pretty obvious from the material here.
I 10 It does trigger a thought with respect to 11 .some procedure modifications in the -- on the SPR i ~12 form, and I don't know whether that occurred after ( 13 this or not, but I am aware that at one point in time 14 on the SPR we-called out on the SPR, provided a block 15 for determination of whether a PSC, you know, is-16 .this a safety concern, because that would then. tie
-17 it to the PSC procedure. I'do-have a recollection 18 of that event. That could very well have been a 19 - follow-on to this. My best recollection is that'it was.
20 One of the results of'this kind of thing. But I 21 don't recall the meeting with.these people at all. F 22 I just don' t recall 'that particular meeting with them.
- 23 .Q Are you-saying that you do' recall that at
'24 some point prior to the Three Mile' Island accident 25 =a" decision was made to change:the, site problem report-I c
- y . - < , - a - w, r- -
, - - - ~ , .-,v , --r, ,= ----m,.-n&
.,~ :
.j
)
~
1 Roy 258 _ H G = k-) , 2 form to require the originator of the form to indicate
'{s !
)
f 3 whether there was a significant safety problen j
.I 4 spotted in the site problem report? ~
~
ll) 5 A Yes. It w a.s to provide an opportunity, 6 a more clearly designated place on the SPR to~ indicate i 7 that this particular issue, whatever the SPR was 8 addressing, was a safety' concern, and then that would 9 coupleLit to the PSC. If-it was a safety concern, I
- 10. he would write up a PSC for it.
~
11 I belleve we did, we thought that that 12 ; would be a useful enhancement of the SPR form, to (~% ' (,) 13 provide that particular blocks on the SPR. 14 Q Did you agree with that modification? l 15 A. Yes. [It in no way excluded using the SPR. l 16 We would go ahead and use the PSC but this pernitted 17 ^ it to be called out very specifically on the SPR form 13' liself, and I thought that was a good enhancement. 1
- 19 Q Was the intent that the originator of the l 20 SPR would probably be one of the most knowledgeable 21 people about any safety concerns that arose in 22' connection with the site problem?'
23 A 'That the originator would have the first 24 opportunity to make that determination and check it. {~
-s_-
25 That wouldn't. preclude, you know, others that might 4M F r I- -..i-
.m i- .
'l 1 Roy 259 I
v) 2 receive it from identifying it as a safety concern. 3l Q After that modification in the site problem 1i 4 report form was developed, was there any program that ggg 5 you are aware of to go back and review prior site 6 problem reports and see if they had raised'significant 7 safety concerns that should be called out? s 8 A I have no recollection of.that occurring. 9 Q Do you see the third sentence in the 10 second paragraph of GPU Exhibit 455 where it begins, 11 "It was the consensus..."? 12 A Yes, I see that. [ ; 13 Q You said, "It was the consensus of those L/ \ 14 in attendance that encouraging service managers, 15 licensing engineers and plant integration personnel to 16 obtain the advice and consultation of safety analysis 17 personnel at the front end of any significynt site ' 18 problem would do much to help alleviate the problem." t 19 Was it in part to implement this consensus 20 that you drafted the action items on page 27 21 A well, I don't recall the meeting, but O 22 these action items obviously followed from that, from i 23 these observations that I have summarized here in f _- 24 the minutes. 25 Q on page 2 where you described the action
.-y
-s
=g 4
1 Roy 260 p 2 items generated during the discussion, you have listed 3 names after each of'the four action items. 4 were the individuals named the ones who g? 5 vere. supposed to follow up on the action item listed? 6 A Yes, I believe that would be the meaning t 7 of those names.
,_.- 8 . Q The first action item states, " Assure that i t 9 a Plant Integration is reviewing SPR Forms for potential t
7 10 safety impact and refer to safety Analysis as 11 ,
'apprrpriate."
s Eight dots and then "D. H. Roy."
# 4
.t 4 12 You were the one.and you assigned to
,' 13 yourself the responsibility for assuring that plant 14 integration was reviewing site problem report forms N, -
and looking for potentia'l safety impact; right? 15 / w
~ .
Q~
-- 16 A Correct.' w s- i *
'% s
\ '
~ , , , 17 Q And you vere going to check up to see that 3,
1,
, 18 they referred potential safety impact questions to the t .
,, I p, 19 safety analysis unit; right?
20 A Correct. 21 Q Did you delegate to anybody else the 2* k responsibility.for following up on_that action item? 23 A The only recollection I have of any s. 24 l follow-up on that, and I am not,sure how it was
'J
~,
25 implemented, it seems to me that at some point'.in ' is i- ., 4 ib
'5 5 i -
2 .-
. [. %
,- **= '
;, , . -.x- _,. .n - -. . - - --
I Roy 261
,m I s 2 time I either sent a memo which would require that 3 SPR's go to plant integration, or certain SPR's --
l 4 I don't think we captured that in the SPR procedure -- ll) 5 but I do recall some follow-up in either encouraging 6 or directing that certain SPR's go to plant integration 7 and whether I delegated that or wrote it myself or 8 what, I don't recall. 9 Q That gets the form to Bruce Karrasch's 10 integration group. 11 Do you recall doing anything to make sure 12 that plant integration reviewed the site problem () 13 reports for potential safety impact problems? 14 A I don't recall any action that I may have 15 taken on that element of the item, the action item. 16 Q Do you recall doing anything to follow up 17 and make sure that plant integration would refer any 18 safety impact problems to Danny LaBelle's safety 19 analysis unit? l l 20 A I don't recall any action I may have taken i. 21 with respect to that portion of the item. 22 Q If you took any action in writing, we , 23 .should be able to find.that in your chronological file: J r^s 24 right? ( 'O j 25 A It should be, yes. You should be able to ,- p _. ._s_
4 1 Roy 262 yes. 2 3 Q I take it at the time you thought potential 4 safety impact problems should be referred to safety lll 5 analysis, you weren't excluding the appropriateness 6 of sending some safety impact problems to ECCS 7 analysis, were you? 8 A That's correct. This sequence didn't 9 suggest to me that, you know, to my own personal 10 knowledge that safety analysis was not sufficiently 11 involved, you know, when there was a specific item i 12 which they should address on it. It just didn't -- () 13 that is not the thought I had in mind on it. 14 Q O.K. 15 MR. SELTZER: Shall we take a brief recess? 16 MR. FISKE: Fine. 17 (Recess take'n.) 18 BY MR. SELTZER: 19 Q You have said at various points in your 20 . testimony that you were concerned about improving the 21 acquisition of information about. problems in the fields 22 right? IG A Yes, specifically we and I consider that
' ~
24 .an importan't element in terms of building an effective
-t O) -
\_/
25 availability improvement' program, i
1 Roy 263 (D (_) 2 Q Did you ever develop any program for i 3l encouraging site personnel, B&W site personnel, to l 4 increase the accuracy and flow of information that
) 5 they were sending to B&W7 6 A I don't recall any initiatives along that 7 line that I am aware of.
8 The principal thrust of my view on that 9 flow of operating experience is that again we get the 10 customer to join with us in a program because he has 11 the data and could permit the access that could allow 12 us to get the data, so my principal thrust is in terms 13 of getting it, a coupling with the customer so we can 14 approach these problems. That would be more what I 15 had in mind. 16 Q So are you saying you ignored any attempt 17 at improving the flow of information from sites 18 where B&W had its own site representatives who could J 19 get the information? l 20 A No. I am not aware of any initiatives 21 along that line. There may have been some, but I am 22 not aware of it. 23 MR. SELTZER: I would like to mark as GPU f l.
- ~g 24 Exhibit 456 a memorandum from Dr. Roy to L. %.) l l 25 Fahland, Setpember 11, 1978,
Subject:
)
I 1 l L
1 Roy 264 (~' - _ (_)\ 2 Availability Information Flow. j i 3 (Memorandum from D. H. Roy to F. R. ! 4 Fahland, dated September 11, 1978, with lll 5 attachments, was marked GPU Exhibit 456 for 6 identification, as of this date.) 7 Q Is GPU Exhibit 456 a copy of a memo that 8 you sent to Fahland in or about early September 1978? 9 A Yes, it is. 10 Q You were suggesting follow-up on the 11 " Availability Improvement Action Plan" that you attached 12 to the memo; is that right? () 13 A Yes. I don't recall this particular memo 14 or -- I don't know if there is an attachment -- 15 Q In the first sentence you said, "The 16 comments put together by Bob Burnley, Harry Honig, 17 and Larry King dealing with the 'information cycle' as 18 related to availability problems in the field is 19- deserving of followup." 20 Had those gentlemen put together the 21 attachment, to your memory? 22 A I can't recall. I don't have a l 23 i recollection of these comments and whether that is --
.te g- 24 comments as referred to here in my memo of september O 25 11th relate to this attachment.
1 Roy 265 O k- - 2 Q Are you saying that although you sent GPU 3 Exhibit 456 to Fahland in September of 1978, you 4 don't remember -- lll 5 A I don't recollec.t this construction. 6 Q Do you remember putting this attachment on? 7 A No, I don't. 8 Q Did you agree in September of 1978 tl.a t 9 the SPR procedure "has not been efficient or effective"? 10 MR. FISKE: Where is that? 11 MR. SELTZER: Page 2 of the attachment. 12 MR. FISKE: Page 2 of the attachment? 13 MR. SELTZER: Yes. 14 Q Right here (indicating). 15 A I don't recall coming to that conclusion 16 with respect to this memo or in general. The basic 17 _ thrust of the information flow, of which the SPR 18 can only be a part of it, that for an effective 19 availability improvement program you will need more 20 than just SPR's to address the problem, so I don't
- 21 recall coming to this conclusion for this, and I
, 22 wouldn't necessa'rily say the purpose that the SPR
-23 serves might not be efficient or effective for its - [~N.
24 specific means, but if you are going to undertake an Q) 25 availability improvement program, you've got to be
1 Roy 266 N 2 supplementing with a wider base and a larger program 3 to acquire this information. I 4 Q Take a look at the SPR checklist that'is 5 included in your attachment,- page 29 in the lower
)
6 right-hand corner. 7 This is someone's proposal for a checklist 8 tg aid the originator of a site problem report in ^ 9 gathering facts required to arrive at a meaningful 10 resolution. 11 Take a look at item 10 and tell me whether 12 in 1978 you formed any conclusion as to whether the w
) 13 items that are described in paragraph 10 were items 14 that you thought should be gathered by the originator of 15 a site problem report.
16 A I have no recollection of the checklist, 17 nor have I drawn any conclusions with respect.to item 18 10. 19 Q Was it your understanding that a site 20 problem report should indicate if equipment was not 21 operated in a n.anner that satisfied the inte1ded 22 B&W purpose? i .- l 23 A I don't have any recollection of reviewing l f'%g 24 the scope of the SPR and making any conclusion of what
'NJ -
25 it -- what it would or woul'd not contain. _l I
"4-
+- ,- , , , ,,y, . - . . ,- pw, -w. -
.<ew- - e g
. _ -. ~.
1 Roy 267
/~N -
(_) 2 Q I am just asking for your understanding of 3 site problem reports in 1978. 4 Did you think that if there was an ggg 5 instance where B&W equipment was operated in a manner 6 contrary to how B&W intended it to be operated and 7 that had contributed to the site problem, did you think 8 that that should be included in a site problem report? 9 A The originator makes that kind of 10 determination, you know, at the time he is preparing 11 it. The originator is generally the representative 12 at the field. If he thought that was, you know, a [J)
, 13 significant element in quantifying a problem, 14 characterizing the problem, I certainly would have no 15 knowledge of any limitations or information which 16 would cause him to_not, you know, put that in or 17 preclude him from. putting it in.
18 Q How did it come that you were an attendee 19 at an NRC research review group meet.ing early in 19787 20 A You will have to refresh my memory some. 21 Q Saul Levine presided, Steve Hanauez
@ 22 attended and spoke. Levine was director of..NRR at 23 the time.
- - 24 A I don't recall that. specific meeting. It
\/
25 was fairly traditional for B&W to conduct what c-<
1 Roy 268 2 we called review and research programs, I think, then l 3 , n an annual basis, as best I can recollect, regarding 4 research programs that we had under way at B&W. I g 5 don't recall -- 6 Q Y u w uld review them for the NRC? A Yes. 7 I don't recall that specific meeting. g Q Let me show you a memo you wrote to Jim g Taylor,
Subject:
NRC Research Review Group Meeting, 10 which I will mark GPU Exhibit 457. gg (Memo to J. H. Taylor from D. H. Roy, 12 dated January 20, 1978, with attachment, was 13 marked GPU Exhibit 457 for identification, as of 34 this date.) 15 Q You look as though you have been hit by l 16 a thunder clap. g7 A No. This was a meeting entirely different ! 1 18 from the one I referred to earlier as those routine _gg reviews of NPGD research programs. This is a meeting 20 I was invited to participate in to review the NRC's 21 plan f r resear h projects related to the development 22 f new and improved systems for nuclear power plants, 23 and.I remember being a member of that, and I think 24 probably a task force would be a good name, ad hoc 25 **Vi*" 9# "P*
4 1 Roy 269
.2 Q At the time of this meeting, did you think 3l that B&W nuclear plants were already safe enough and 4 didn't need any more NRC suggestions for further gg) 5 safeguards?
6 Q I believe the B&W nuclear plants are safe, 7 at this time and at the time of this meeting, yes. 8 Q And you didn't think there was any need 9 for further NRC recommended safeguards; is that right? 10 MR. FISKE: I don't quite understand the 11 question. Do'you mean that he wouldn't be 12 willing to listen to proposals or evaluate ex (_) 13 them, or do you mean without hearing any proposal, i 14 he didn't think there was any need for any l 15 further proposal? l 16 Q Do you want to respond to the question? 17 A This meeting, I was asked to join a group 18 which the NRC had. convened to advise and counsel them 19 with respect to their long-term research plans that 20 could lead to improved safety systems as well as 21 analysis approaches. One of the items, for example, 9 22 we considered was the establishment of a safety goal, 23 using a probabilistic approach was one of the items 24 . we discussed, so there were broad ranging issues as 25 the NRC look'ed ahead to the kind of research they ought
1 Roy 270 2 to be conducting in fulfilling their requirements for 3 the Atomic Energy Act, and I served on that ad hoc 4 review group. Probabilistic analysis picks up on an l 5 Q 6 idea that Norm Rasmussen has proposed in WASH-1400; 7 right? 8 MR. FISKE: Wait a second. Can I hear that 9 again? 10 Q You referred in your last answer to 11 probabilistic analysis. 12 A Yes. () 13 Q You said that was something you discussed 14 at this NRC meeting. 15 A Yes. 16 Q Probabilistic. analysis means that you 17 weigh the probability of something happening in a 18 nuclear plant; right? 19 A Yes, it is a broad issue. As a matter 20 of fact, it is a broad discipline today. WASH-1400
.21 uses -- used some of the methodology and techniques 22 that fall under the umbrella heading of probabilistic 23 , assessment.
I 1
/~ 24 Q Fault tree or event tree analysis places li
(_}/ 25 heavy reliance'on probabilistic analysis, doesn't it? a
- e. , , , , , --
- , . , - -, , ,,e -, + e -.
~
l ( 1 Roy 271 l 0 (_j 2 A That is one of the elements of this 3 discipline, yes, techniques, methodology of that 4 discipline, g 5 Q Is it fair to say that B&W has gone much 6 farther toward embracing probabilistic analysis since 7 the Three Mile Island accident than you had gone 8 before the Three Mile Island accident? 9 A No. We have used event trees and fault trees 10 in terms of analyses associated with system reliability 11 in looking at various plant scenarios, but in terms 12 of the context you are talking about and playing a role () 13 in licensing and changing of NRC criteria with respect 14 to FSAR or PSAR acceptance criteria, it still is not 15 playing any bigger role afterwards, in my opinion, 16 than before the accident. 17 Q Put aside licensing because I wasn't 18 talking about licensing at all, I was talking about 19 engineering analyses. 20 A. Yes. 21 Q Attempts to anticipate what could happen 22 during a casualty and whether the system response 23 is going to be adequate to deal with it or whether g 24 operator response is going to be adequate. 1 J 25 Are you using more probabilistic analysis l l i _ _ , - - _- _ - - - J
~1 Roy 272 e
' k_m) 2 now than you were using before the accident?
3 , A I wouldn't say it was a great deal more j- 4 than before the accident. llg 5 Q You sure weren't using very much before 1 6 the accident, were you? i 7 A Well, probabilistic analyses, they are 8 really only effective if there is going to be an 9 acceptable technique for making judgments about 10 acceptability of systems or unacceptability, and that 11 judgment has to include the acceptance of the 12 regulatory authorities. Since that is not embraced (/~)j 13 to any great extent by them, it wouldn't play a very 14 big role until that becomes a more formalized method 15 in the safety reviews conducted by the NRC, in my 16 opinion. 17 Q Why do you need to have the NRC tell you 18 what's a good way or not a good way to analyze , 19 casualties? 20 A well, you don't -- if you are going to be 21 effective at those analyses affecting the design of 9 22 the system and you are going to propose'a system for i 23 which you have justified it or defended it on the basis f-$ 24 of probabilistic risk ~ assessment, O.K., it is not N-) . 25 going to find-its way into the plant or into service i
1 Roy 273
-3
_ ) 2 until that approval body can accept that kind of 3 analysis. You just can't do it, you know, unilaterally. 4 Now, I can use it in the design process lll 5 'to help me make selections from alternatives available 6 to me, but ultimately if I am going to rely upon 7 those methods to qualify my system, then both the 8 methods and the justification I put forward is going 9 to have to be accepted by that approval body, so you-10 can't do it unilaterally and have it as effective as 11 it can be. 12 Q I understand it is a question of degree. ( 13 A Yes. 14 Q Are you saying that because the NRC, 15 according to you, has not accepted probabilistic 16 analysis to a great extent, B&W has therefore not 17 made much use of it? 18 A In terms of proposing, you know, 19 modifications or using it to justify and seek approval 20 for some system or system modification, that is true. 21 Q What about just to design a better plant? 22 A Yes, we can apply it, as I say, 23 unilaterally to help us make decisions -- _ ,em : 24 ,Q Do you? A v
]
25 A -- on it.
i 1 Roy 274
"%- 2 Yes, we have used system reliability 3 technique which is very close to PRA techniques, f
4 probability risk analysis in examining the reliability lll 5 of various system designs. 6 Q Are you doing more of that since the 7 Three Mile Island accident than you did before the 8 Three Mile Island accident? 9 A Some more, but it is not greatly different. 10 It is not highly significant in my opinion. 11 Q WASH-1400, entitled the " Reactor Safety 12 Study," was a report published by, issued by, 13 circulated by the' Nuclear Regulatory Commission; 14 right? 15 A Yes. i 16 Q It came out in 1975. 17 What are you' relying on when you say the 18 NRC has never endorsed the use of probabilistic 19 risk analysis? 20 A I will use your words, in any significant 21 way. It has been used in isolated instances and 22 reviewed by the NRC on that basis. 23 My basic opinion for that is deriven by ('N 24 what I know to be an ongoing dialogue,'which this.
.() .
25 meeting was part of that ongoing dialogue, as to what l
1 Roy 275 (,) 2 role probabilistic risk assessment and the 3 establishment of quantitative safety goals should 4 play in the licensing process, and I know that that ggg 5 is an ongoing dialogue withi-n the NRC and within the 6 industry. 7 Q Did you ever point out to any of the NRC 8 people at this meeting that Professor Rasmussen had 9 put together a major report published by the NRC on 1 10 probabilistic risk analysis? l 11 A I don't recall pointing that out. This 12 body of advisory personnel which was convened by ( 13 the NRC would be quite aware of that, in my opinion. 14 Q Did you get the impression that people 15 thought WASH-1400 was really off base and not very 16 helpful? 17 A No. I don't recall having any impression 18 one way or the other about WASH-1400 in this meeting. 19 Q Well, the meeting sounds, from your 20 description of it, as though it had rejected a lot 21 of the fundamental tenets of WASH-1400. O 22 A No. This meeting dealt with identification 23 j of areas which this advisory board was to counsel the i
~
24 l NRC on in terms of opinion as to what areas the NRC 25 in the long term would find useful to pursue as they l
l, 1 l 1 Roy 276 l
/ 2 laid down their long-term research plan.
3l Q Have you ever had any significant involvement 4 in the BBR design for the MK plant in Germany? ggg 5 A Yes. A substantial amount of the design 6 activities task which we were performing for BBR were 7 -conducted in the plant design section when I was 8 manager and in the engineering department since I 9 was manager of engineering. 10 Q Are you aware of more use of PRA, or j 11 probabilistic risk assessment, being employed for that 12 plant than B&W was applying to its domestic nuclear w' 13 plants? 14 Q Not in any experience I had and tasks I 15 we were dealing with for BBR that I recall. l 16 Q According to page 2 of your notes in GPU 17 Exhibit 457, Steven Hanauer led a discussion on the 18 work "needed to help eliminate operator error prior, 19 during,'and after an upset or accident event." 20 Do you remember that discussion? e 21 A I don't recall the specific discussion. O 22 That is an area of Dr. Hanauer's expertise. I know 23 he is especially expert in that. area, but I don't rw 24 recall his particular discussion of that at the meeting. 25 Q You wrote this memo just a couple of weeks
l l 1 Roy 277 i D (- 2 before you got Dunn's memos about the operator 3 termination of high pressure injection at the Davis-4 Besse plant. ggg 5 When you got Dunn's memos about that 6 operator action, I am sure you thought about the 7 spirited discussion that you had just been in with 8 Steven IIanauer about the need to avoid operator error; 9 isn't that right? 10 A I don't recollect making any connection II or drawing any conclusion like that at that time. 12 g Did you think that Hanauer was correct in
) 13 pointing up the siorificance of studying operator 14 errors related to reactor safety?
15 A I don't recall the conversation, but I 16 would support that as a useful objective. 17 As a result of this meeting or at or about Q 18 this time, do you know whether B&W had under way any 19 projects to analyze opportunities.for minimizing 20 operator error? 21 A I don't know about the time span, but one 9 22 of the consider 5tions in an overall availability 23 improvement program, I had mentioned previously the I 24 (~s identification that we had of various contributing areas
'V 25 to lost capacity days, the equipment failures or what l
_~ l l r . , - . - - - - , _ _ . - .,
, - - . . . _ _ _ _ ~
1 Roy 278 p) t (, 2 have you, refueling outage, and the broad term human 3 error is one of those classifications for contribution 4 to lost capacity days. I don't recall whether it was 5 in this time frame or not, but in.looking at that 6 broad classification, which includes everything from 7- operator error or maintenance error or what equipment t 8 modifications could help reduce that contribution, 9 we had given consideration in a-broad strategic plan 10 for availability issues to try to find out what 11 specific items were occurring that make up that broad 12~ category. I 13 We had very little information on that. So
\_)T 14 we addressed it strategically and began to concern
~15 ourselves that at some point in time when we are 16 putting together ideas for availab'ility improvement 17 that that would be an area that we would like to learn 18 more about. Again you can't do that unilaterally.
19 you had to work with a customer, and that is a 20 particularly difficult one; you've got to have the 21 customer on board to do that. 22 -Q Was there any group or individual at B&W 23 fwhowas given particular. responsibility for reviewing gs 24 the possibilities for reducing operator error?
- ()- 25 A My recollection is that -- well, a
4 1 Roy. 279 + ( 2 gentleman named G. M. Olds was the -- you might say 3 the point contact for the strategic availability plan. 4 Whether there was a specific individual assigned 5 to help, you know, review, look into that particular i- 6 item, I don't recall, but that would have been in the
.7 service department.
8 Q Olds' area included improving availability .Ji
- 9 of valves, pipes, any piece of hardware also; right?
10 A well, in this particular assignment that.
- 11 I know Mr. Kosiba gave to Andy was to be a coordinator 12 for many elements of the availability improvement
! 13 program. So he would be overlooking whatever 14 activities or initiatives were under way to try to i 15 address all of the potential areas for improvement. 16 Q Right. . 17 A The principal responsibility for an
~
18 equipment type of issues like pump seals -or this kind 19_ of thing would be primarily in engineering, but again 20 it is not a unilateral thing. We_have to work 21 through customer service to acquire information and i~ 22 to put together statistics; for example, the statistics 23 -were prepared in the customer service area. 24 Q Other than' developing a matrix,-you' don't 25 know what, if any,~ work was done to. review opportunities
1 Roy 280 2 for minimizing operator error, do you? 3 A well, I don't have any recollection of a 4 matrix associated with human error; O.K.? Just that g 5 was a broad category contributing to lost capacity 6 days in which you can put a big question as to what 4 7 are the specific elements of that. You have to know 8 that bef' ore you can conceptualize ways to help your 9 customer get that out and help drive it down, and I 10 don't know of any specific initiatives. I know that 11 most, if not all, of our customers were approached 12 with this overall strategic plan of addressing 4 /~N ? 1 ,) 13 availability improvement, suggestion to set goals. 14 Aside from that, I wouldn't know of any --
- 15 I am not aware of or have any recoll'ection of any 16 specific initiatives along the human error line.
4 I 17 Q In the discu'ssion'that you had at the NRC 1 18 did you. consider to what extent, if any, operator 19 error was included in transient analyses? 20 A I don't recall that being an element of 21 the discussion up there. It was a very broad topic O 22 we were dealing with up there. 4 23 - Prior to the Three Mile Island accident Q I 24 i do you. recall any discussion or analysis at B&W of 25 to what' extent operator error should be taken into
<s.
9 ..r , y ,. , , . - -4 , . - , , . , , . - .
1 Roy 281 i ( ,). 2 account in transient analyses? 3 A No, I don't have any recollection of that 4 coming up as a specific item in the performance of g 5 the safety analyses that are.-- formed the basis for 6 licensing, you know, of nuclear power plants. 7 Those criteria are laid down in the 8 regulations prepared by the NRC in a very conservative 9 way. They bound a provision for a lot of uncertainties 10 or unknowns. That is why they postulate, you know, 11 combinations of failures and require that it be done 12 in a conservative or bounding way. So with respect ( ; 13 to that coming up as a specific topic, I don't have
%)
14 any recollection of it doing so. 15 Q Do you know Zoltan Rosztoczy? 16 A Yes. 17 Q Did you ever discuss with him at any time 18 the problem of taking operator error into account in 19 doing accident analyses? 20 A Prior to Three Mile Island? 21 Q Prior or post. O 22 A Not prior to Three Mile Island but -- no, 23 I don't recall a specific conversation myself, but i 24 with respect to the blue volumes, small break
'~
25 guidelines, if my recollection is correct, there some
1 Roy 282 k_ 2 of the scenarios would, as I recall, embrace some l 3 consideration of operator action or misaction. I 4 think that was an element of consideration, and I ggg 5 don't know to what extent, in the preparation of the 6 blue volumes. 7 Q Was there any consideration of it before 8 the Three Mile Island accident? 9 MR. FISKE: Any consideration of what? 10 MR. SELTZER: Operator error. 11 MR. FISKE: In what? 12 MR. SELTZER: In drafting instructions (uj I 13 for operation of the plants. 14 A I wasn't involved in drafting instructions 15 for operation of the plant, so I don't know about that. , 16 Q You referred earlier to having two high 17 pressure injection trains, and in all of the safety 18 analyses that were done you assumed that one train is 19 not going to work and that you will only have one 20 train working. - 21 Did you ever in your safety analyses make O 22 the further assumption that there might be operator 23 error that would incapacitate the second HPI train? l
,- 24 MR. FISKE: By --
25 Q Prior to the accident, I am talking about.
1 Roy 283 2 MR. FISKE: By safety analyses, Mr. 3 Seltzer, I assume you are using that as a term 4 of art. 5 MR. SELTZER: Yes. Let's take it first g 6 as a term of art. l 7 Q Analysis done to comply with the ECCS 4 8 cooling criteria of 10 CFR 50.46 and Appendix K. 9 A No, that wasn't a requirement of the 10 regulations to totally fail the -- you know, the safety 11 system. The single failure was considered to be a 12 conservative assumption for the purpose of doing that 13 analysis. 14 Q Now, taking it a different way, for any 15 other purpose other than for analysis to comply with 16 10 CFR 50.46 and Appendix K. did B&W ever analyze the 17 impact of possible operator error, shutting off the 18 second high pressure injection train? 19 MR. FISKE: You are talking about Dr. Roy 20 as head of the plant design and then the 21 engineering department? O 22 MR.. SELTZER: I am not excluding his 23 hearing about it'in any other way, but certainly _ . 24 that's who he was prior to the Three Mile Island l 25 accident.
.,.v -.
. -.u,. . . ,
1 Roy 284 i im I) 2 MR. FISKE: Could I hear the question 3 again, please. i- 4 (Record read.) i 5 MR. FISKE: Can you answer that? 6 A Not to my reccliection, no. 7 Q Was any analysis done before the accident 8 to consider the impact of operator error on transients? 9 MR. FISKE: Do you mean in any analysis 10 of any transient, was any consideration ever 11 given to operator error? { 12 MR. SELTZER: That is where we are starting. () 13 Does Dr. Roy know of B&W appraising the 14 significance of operating error before the Three 15 Mile Island accident. 16 A Well, the scope of these analyses that I 17 was responsible for and am responsible for is for the 18 qualification of the safety systems, that they have 19 adequate margin to mitigate a transient, to the L 20 guidelines of the regulatory requirements. t 21 So no, I don't recall any specific 9 22 analysis of.th.: :im p a c t of operator action or misaction 23 as part c: the analysis scope that we worked on, l gy 24 either in the benchmark case or the performance of l IO 25 analyses to license a safety system. t k
, . _ . _ _ a
f 1 Roy 285
/Y 2 Those analyses have a number of V
3 e nservatisms, i~nitial conditions, failures are
, 4 conservative and provide a wide margin for that the
- e 5 6
*Xpected condition would be generated in a plant versus that the safety system could be capable of
~
7 handling. 8 Q All right. You would study almost every 9 conceivable type of pipe break, valve failure, pump 10 failure, power failures right?
~11 A No, not every conceivable. You --
12 Q Almost every conceivable way? ' 13 A ~ You do sufficient sensitivity studies
~
14 within the scope of the-regulations to provide 15 reasonable assurance that you have bound the worst 16 case, and that is done in a conservative way; O.K. 17 Q Is it fair to say that you did analyses
~
3 18 to anticipate the consequences of any significant 19 pump failure, valve failure, pipe failure or power. . 20- failure, whether you bounded it and considered it,- 21 anticipated in that way or you studied it directly? 22' A Yes,.the -- let me state it this way. 23 The basic approach to safety. analyses is 24 such that adequate conservatisms are assumed so that' A s- 125 the scenarios that you would'put together in a l l I - 4 s
-- > - - - ,e .~ , , - - - --,n ,-.n.-,--.,y.--. - , - - - , ,
e l l 1 Roy 286
\
(_)' 2 sensitivity study which have major degrees of I 3 conservatism and then which you further identify the 4 worst case within that bound, O.K., would envelop o.K., gg 5 a wide. variety of much.more specific kinds of 6 scenarios that you might come up with. 7 Q So by studying the worst case, you feel 8 you have anticipated and covered failures and problems 9 that are-less severe than the worst case; right? , 10 A And provided a substantial margin to cover 11 all the scenarios you are not going to be postulating 12 and coming up with. [_) 13 MR. FISKE: Are you going to ask him whether 14 they analyzed the-hypothetical worst case of 15 operator error? 16 MR. SELTZER: Let's not try and make a 17 joke out of something very serious. Let me 18 proceed. .
~
19 MR._FISKE: .Sure. 20' Q . 1 You are saying that by studying the worst 21 case..and envnloping cases that are less severe, 9 22 you anticipated the consequences of any significant 23 ! valve failure, pipe failure, pump failure or electrical 24 .f a'ilure s. . i s t[ hat right? s N)' A It'ir not-specific. 25 What the. intent i's - tl ,. 1 j .,
- t ; '
\ '
;)3 .,.y _ a f . 3$ = - - -, - ~ '- ' ~"' ' ~ ~ ' '
1 Roy 287
) 2 is that adequate margins in the safety systems have 3 been provided to cover, with reasonable' assarance ,
4 those other kinds of scenarios, and there are an 5 infinite number you can postulate. You can't cover ggg 6 all the infinite variety, but reasonable assuranco 7 that those kinds of things which would decur~B o.K., , 8 will be accommodated by the, safety system. 9 Q Is the answer to my question yes? 10 MR. FISKE: I think he answered'the ques' tion. 11 MR. SELTZER: I think he walked around t 12 it many times and maybe the answer sits in t h c-rN 1 (m ) 13 middle. I think he indicated that\the answer 14 l is yes, I would just like to unders .'and whe the'r 15 it is yes. If I understand that, the answer is 16 yes; if I misapprehended -- 17 MR. FISKE: Read the question and the 18 enswer back, and then he will answer your 19 ,; question. s t 20 f (Record read.) 21 MR. FISKE: Mr. Seltzer, if youswant to I 22 ask him the question again and he can answer 23 it yes or no, he will answer it that way. , j 7~ 24 I spent six days with one of the Three 25 Mile Island operators and in the entire six days j I i v 1
th 4 . 3 1 Roy 288 k f- . T * ' 2 he never once answered a question yes or no,
" 3 although he was repeatedly asked to do so. I q
ar. ,\ l
'4 was repeatedly told t h'a t he had answered the g
5 question and! that he didn't have to answer it
) ".
6 .' yes or no if he couldn't, and indeed he even i 7 \ told me on one occasion he couldn't answer a t.
~8 ques, tion yes or.no before I even put the question g %
9 to him. So I; kill be perfectly happy to let
.10 - Dr. Roy be the judge. If he can answer your 11 question yes or no, I would be happy to have 12 him do it, but he is under no obligation to do
() , 13 it i it is not susceptible to that kind of an
~ 14 ,; an s, wor )
-Q 15 MR. SELTZER: I will stipulate to that, i
16 that is a very good statement. If a witness says 17 he c a r. ' t answer a question yes or no, it is
't 18 certain'ly not fair for you', Mr. Fiske, or for
~:
19 me to\ \insist that a witness make an answer that 20 he says he can't make. 21 MR. FISKE2 I agree with that. The witness O '22 has to look himself in the eye and decide
.._. 23: whether he can do it.
v i 7 .;pg 3
' 24 ' A The way that question is posed, I - .( b J ' ',.
1 7 25 respectfully submit I can't answer yes or no, ' 3 I
'Z, (4
l j
1 Roy 289 4.
'(~%
c (_) 2 Q O.K., that is an answer. 3 A It is too broad. 4 Q O.K. ggg 5 Using the bounding analyses that you had 6 done, did you believe that B&W had determined that 7 there was reasonable assurance that the plant could 8 respond safely to any credible pipe failure, valve 9 failure, pump failure or electrical failure? 1 10 A What do you mean by " credible"? 11 Q I mean what 10 CFR 50.46 means.by credible. J 12 A Within the stipulations of 10 CFR 50.46
} 13 and Appendix K, yes, we perform analyses in compliance 14 with those regulations and have safety systems and 15 core designs that will comply with those regulations.
4 16 Q So you do believe that you had done 17 sufficient analysis to give reasonable assurance that 18 the plant would be able to respond safely to all 19 credible pipe breaks, valve failures, pump failures 20 or electrical failures; is that right? Can you 21 answer that yes or no? 9 22 A Within the prescriptions of Appendix K, 1 23 again Appendix K and 10 CFR.50.46 are the same kind i I I es 24 of analysis.
\.j ' .. >
.25- Q I understand. 1
.J l
L _
1 Roy 290 A d ,) 2 A A bounding kind of analysis, sensitivity 3 studies, or the likely locations of it, you know, the 4 worst case break or size, you don't analyze every gg) 5 single place, O.K. To provide an ECC3 system that will 6 envelope a wide variety of breaks and locations 7 and core conditions; that is the intent of that 8 regulation, to do that. 9 Q Now we have on the table what the intent 10 of 10 CFR 50.46'and Appendix K is. What is the l 11 answer to my question? Do you believe that based on 12 the analyses you did, B&W had assured itself that the () 13 plant could respond safely to the full range of 14 credible accidents that could incapacitate pumps and 15 valves, break pipes or shut off power? 16 A You have used the words full range of e 17 breaks and pump conditions and so forth, i 18 Q Let me qualify that. I mean the full 19 sp.ectrum'within the confines expected under Appendix K. 20 A Yes. They are ongoing analyses, you know, 21 0.K. So at any give time, yes.
;O 22 Q Now, did you recognize when Steve Hanauer 23 -was leading the discussion in January 1978 that i
gw 24, WASH-1400 had set forth as one of its most significant v 1
\,_/
25 findings that greater attention needed to be paid to.
+ , . ~ , -
n , ,,. . , . - -
}
l 1 Roy 291 l O} 4 2 operator error as related to reactor safety? 3 A In the context of that meeting, I don't 4 recall Mr. Hanauer's -- Dr. Hanauer's comments or ggg 5 discussion of WASH-1400. . 6' Q All right. Had you read WASH-1400 or 7 any portion of it? 8 A I have seen it and read portions of it. 9 Q Had you read portions before the Three 10 Mile Island accident? 11 A Yes. 12 Q Did you know before the Three Mile Island 13 accident that one of the points made in the Rasmussen 14 study, WASH-1400, was that operator error should 15 receive greater attention? 16 A It's been so long that I have read that 17 that I don't have a recollection of what conclusions 18 I made when I was reviewing the document. 19 I don't recall deriving that conclusion 20 from my reading of it at that time. h' 21 Q Did you ever suggest to anyone that B&W / 22 should include within its transient analyses some
~23 consideration of the probability for operator error?
- 24 l A I don't recall making any such suggestion s
25 within the context-of my knowledge of the kinds of.
1 Roy 292 e
.is .
- s. 2- conservatisms that are built into the analyses 3 , to qualify safety systems, with that kind of knowledge 4 of the margins that are there, that wouldn't have --
lg) 5 ~ wouldn't and didn't occur to me. I don't recollect 6 raising that issue or making that suggestion. 7 Q A lot of B&W's transient analyscs were 8 done by people who reported to you, both when you were 9 the head of the design section and the head of 10 engineering; right? 11 A Yes. Prior to Three Mile Island? 12 Q- Yes.
.( 13 A My answer to your previous question was 14 with respect to prior to Three Mile Island.
15 Q And the answer is.yes?
-16 A The answer is I don't recall making that f
j 17 suggestion prior to Thre'e Mile Island, or that coming 18 up -- 19 Q All right. 20 A -- as an issue that should be factored 21 into the safety, analyses associated.with, you know, 22 licensing and qualifying the safety systams. 23 Safety analysis -- by the way, when we t
(
.q, '
24 talked about the safety analysis unit, that is their j 25 mission. They. don't deal with just all safety issues, l
1 Roy 293 O)' (_ 2 they are dealing with safety analyses within the 3 regulations directed towards licensing the safety 4 systems and supplying appropriate setpoints. i-j 5 Q Dunn's group is involved in that also; 6 right? 7 A Yes, and they are on the ECCS side. 8 Q That is safety analyses dealing with
- - emergency core cooling systems; is that right?
9 10 A Yes, sir. 11 Q So you had two units that were doing 12 transient analyses reporting to you; right? 13 MR. FISKE: I have no objection to that [%s} 14 question -- 15 MR. SELTZER: Lgt's just get the answer. 16 MR. FISKE: Yes. 17 A Yes. 18 MR. FISKE: I think it would be helpful, 19 and I don't think it made any difference to the 20 last answer, but when you use ' transient a n aly s e s ," 21 whether you are using it in the context of 9 22 qualifying it for the licensing authorities 23 or some broader basis. 24 Q- Let's forget doing things for licensing 25 authorities because I am not limiting my inquiry to
= _ _
\
1 Roy~ 294
; .. j 2 that. I want'to know if you and your people were 3 doing analyses for any purpose that considered the 4-
'4 impact of operator error; that is the area of my h 5 inquiry. -
- 6- MR. FISKE
- One reason I think this would 4
i 7 be useful is because I think there has been a 8 lot of testimony here about safety analyses, t
/
9 and I am a little concerned when that gets typed i.
- - 10 ' up --
as you indicated once, it is important i 11 whether that is capitalized or not capitalized.
~
12 MR. SELTZER: Right. ( 13 MR. FISKE: But the point has been made ~ 14 by Dr. Roy that the safety analysis group was 15 a group whose- responsibility was limited to the i 16 type of transient analysis which had'to do with
- 17 qualifying for the l'i c e n s e , .and as long as,that I
18 is understood, then I think there is not as big 19 a problem, but these terms sort of flop around i 20 and I think there is a.resk of misconstruction
. 21 later unless.we all understand it.
22 MR. SELTZER: I-don't want anybody flopping; 23 here. 24 . MR. FISKE: Right.. 25 Q Mr. F'iske just made a statement that1he I
- ~ - .
m -g 'Tv-9 mww=.W*-ie- eW+q ums +m$ g yyp&-ete* ,tgy-v-t?--4 W MTm+-yr+pp +., wp p ge>**yy y eeFVW-N'f -* 7FM TTY-'Y#" $
1 Roy 295 1' b N,j 2 thought'the safety analysis group only did analyses 3 to support licensing prior to the Three Mile Island 4 accident. lll 5 Is it correct that they, to your knowledge, 6 never analyzed anything except as it related to 7 qualifying plants for licensing? 8 A Let me add one other class of analyses l 9 which in a way are supportive of, related to the 10 primary mission of doing the licensing safety analyses. 11 I don't recall any specific instance, but 12 in the process of benchmarking the computer codes --
-[% )\ 13 we talked about that yesterday -- if a tran'sient at 14 an operating plant were selected to be used as a 15 benchmark or to be reviewed or evaluated for use as 16 a benchmark, if there was some action that materially i 17 affected the course of the transient, action on the i
18 part of'the operator or some other equipment 19 malfunction, they would have to be modeled, you know, 20 t6 get a ~ fair benchmark. 21 Q Yes. 22 A Is that clear?
- 23 Q Yes.
g- 24 When Vosburgh's memo came to your attention ( ' 25 and he suggested that safety analysis ought to be
1 Roy 296 O
-(_) 2 involved in more operational problems, did you 3l understand that his concern was only to qualify 4 equipment for licensing?
lll 5 A I. don't recall coming to that conclusion. 6 Q He had a broader purpose, didn't he? 7 He wanted to make sure the plants were safe; isn't 1 8 that right? 9 A We have only his memo. 10 Q Where is his memo? [ t 11 (Document handed to witness.) 12 Q From his memo, isn't he concerned with () - 13 safety in a more general sense than just qualifying. L 14 equipment for licensing? l 15 A Well, I would derive certainly-one thing 16 'from this that Mr. Vosburgh had in mind was that he 17 desired more timely involvement, and one of the 18 purposes.he calls out here is that he can compare the 19 kinds of transients that are occurring in the plant 20 . with the assumptions that he has made in his safety 21 analyses. That is one that I read yesterday. 22 Q Take a look at item 3. Doesn't he say 23 he wants to assess whether a site problem constitutes
. 24 a safety issue and/or a violation of the safety analysis L-25 assumptions under which'the plant is licensed?
4 . . , , . . . . . _____.______m_____ _ _ _ - - _
1 Roy 297 p)x ( 2 A Yes. 3 Q Doesn't that indicate to you he was 4 concerned not just_with compliance with the licensing r
$ 5 criteria?
6 A Yes. 7 Q Didn't it seem anomalous to you before the 8 Three Mile Island accident that your engineers were 9 doing such an extensive job analysing equipment ! I-10 failures but were not incorporating at all into their i 11 explicit analyses the possibility of operator error? 12 A With the very conservative a s s ump tio n s ( 13 that are made in assessing the adequacy of a safety 14 system, that did not appear to me as an anomaly then 15 or today. - 16 Q The conservative assumptions were 17 assumptions about equipment failure; isn't that right? 18 A Not limited to that, no. All initial 19 conditions on the conservative side for a given 20 particular transient: available flows, core linear 21 heat rates, inlet temperature, outlet temperatures. 22 All of those many parameters that constitute the 23 input to analysis would be selected with margins to l
~ 24 be conservative, to give a conservative representation
,O 25 of the response of that plant, not just limited to
1 Roy 298 O
\s-)- 2 equipment failures.
3 Q Do you know whether any of those 4 conservative margins were selected based on an {lg 5 explicit appraisal of the types and probabilities 6 of human response? 7 A No, I don't recollect that that is a 8 specific consideration. For example, if you in the i 9 safety analysis, if you assume you fail a high 10 pressure injection train, from the analysis standpoint 11 it doesn't care how you got there. It might have lost 12 power or the pump didn't come on or whatever way, but O () 13 only in that context, but not explicit consideration 14 how it got there. 15 Q All right. So there was no explicit 16 consideration that you know of of the type or 17 probability of human action? 18 A In the performance of these analyses, 19 for qualifying the safety systems, no. 20 Q What about for any other analysis done 21 by your people studying accidents? O 22 A I have mentioned the only other one that. 23 occurs to me right now, and that is if I were r~} 24 benchmarking a transient in which operator action
.v/
, 25 played a role, then I've got to model that, whatever r i
' 1 1 Roy 299 l l l l-2 that role wasr O.K.? 3 Q What does the phrase " common mode failure" 4 mean to you? jgg 5 A A common mode failure would be where a 6 single element in a system could fail and have 7 consequences in failing other systems, one or more 8 other systems. In other words, the failure will not 9 just affect the system which contained the element, 10 but could lead to a sequence of. failures perhaps of 11- other systems. 12 Q Since the Three Mile Island accident,
) 13 has there been analysis done by your engineers of 14 the possibility of common mode failure causing all l 15 redundant trains of a redundant safety system to be i 16 incapacitated simultaneously?
17 MR.-FISKE: Could I just take two minutes 18 to get some air? 19 MR. SELTZER: Yes, sure, i I
- 20 (Recess taken.)
21 MR. SELTZER: Will you please repeat the 22 pending. question. , 23 . (Question' read.) f~g- 24 A I am not aware of any analyses that would 25 postulate common mode failure, no.
1 Rey 300
-m k_) 2 . Q So your group doesn't consider common 3 mode failure; is that right?
4 A Not explicitly from the standpoint that 5 equipment single failure is assumed to occur for llh 6 purposes of doing analyses. That that may be produced 7~ by some common mode failure is a possibility, but
. 8 that is not an explicit consideration. You don't look 9 ~ at the various modes by which, say, an HPI train 10 could be incapacitated. It could be by some common 11 mode failure, but that is not an explicit consideration 12 in doing that analysis. That analysis bounds all the
( 13 possible causes that enuld cause that to fail. 14 Q So to your knowledge, since the Three Mile 15 Island accident B&W has not explicitly set out to study 16 any common mode failures; is that right? 17 A One consideration that involves a 18 postulated common mode failure which has been under 19 analysis since 1969, that is ATWS. Anticipated transients without scram is a. common mode failure 20 21 postulated by the staff has been the basis of analysis
-22 for quite some time. That is a very explicit one, i
23 ! I can't think of any others. I y3 24 Q To your knowledge, did the WASH-1400 N-25 report cause B&W to undertake any types of analyses
l 1 Roy 301 l [(,)\ 2- of nuclear plant events that B&W was not already doing? 3 And I am focusing on the period just before the Three 4 Mile Island calamity. l 5 A None that I can recall. That analysis 6 was not of a B&W plant, as you know, and I don't 7 recall the initiating of any similar kinds of analyses 8 for B&W plants. 9 -Q What plant was the analysis of? I didn't 10 realize it was a plant-specific analysis. 11 A Yes, it was modeled, and I don't recall the 12 plant name. There may be some composite features [% )) 13 but there were certain specific plants, to my best 14 recollection, that model -- for example, as I recall, 15 it modeled a plant with a recirculating steam generator. 16 There may have been some composite features, but it 17 was modeled on a specific plant type. 18 Q It is a fact, isn't it, that many of the 19 analyses in WASH-1400 were applicable to the B&W PWR 20 design: isn't that-right? 21 , A No. Many of those assumptions -- I can't 22 give you all the details, but many of those assumptions 23 very specifically depend upon the equipment and g 24 failure rates which were assumed in WASH-1410 were
%)
25' specifically oriented to.the plant they modeled. You
-- . I
1
, 1
. l 1 Roy 302 (O,f 2 cannot directly apply any of those results to a plant 3 with any significant deviation from the equipment 4 type or failure probability.
lll 5 Q Are you saying that there were not 6 significant areas of generic applicability from 7 WASH-1400 to the B&W product? 8 MR. FISKE: Are you asking Dr. Roy what 9 he understood or what he recalls now of what he 10 un'derstood at the time? 11 MR. SELTZER: Right. 12 Q When you first got WASH-1400 and were () 13 reviewing it prior to the Three Mile Island accident, 14 did you recognize that there were areas of generic 15 applicability of that report to the B&W product? 16 A I didn't at that time draw any conclusions 17 about that. 18 Q Since the Three Mile Island accident, 19 have you recognized that there are areas of generic 20 applicability of the WASH-1400 study to the Babcock 21 & Wilcox product? O 22 A No, I have had no occasion to revisit 23 -WASH-1400 and make any comparisons of that sort s 24 of thing. I was - speaking in the context today. N, 25 Q When you got back from the NRC meeting that
1 Roy 303 O, (, 2 is reported in your notes in GPU Exhibit 457, did you 3 sugges't to anybody that B&W should reexamine the 4 impact'of operator _ action on reactor safety? lll 5 A I don't recall doing that, on the basis 6 of this meeting or any other that I can recall right 7 now. 8 Q You said that your face had flushed and 9 your heartbeat accelerated rapidly when you stood up 10 to speak about safety systems and criteria being used 11 'and considered in Germany. That is page 3. 12 A I don't recall the meeting, but I see the O V 13 words. 14 Q At the time you wrote this, you knew thet 15 nuclear plants being constructed in Germany required c 16 higher degree of safeguarding and installation of 17 safety systems than U.S. plants were required to haver 18 right? 19 A Regulations in Germany are quite different, 20 both in content and in philosophy, in Germany. They 21 are quite different. They lead to system designs 22 that are quite different. 23 Q Increased safeguards, right, among other m 24 ' things? 25 MR. FISKE: You mean.in terms of equipment? Y
-- - -- , , , - _ . ~ w.,. - - % ... *-
1 Roy 304 ('h
-s m) 2 MR. SELTZER: Yes.
3 A In many safety systems they will require, 4 as a consequence of failure assumptions, greater lll 5 redundancy in some of those systems. 6 Q Why was your heart beating so fast and 7 .your face getting flushed when you rose to respond 8 to this German question? 9 A I don't recollect the meeting so I don't -- 10 can't recall my basis for this statement. 11 Q Did you feel at the time that the German 12 approach to safety was excessive? ID
%) 13 A Yes, to the extent that some of the I
( 14 analyses and assumptions that we had to make,
- 15 particularly with respect to changing, rapidly 16 changing regulations back in that time period, that I 17 thought that relative to the kinds of margins and 18 the regulatory approach here that they were excessive.
19 Q You talked earlier abour running a 20 comparison on your computer codes of what the Three 21 Mile Island accident would have been like if there had Le 22- not been an eight-minute gap in.the operation of 23 auxiliary feedwater. 24 Do you recall the fact that B&W did some- [h U 25 review of what would have happened if the Three Mile
1 Roy 305
' /}j-I 2 Island sequence of events were to have occurred in 3 the plant you were designing for Germany under your J
4 joint venture with BBR? lll 5 A First of all, we do various design tasks 6 for BBR; they are the designer of the plant. Just 7' a qualification point. 8 Second, I don't recall any analyses along 9 that line. 10 Q Don't you recall that there was a report 11 that came out in May of 1979 that compared Three Mile 12 Island Unit 2 with the MK plant?
. !T 13 A I don't recall such a report.
Q) \ 14 Q Let me ask you this -- maybe it will 15 come back to you as we talk more about it. There is a I 10 particular interlock feature in the MK plant that would I 17 have arrested the Three Mile Island 2 transient 18 very, very early; isn't that right? 19 A Can you give me some additional guidance 20 here? 21 Q Aren't you aware that substantially 22 before the Three Mile Island accident B&W engineers 23 in Lynchburg had designed an automatic interlock to 24 close the block valve on the MK plant if the pilot j
. f'\ \
~ '
<- 25 Operated relief valve failed in an open position i
1 Roy 306 2 and pressure in the reactor coolant system fell below 3 the PORV setpoint? 4 A Do I have a recollection or knowledge of that prior to the accident, or after the accident? l 5 6 Q Do you know today that such an interlock 7 had been designed by B&W in Lynchburgh and that it had 8 been designed prior to the Three Mile Island accident? 9 A I know that that feature is there. I am 10 not aware of our involvement in doing that but I i 11 know the feature is on the plant. 12 Q Do you know who in Lynchburg designed it? 13 A No, I don't specifically know who designed 14 it.
. 15 Q What unit designed it?
16 A I don't know what unit designed it. 17 Q Who would you talk to if you wanted to 18 find out what unit designed it? 19 A I would talk to Mr. Carlton first. Mr. 20 Carlton was the manager of international projects 21 engineerirg for a while. Dr. Womack. That would be O 22 a starting point to determine what role we had in that 23 , design. I e- 24 Q Are you aware today that under the design wJ 25 for the MK plant,-if a TMI-2 type accident were to l
._. I
i l 1 Roy 307 r.
.(-
( 2 occur at that plant, the block valve would have l 3 been closed within two minutes automatically? 4 A With that feature -- it. works, it is lll 5 going to close the path automatically versus, say, a 6 remote manual case, but if it's got an automatic
; 7 interlock that works, it will close the path.
8 Q Have you ever examined why that automatic 9 interlock was not offered to. Met Ed for Three Mile 10 Island Unit 27 11 A I have no recollection of it coming up as 12 a possible offering.to the customer, owner / operators ( 13 of our plants with our NSS. 14 Q So you don't have.any information as to 15 why it was not offered to Met Ed; in that right? 16 A Yes, or that it came up at all for any 17 consideration on it. 18 Q Do you know that one of the things which i i
. 19 Toledo Edison indicated it wanted after the Davis-Besse 20 September 24, 1977 transient was an automatic interlock 21 to close the block valve in the event of a future i
22 failure of the pilot operated relief valve?
- 23 l A I am not aware of that. I have no I
24 ' recollection of that. f' 25 Q You never heard that Joe Kelly came home t I
I Roy 308 () 2 fr m visiting the Davis-Besse plant after that 3 transient and his notes indicated Toledo Edison wanted 4 such an automatic interlock? A I.have no recollection of being aware of lll 5 6 that at all. 7 Q Did you have any mechanism within the g plant design section for customer requests such as , 9 that being processed in any fashion? 10 A If a customer desired that piece of it equipment or that modification, he can request a bid 12 or have some form from the service manager. It would c me through the service manager, and then if the ( ) 13 14 customer awarded the task to B&W to do that, then the 15 service manager would authorize whatever tasks were 16 appropriate for engineering. 17 Q Was there any route at_that time by which 18 Joe Kellys having written this request down in his 19 notes, could have raised it? 20 MR. FISKE: I will object to the form of 21 the question. O 22 Q Y u may answer. ; 23 A Nothing would preclude hin from -- if it l
'l
~
24 ccurred to him, from suggesting that as a possible
<('~' l
- 25 service that could be rendered to that customer, which i
1 Roy 309 , (3 3,,/1 2 might or could be followed up by a service manager. 3 It would be one appropriate path that I could think of. 4 Q Do you know that the BBR design for the MK plant explicitly included common mode failure in
' lll 5 6 its accident analyses? .
7 A Now you are reaching a little bit beyond 8 my detailed knowledge of those regulations. 9 MR. FISKE: Do you mean --
- 10. A No, I don't remember whether it did or 11 didn't. I am not that knowledgeable right n_o w on 12 those German regulations.
D 13 Did the engineering department, before the 4 Q w . 14 Three Mile Island accident, ever review the safeguards 15 that had been designed into the German plant and make 16 a determination whether any of those safeguards should 17 be brought over to be offered to United States 18 customers? 19 A I don't recall any explicit interchange of 20 that type on a specific modification or feature.
'21 These tasks for BBR, which is what they were, they are 22 specific design tasks to support the BBR customer in 23 l the construction and licensing of the MKA unit,'and g- 24 I can't recall any explicit circumstance where we
(_.)g 25 said, hey, we will take that feature-and make a
. _ - . ~ _ _ . . _ .-. _ _
1 Roy 310 t 3 '2 proposal and go out to all our customers with it 3 .as something they might desire.
,. 4 Q Do you remumber any --
]lh; 5 A I am not saying-it didn't happen, but l 6 I don't recall any-circumstance where that happened.
7 Q Do you remember any specific project to 8 review the MK plant safety features.to assess whether
- 9 it would be useful.to incorporate them on U.S. plants?
10 A I don't recall of any reason. I was 11 hesitating - in the TPEP program, whether Mr. Quale, f 12 as part of that charter of the TPEP, would factor in (' 13 some of those features and give consideration. That 14 is one possible avenue, but I don't recall any 15 activity along that line. 16 Q That was after the accident? 17 .A 'Yes, after the accident. 18 Q You don'.t remember anything along the lines 19 I was asking about occurring before.the Three Mile 20 Island accident, do you? 21 A .No. 22- Q After the Three Mile Island accident, 23 'did you study the extent to:which the pilot operated f -~ 24' relief valve at the.Three Mile Island plant had.been "x/ l 25 qualified for the types of service that it was exposed-
]
4 I l
..- , - . . - - , , _ , . .-y. m. , , _ ., . ,,- .rr,,,.o
,.,f.., my, , _ , .s .m__ ,,y__-nw -
=.
1 Roy 311 ('\ k_) 2 to? 3 A Excuse me, may I have the question again? 4 Q After the Three Mile Island accident, did yud lll 5 any examination into the extent to which the 6 Pilot operated relief valve installed at Three Mile 7 Island had been qualified for the type of service g that it was exposed to? 9 A The only explicit thing that comes to my 10 mind now is involvement with the EPRI program, that 11 is one, where they were conducting a series of 12 operability and reliability studies for relief valves. 13 The second is the NRC required that the valve and 14 associated power systems be 1-E, and we have approached 15 some of our customers, certainly the backlog, I am 16 not sure about the oeprating plants, about providing 17 that service for them and meeting that requirement. 18 Q What is 1-E? 19 A safety grade. 20 Q Prior to the Three Mile Island accident -- 21 A No, this was after the Three Mile Island 22 accident. 23 Q I am asking a new question.
-} 24 Prior to the Three Mile Island accident, e '%) .
l 25 B&W had not specified that the vendor selling the
1 Roy 312 p-5.m) 2 pilot operated relief valve to B&W comply with nuclear 3 applications criteria for that valve; is that right? 4 It didn't have to be an N-stamped valve, did it? lll 5 THE WITNESS: Excuse me, can I ask counsel 6 a question? 7 MR. SELTZER: Sure. 8 (Witness conferrring with counsel.) 9 A Prior to Three Mile Island, I had no 10 knowledge of what the equipment specifications for 11 these valves called for. Most of those were prepared, 12 you know, before I got into that sort of thing. ( ,) 13 Q Did you learn sometime after the accident 14 that B&W had not specified that pilot operated relief 15 valves meet 1-E standards? 16 A oh, yes, I knew -- 17 MR. FISKE: Other than in discussions with 18 counsel? 19 MR. SELTZER: Right. I will show him l 20 l some documents that he got in April of 1979. l 21 A That's right. After the accident -- 22 MR. FISKE: I am just excluding that. 23 A It became clear that after the NRC required 24 that certain power supplics in the valve be upgrated
' ~. /
25 _to safety grade and 1-E.
I 1- Roy 313 l (~)- (. j 2 MR. SELTZER: Let me ask to have marked 3 fr identification as GPU Exhibit 458 a memo 4 from Fahland to Lee, April 15, 1979,
Subject:
Data Requested on TMI Elect.romagnetic Relief llh 5 6 Valve for D. H. Roy. 7 (Memo from Fahland to D. E. Lee, dated 8 April 15, 1979, was marked GPU Exhibit 458 for 9 identification, as of this date.) 10 MR. SELTZER: I would like to mark as 11 GPU Exhibit 459 a memorandum from Lee to Roy, 12 dated April 15, 1979,
Subject:
Data on ( 13 Electromatic Relief Valve. 14 (Memoranditm from D. E. Lee to D. H. Roy, 15 dated April 15, 1979 3 -was marked GPU Exhibit 16 459 for identification, as of this date.) 17 Q Is GPU Exhibit 458 a copy of a memorandum 18 which you received in mid-April 1979? 19 A Yes, it is. 20 Q And is.459 a copy of a memorandum that you 21 received in mid-April 19797 22 A Yes; it is. j!3 - Q Had you asked for information on the 24 specifications for the Three' Mile Island pilot ' operated 4 I V. 25 relief valve? i _ _ , . . . . _ - . . - s , ,_ . . - -
i ML , l n : .j 1 Roy 314 l
~-
1
-i
( , 2.
~A Yes.
3 Q And did you learn in mid-April 1979 that l 4 the valve specifications written by B&W had not { g- 5 specified "the operating conditions cuch as temperature, 6 humidity and radiation"? 7 A Yes,.that observation is made by Mr. Lee 8 in his memo to me. 9 Q Did.you also learn that B&W hadn't
~
.. u 10 specified how many times the valve must be called on 11 to work over a given period? 12 A Yes, that observation was made by Mr. Lee () 13 to me in this memo. j 14 : Q Did you also learn.that B&W did not 15 require Dresser Industries, the manufacturer of the ,
-16 TMI pilot operated relief valve, to perform any 17 qualification testing to insure that the valve will 18 open or-close when asked?
/.
19 A The second page? , i 20 0 The second page, point 2. -
; 21 MR.-FISKE: Are you reading from the memo?
22 We will stipulate what is in the memo. It
+
23 says' what'it says and it was addressed to Dr. I l 24 Roy. 25 .A -YcA, these are based on Mr. Lee's review g . s 4 N- w e -,- - , e a
1 Roy 315
-) 2 of the specification and his observations as he saw 3 them-4 Q Did you ever hear anything or learn anything lll 5 that was inconsistent with what Dr. Lee has reported 6 to you on April 15th regarding the matter that I just 7 read?
8 A I don't recall arriving at say concitsion 9 one way or the other. 10 Q All I asked you is, did you ever get any 11 information centrary to what Lee told you, namcly, 12 that B&W did not require Dresser to do any qualification 13 testing on that pilot operated relief valve to 14 determine whether it will open and close when asked to? 15 A No, I haven't received any inform 6 tion 16 Personally myself that would say, you know, that 17 wasn't a right statement. 18 Q O.K. 19 A But -- 20 Q That is all I asked. 21 A I haven't arrived at any opinion myself 22 of whether, you know, what qualifications versus 8 23 some other kind of testing. l 24 Q Did you also learn that B&W did not 25 include in its specifications for the pilot operated 4 1 l
n 4' s
- g en n ' .\ \
jj s. ). N.
.- - p ..
v ~g i.Is '.-,' I1 . Roy 316 i , ,. f~' g i 2 relief valve that it be qualified to pass water in i 4'.
-'\4 3 addition.to steam? I a
4 A Yes, I have heard that about the specific h 5 specification for the valve. Since Three Mile Island.
'6 ; See, this,whole -- I i
.., j 7 . Q Speaking as the manager of engineering,
,c\
- J= i 8 which you were at the time of the Three Mile Island g w , ~
- p. - .. g accident, what justification, if any, are you aware of
~
l
- r+y.
j, , , . j O ,, for failing'to specify that the pilot operated relief
- n. ,.~~
~
9 b 11 valve be tested by the manufacturer to see'whether it I 12 would open or close? i-13 - MR. FISKE: I will object to that question, 14 Mr. Seltzer. Dr. Roy wasn't involved in
}S ' issuing these specifications at the time they i..
to were issued.
~
17 MR. SELTZER: I understand. ik
, i 18' MR. FISKE: Whenever.that was, back in 19 1969 or 1970.
1
)
i
=1 20 MR. SELTZER: 'I am just asking if.he has 21i any understanding from conversations with
' ' e; others, from documents that'he'has received.
,. 52
~
2C, If. .he doesn't, he should say he doesn't. 1 \U<. l - s Ma.:FISKE: In~other words, does:he have p 24 @u] % - any understanding as to why B&W didn't do that? 7, . t_Pev 25 __
"') g
( Y. py
, _ . . _ . . _ , c_. _ _ . . _ . _ . , . _ . , , _ 1_ . . _ _ . . .. _ _. . . . . . , ,_
g Roy JJ7 (~N (_) 2 MR. SELTZER: Right. I am not asking for 3 any cerebration on the spot today. 4 Q I am asking whether you had any i 4 lll 5 understanding or background at the time you learned 1 6 that B&W never required Dresser to test this valve 7 to see whether it would open or close. I 8 MR. FISKE: Do you have any knowledge -- 3 9 A First of all, you have a characterization 10 and look specifically at what kind of tests are 11 performed versus qualificat' ion testing. That doesn't 12 encompass the full range. There is leakage, stroke j ) 13- testing, this kind of thing, done by the vendors. 14 And secondly,.I don't recall arriving at 15 any judgment at all about that specification for that 16 valve. That valve is provided -- that is an isolatable 17 flow path. That valve has a capability of isolating 18 it upstream. You are looking at the context of-the 19 valve service and the system that it is set in, 20 and I didn't draw any conclusions one way or the other 21 about the test program for'the valve.
~
22 Q This was.a valve that was supposed to open 23 .and shut when it was asked to, wasn't it? A Yes, that would be the. desirable mode of c'( 24 . s / 25- operation.
1 Roy 318 ('% (,) 2 Q Didn't you think-it was unusual for B&W 3 to have specified a valve that was supposed to open 4 and shut but didn't ask for qualifications to see lll 5 whether it would perform its intended. function? G MR. FISKE: Are you asking whether he , 7 thinks that today? 8 MR. SELTZER: No, whether he ever thought 9 that before today. 10 A Drawing that conclusion about the 11 specifications for the valves? 12 Q Yes. O 13 A No, I haven't drawn any conclusion. j 14 Q Do you know whether any qualification 15 testing was specified for the block valve to see if 16 it could pass water successfully? 17 A I am not aware of that specification. 18 MR. FISKE: You mean when it was closed?
. 19 MR. SELTZER: No. Before it was closed.
20 MR. FISKE: I don't-understand. You mean 21 ,when it was open, would it allow water to go
- 22. through?
23 MR. SELTZER: Yes, and then not jam up, 7- . _ , 24 freeze up so it couldn't close. That is the I
..Q ,/ -
25 real inquiry, but he said he.doesn't know about g , w 9
I l L 1 Roy 319 l 2 qualification testing. , 3 A I d n't know what the testing I 4 specifications were for that valve. lll 5 Q So y u don't know whether it even had 6 to be tested to see whether it could close against a 7 water flow, do you? 8 A I don't know whether it was or wasn't. . 9 MR. FISKE: You mean if somebody tried to 10 close it, it wouldn't close? 11 MR. SELTZER: Right. 12 A If you are asking me what is my knowledge () 13 of the test program which is prescribed by B&W, I
}4 don't know what that is for that valves 0.K.? That 15 valve sees a lot of service.
16 Q - Was the engineering department required to 17 function on the equipment specification for pilot 18 operated relief valves prior to the Three Mile Island
- 19. accident?
20 A Equipment -- the engineering department
~
21 prepared the equipment specifications-for procurement O 22 of such valves-if they were part of our scope of 23 supply,.and.to my knowledge.it was. 1 "NL 24 Q .Who or what unit-in engineering would have J
- 25. prepared.the specs for.the pilot operated relief valve?
w.
g Roy 320 2 A The procurement specifications would have been prepared by -- well, you've got to identify the 4 time frame. Today that section is called equipment '
$ 5 engineering. As far as I can recall, that was the 6 name it went by before, while I was in plant design 7 and I became engineering manager.
8 Q Since the Three Mile Island accident, do 9 y u know if the engineering department has upgraded 10 the specifications for the pilot operated relief valve? gg A I can't recall that explicitly. The reason 12 I am hesitating is that in conformance to -- response 13 to a customer, and I believe it's consumers, we had 14 ffered to them the service of upgrading that valve 15 in compliance with the action plan requirements. 16 Q Yes. A g7 But I don't know the extent to which the 18 equipment specifications.have been prepared and gg upgraded, so the requirements are in compliance with 20 the NRC 0737. I can't. recall any other instance. 21 Q What, if you know, caused the TMI-2 pilot h 22 perated relief-valve to fail open? ; g A .I have no idea what caused it to' fail open. 24 ~E' Have-you discussed it.with'anyone? 0: ,, A no. vossiul. caus.s,
1 Roy 321 j 2 Q Yes. 3 A No. I have no recollection of doing that. 4 Q Do you think it was the same problem that jlh 5 caused the Davis-Besse valve to fail in the test 6 that b&W ran in the fall of '777 7 MR. FISKE: You are asking whether he is 8 speculating on that? 9 Q Has it ever occurred to you that it was the 10 same as what your own tests had shown in 1977 that 11 caused the Davis-Besse valve to fail open? 12 MR. FISKE: Go on and answer. Ij 13 A I haven't speculated on it at all. Two 14 different valves, you are talking about those two 15 plants, quite different, but I haven't sat down and 16 thought about all the possible modes or combination 17 of events. 18 Q I just asked you about that one mode. 19 You haven't thought about that? Have you discussed it 20 with Paul Perrone, let's ask you that? 91 A No. 22 MR. FISKE: I will say on the record, 23 j Mr. Seltzer, that my understanding is that that r 24 valve is still in the plant, inaccessible because t 25 of radiation, and we would join with you in any
1 Roy 322
'(N ).I
'2 . objective effort to determine what the cause 3 of its fail-open was, but to my knowledge, that 4 cause has not been determined.
l lll 5 MR. SELTZER: We are waiting for a Rule 34 6 request from you to inspect the valve. 7 MR. FISKE: We are drawing to see who gets 1 8 the short straw to go and see who will do that. 9 MR. SELTZER: I would like to mark as GPU
- 10 Exhibit 460 a meno from Sullivan of PS&P to 11 Roy, Manager of Plant Design, dated July 22, 4
12 1977,
Subject:
Draft NPG-0151-81G. 13 (Memo to D. H. Roy from F. W. Sullivan, 14 dated July 22, 1977, with attachment, was 15 marked GPU Exhibit 460 for identification, as 16 of this date.) 17 MR. SELTZER: I would also like to mark 18 as GPU Exhibit 461 Liz Brooks's memo to Fred 19 Sullivan,, dated _ August 29, 1977, attaching Dr. 20 Roy's comments on GPU Exhibit 460. 21 (Memo from Liz Brooks to Fred Sullivan, 22 dated August 29, 1977, with attachment, was 23 marked GPU Exhibit 461'for identification, i
~ 24 '. as of this date.).
s/' 25 Q_ Is GPU Exhibit 460 a ~ copy of a memorandum
, , _ . . _ . ..- =_ _ _. . _ _ _ _ .
L 1- Roy 323 4
. m
_): 2 which you received from Sullivan in July of '77 with ' 3 s me attached materials for you to review? 4 A Yes, it is. lll 5 Q And d es GPU Exhibit 461 contain a copy J 6 f some responsive material which you prepared after I 7 reviewing GPU Exhibit 460? g A Yes, it does. l g Q Can you identify what the attachment is , 10 to GPU Exhibit 4607 1 , 1 11 A It isa summary of the objectives and 12 responsibilities of the plant design section. () 13 Q The purpose of this was to set forth in g4 B&W's manual of policies and procedures what the scope 1 15 f responsibility was of each of the units-within 4 16 the plant design section? i 17 A Yes.
}g Q And you reviewed it'to see that you thought 19- it comported with what you believed those i
20 responsibilities should be?' 21 A Yes. g. 22' O. Take a look at the pages that refer to the --,
- 23 ECCS analysis unit, beginning at'page 79. That'is l
-page 11- f the procedure.
. (^g -24. -
l U 1-Is that the' portion where the r -
. 25 A
's,y---' %-e>g ;- -gi--+ .__.p.p-.. w ,e v wW ,,m,...,9 .w,g4, 9w 9 -y *e..,
w rn - 9 y e-,---,.-
-. . . _ _ . _ = . - . - _ _ _ _ . . _ . _ _ _
i Roy 324 A. ,-j _) 2 . responsibilities of Bert Dunn's emergency core cooling 3 systems analysis unit are set forth? , 4 A Yes, it is. W uld y u turn to the second-page. (ll 5 Q Do I 6 you.see item 4.67 7 A Yes. ' .i g Q Is it correct that one of the g responsibilities of the ECCS analysis unit was to i. 10 " review test and operating procedures for ECCS .; 11 systems and equipment"? 12 A That is so stated here. i 2 ( 13 Q D es that mean that operating procedures 14 being prepared by B&W for utilities were to be reviewed 15 by the ECCS analysis unit if they related to emergency 16- core' cooling systems and equipment? r .. 17 A This was a capability which was provided 18 by that-expertise in that unit. The responsibility-19 for any procedural activities vere in the' customer l service department, whatever obligations we had to a-20
~21 customer, depending on.the contract.
.O 22 But this was a capability that was provided,
- , 23~ could be provided by this particular unit so that if I
.,r
.m. .og in'the preparation of a startup test procedure or
~
25 'whatever,1the' service wanted -- the service department'
~
r, , , , , , _ _ _ _ ,, . _ , _, ,, . . _ . . __ ___ _ , _ _ _ , [_
1 Roy 325 C
\-)\ 2 wanted to get that reviewed, that the scope -- the 3 scope of'those tests that were related to ECCS 4 systems would be the responsibility of this unit.
llh 5 Q To what extent was it the practice of B&W G to ask ECCS analysis to review operating procedures 7 affecting ECCS systems and equipment prior to the 8 Three Mile Island accident? 9 A I have no personal knowledge of the extent 10 of that kind of review that went on, if any, between 11 the two organizations. 12 Q Let me show you an " eyes only" memo drafted 13 by EAW in July 1979. 14 MR. SELTZER: We will mark it as GPU l 15 Exhibit 462. It is entitled " Performance . 16 Engineering Concept, Internal Use Eyes Only." 17 Q Let me ask you while we are going to dig 18 out a copy of Exhibit 462 what the expression " eyes 19 only" means to you? 20 A That is a new one to me. As an author 21 might stamp it on a memo or something? O
- 22. Q or have his secretary type it on.
23 A I could only surmise. It must have been 24 ;the eyes -- all right, I won't' surmise. I don't know. 25 I don't have any image of what that might mean.
-m _ _ . _ _ _ - - _ - . -
1 Roy 326 ( 2 MR. FISKE: If you want to show him a 3 specific document and ask him if he has 4 ever seen that document, that's fine. 5 MR. SELTZER: Off the record.
= lll 6 (Discussion off the record.) l 7 (Document headed " Internal Use Eyes Only -
i 8 Performance Engineering Concept" was marked 9 GPU Exhibit 462 for identification, as of this 10 date.) 11 Q Is GPU 462 a copy of a memorandum you l 12 received in or about July of 19797 () 13 A As I was reading it here I am try3ng to 14 refresh my memory of ever having seen it. 15 I don't recall receiving or studying this 16 document right offhand, right now. 17 Q Have you finished reading it? 18 A Yes. 19 Q Dr. Womack reported to you in the summer 20 of '79; right? 21 A Yes. O. 22 Did you have discussions with him about Q 23 the idea that he is principally putting forward in l 94 GPU Exhibit 462?
. (/
N~. 1 25 A We had discussions of establishing a a
1 Roy 327
/
(/ 2 performance engineering unit in the plant design 3 section. 4 Q What disciplines would have been part of the performance engineering unit? llh 5 6 A The concept that we had was -- for that 7- particular unit was that it would be a point contact 8 that would interface with our service department 9 on issues that could arise with respect to operating P 10 plants in a broad front, that they would be a place, i 11 unambiguous contact point that issues arising with 12 respect to operating plants, that people could make , r
- 13 that contact point there and the performance engineering 14 people would know the particular disciplines. In the 15 rest of the engineering they could go to and counsel --
16 - for counsel and help but one of the concepts for the 17 disciplines that would be in performance engineering 18 would be people that would have field experience, 19 knowledge of the -- to the greatest extent possible 20 of the systems that were then existing on the 21' operating plants and could be a coordinating unit 22 for' issues which would involve the operating plants. 23_ Q Is that an idea that was suggested as a jr' 24 _' result of'the Three Mile Island accident?
'L/
25 A I can't say what the -- I don't recall 1
r 1 Roy 328
.fM (s / . 2 what the driving force for that was right now, what 3 the initiative. Certainly it had some impact on it.
4 Q How? lll 5 A In looking at -- you recall the three-6 volume bluc bcok. One of the sarvices that we 7 performed for our customers was the preparation of 8 guidelines for managing the small break in compliance 9 to bulletins and orders from the NRC and -- 10 Q You are referring to the small break 11 operating guidelines? 12 A Yes. One of the extensions that we had I) 13 in mind of that service was that'we would extend that 14 to other kinds of transients and we wanted -- we ! 15 were focusing the responsibility for that kind of 16 effort in -- for the coordinating of that kind of 17 effort in plant performance engineering, so that way 18 indirectly, you know, the three-volume blue books 19 were a consequence of the bulletins and orders 20 following that, and that provided an opportunity to 21 do that kind of service for a broader range of O~ 22 transients. 23 Q You have talked about having a performance i jr 3 24 engineering unit.
'G 25 In paragraph 2 of GPU Exhibit 462,'EAW,
, . - =
l l 1 Roy 329 l 1 2 whom we might assume is Allen Womack, says "the 3 performance engineering responsibility will be 4 assumed by the unit which has the responsibility for lll 5 the system analysis of the particular transient under 6 the present organizational arrangement." 7 Do you recall discussing with him that the 8 ECCS analysis unit that already existed within plant
~
9 design-would be responsible for overall performance 10 of B&W plants and response to primary loss-of-coolant 11 accidents? 12 A Yes, specific issues that would arise (s) 13 for the ECCS system would then be directed -- it may 14 be the responsibility of the ECCS unit, the specific 15 functional area that had that cognizance and expertise. 16 Q Was one of the thoughts that the ECCS 17 analysis unit would be expected to function on the 18 drafting of procedures for the operation of the 19 high_ pressure injection system, to the extent that 20 Bsw functioned on drafting such procedures? 21 A With respect to the service that we O 22 performed in drafting guidelines for the small break 23 analysis, there is an issue that specifically related to the.ECCS system. They are not drafting procedures,
/~ $y D
25 they are guidelines and other descriptive material
1 Roy 330
~s
) o
~ which we related to and would support the preparation 3 f those guidelines, that the principal responsibility 4 f r performance cf any detailed analyses to support lll 5 that and the supplementary material would be ECCS 6 analysis unit and was the ECCS analysis unit.
7 Q What do you mean was? g A For the small break guidelines. 9 Q Oh. 10 And what was done for the small break 11 guidelines was a pattern that you were considering 12 applying generally to other ECCS guidelines that b '; 13 e uld be developed in the future; isn't that right?
\ _.)
g4 A Other transients. Other transients, not 15 just LOCA. ECCS would be related to the LOCA but 16 other transients as well. 17 Q Was it also contemplated that if this 18 performance engineering concept were adopted, that 19 if operational problems occurred affecting ECCS 20 systems or equipment, the ECCS analysis unit would 21 be expected to participa'te in an evaluation of that O 22 operational problem? 23 A Yes. And one of the units of performance 24 engineering w uld also be informed of such matters 25 and inv lve the ECCS unit where it was appropriate. l 1
1 Roy 331 2 Q Was there any impediment to having 3 IMPl emented this performance engineering concept 4 prior to the Three Mile Island accident? , 5 MR. FISKE: I don't understand what you ll 6 mean by impediment. 7 Q Would it have been feasible to have 8 implemented this before the Three Mile Island 9 accident, based on your managerial overview? 10 A I don't see any particular -- I can't 11 recall any particular impediment that comes to my 12 mind now. of course, as I say, this concept -- this () 13 concept, as best I can recall, is precipitated now 14 by this broader service that we are providing, of which , 15 the small break guideline was our first -- and 16 systematic way, the first approach at doing this, and 17 we hoped to do that for a broader range of transients, 18 so that was kind of the initiating idea that pulled 19 this together, but it wouldn't be an impediment to 20 setting up a unit like that before'the accident. 21 Q You prepared a " Top Ten Lessons Learned"
- 22. memo for John-MacMillan in June of 1979, didn't you?
23 A Yes, I recall such a memo. 24- Q What was the purpose of it? O
'~Y A As best I can recollect, there was a 25 n ,y- e 4- e m ww *9'
\ l 1 Roy 332 I
L o rcquest from Mr. MacMillan, I believe, part of a 3 staf f e:cercise , to identify what each of the staff 4 members thought were the top ten lessons to be learned f lll 5 in a generic sense from the Three Mile Island accident. 6 And I solicited input from the various staff members, 7 and I believe that memo, as best I can recall, 8 represents a summary of those lessons and inputs. 9 Q You recognized after the Three Mile Island 10 accident that it would be desirable to include in 11 transient or casualty analyses of realistic or more 12 prcbably accidents than the worst case accidents that 13 had been the focus of your studies before the Three 14 Mile Island accident; isn't that right?. 15 A Yes, and one of the best examples is in 16 the guidelines. We approached that in providing the 17 service of preparing guidelines for the management 18 of specific kinds of transients. .There you would of 19 course, of necessity, want to model the transient that c 20 was called the best estimate rather than in a bounding 21 sense. 22 Q In other words, by doing analyses of more 23 probably or more realistic casualties, you were able 24 to develop better guidelines for handling the more
-25 probably accidents?
A Yes. And some of the material was to
1 Roy 333 _f~h *
- (,) 2 assist the customer in supporting the guidelines 3 with descriptive material of what the basis for those 4 guidelines were. Then you would be looking at parameters as a function of time. That should ll 5 6 correspond to what the plant will actually do, as 7 best you can estimate it, versus an accumulation of 8 concervatisms which could, not always, but could 9 provide a different kind of representation of what the 10 plant's response would be to that particular accident.
11 Q Before the Three Mile Island accident, the i 12 analyses that were done were based on the accumulation
-( ) 13 of conservatisms?
14 A For the purposes of those analyses which 15 were done to support the licensing of the safety 16 systems, they are done in the bounding analysis 17 approach. 18 Q And that bounding analysis approach has the 19 accumulation of conservatisms that you referred to in 20 your last' answer? 21 A Yes. i 22 Q There was no impediment before the Three 23 Mile Island accident to your doing the type of yS 24 - casualty analyses that were based on more realistic
%)
25 and more probable casualties, was there?
1 Roy 334 fs s- 2 A Yes, I don't know of any impediment and 3 I d n't know it wasn't done. I don't recall any 4 specific instance where we would do best estimates. , l lll 5 There was casual talk o.f best estimate kind of 6 analysis prior to the accident in a generic sense by 7 the NRC, but I don't recall any specific analyses that 8 we did, what we call best estimate. 9 Q After the Three Mile Island accident 10 did you believe that it would be desirable to close 11 the loop within B&W between the people who worked on 12 procedures, the people who worked on analysis, and (~) (_j 13 the people who worked on operator training? 14 A The concept that comes to my mind after 15 Three Mile Island is broader than that. It is closing 16 the. loop in a broad sense from designer, analyst, 17 operator training, or enhancing that loop, you know, 18 looking at what we could do to make that loop tighter 19 and more complete, but not just B&W, reaching out, 20 customer, the operator, their people who did the 21 procedure preparations, analysts, designers, in a 22 generic concept. That is more what I had in mind 23 when you talk about closing the loop after Three Mile 24 Island.. N.J 25 Q Why did you think that would be desirable? i.
1 Roy 335 N/ A Enhance overall plant performance. 2 3 Q What did you think could be gained by 4 greater communication between the B&W elements of the lll 5 lo0P? 6 A Well, I think -- excuse me. Just within 7 B&W, analysts, designers, guideline writers? 8 Q Yes, what would they gain from enhanced 9 communication among themselves as part of the closing 10 of the' loop? 11 A Well, there is the potential in identifying 12 those communication paths more explicitly and being l .O ( ,I 13 able to provide better service to the customer on 14 a broad front: analyses, guidelines. That is what 15 we hoped to gain, if the potential is there, to 16 improve that, realize that potential. 17 Q There was no impediment to B&W closing its 18 part of the loop before the Three Mile Island accident, 19 was there? 20 A I don't have any evidence it wasn't closed. 21 I am talking about -- if we focus on that, there is 22 potential for improving the overall service to the 23 customer. We want to realize that by focusing on
/[ 24 that particular concept.
'g 25 Q You say-you had no evidence it wasn't
l l l 1 Roy 336
}/~Q.
(_) 2 closed, but isn't it a fact that you know of no 3 procedure.that was in effect before the Three Mile 4 Island accident for anybody in the engineering department lll 5 to review the contents of training programs before the 6 Three Mile Island accidents isn't that right? 7 THE WITNESS: Have I testified on that 8 issue? . 9 MR. SELTZER: Not here. I think you 1. 10 testified to that before Rogovin or Kemeny.
.11 MR. FISKE: Just answer the question.
12 Q Why don't you just answer the question. () 13 A There is no systematic or procedural tie. 4 14 You know,- there is no impediment to dialogue. occurring 15 and I know dialogue did occur. 16 .Q Other than random dialogue, was there'any 17 Program -- 18 MR. FISKE: He hasn't characterized it as 19 random dialog. 20 MR. SELTZER: 'Finc. That is why I an 21 asking further. 22 Q Other than random dialogue, was there any 23 . program in place so that engineering department r 3: 24' Personnel would review and approve the content of ()-. ~ the training programs that' existed prior to the Three 25-
- - ' - w W+ e , y w m wy
1 Roy 337 (~) (j 2 Mile Island accident? I am talking about the B&W 3 training programs. 4 A We had no procedural requirement for doing lll 5 that. 6 Q Was it done as a matter of routine, to 1 l 7 your knowledge? 8 A Approving the content of the training 9 programs by the engineering department, I am not aware 10 personally of any approval process or material submitted 11 for approval myself. I am not aware of all the 12 dialogue that went on between the training and the .l l ( 13 functional experts in engineering, so personally I am
.14 not aware of material submitted for approval. ,
15 Q All right. l 16 A Personally. 17 MR. FISKE: I am not sure he finished his 18 answer. 19 THE WITNESS: ~ Personally. 20 Q Are you aware of any other routine by 21 which engineering department personnel systematically 9 22 conferred with training-department personnel about the 23 content of the-B&W training program? s 24 A It was, to my perception, relatively j
' 'N 25 routine for the training people to call upon certain
1 Roy 338 2 experts in engineering in doing lectures, O.K. That 3 would not be a systematic program, O.K.; that would be 4 one instance where, you know, there is dialogue in lll 5 the training department calling on the expertise of 6 the engineering department, but there wasn't a 7 systematic, you know, program for doing that. 8- Q Let me see if I can summarize accurately. 9 While there was no barrier to communication 10 between the training department and the engineering 11 department, and while there was some communication, 12 you don't know of any systematic review by the 13 engineering department of the content of the training [~)T ( 14 l program before the Three Mile Island accident; is that 15 a fair statement? 16 A Yes, that's true. I am not aware of the 17 volume of the informal dialogue that.may have gone 18 back and forth across, but there was not a systematic 10 approval process in place. 20 Q Or even a systematic review, right, of the 21 substantive. content? O 22 A That is true. 23 Q Was there any. systematic review by the 24 engineering department of procedures that were drafted
~ O, 25 and sent to customers.before_the Three Mile Island
I t
. i 1 Roy 339
. O
.O 2 accident?
3 A I believe that it was routine and may have 4 been systematic for -- well, let's see. No, I am lll 5 thinking of test specifications which I know were 6 reviewed frequently and I think on a systematic basis, 7 but not with respect to procedures, that I can recall. 8 Q Other than through Mr. Fiske or other f g counsel for B&W telling you, do you know that before 10 the Three Mile Island accident B&W had written, 11 reviewed and approved an operating procedure for
- 13 loss-of-coolant accidents that included a 50 degrees
( ) 13 subecoling rule along the lines that Dunn had proposed 14 in his February '78 memorandum? 15 A No, I had no knowledge of that. 16 Q You don't know about that even today 17 other than through counsel? 18 A That's true. 19 MR. FISKE: Could I hear the -- well, I 20 guess it doesn't make any difference. The answer-21 was "no." 9 22 I will n te a retrospective objection to 23 the form of the question. 24 - MR. SELTZER: Arc those valid, Mr. Fiske? V MR. FISKE: I-believe in the case where the , 25
. . - __ .= -. - _ . - .
4 1 Roy '340 (^h (_) 2' witness says "I don't know," they are. 3 Q Did you come to conclude, after the Three 4 Mile Island accident, that the procedures which B&W lll 5 had drafted for handling transients were in some cases 6 not ones that permitted quick and accurate access to 7 and interpretation of -- i 8 A I have no -- 9 MR. FISKE: If you are referring to a quote 10 from some document, Mr. Seltzer, I think it is 11 only fair that you show it to the witness, 12 particularly since it is six p.m. and we are (~N trying to get out of here. () 13 14 MR. SELTZER: Let me show how mindful 15 I am of the time and solicitous of Mr. Fiske's 16 suggestion. 17 I will ask to have marked as GPU Exhibit 18 463 Dr. Roy's " Top Ten Lessons Learned - Generic," 19 dated June 8, 1979 and addressed to John MacMillan. 20 (Memorandum from D. H. Roy to J. H. 21 MacMillen, dated June-8, 1979, was marked GPU
, 22 Exhibit 463 for identification, as of this date.) ,
23 ,. Q Is GPU 463 a copy of.a memo that you sent je N 24 to VP MacMillan on or about June 8, 19797 b (
%,/ -
25 A Yes, it is. )
.u . - ,,
1 Roy 341 m 2 Q Do you see in item 3 where you said "Must 3 find some way to permit operator quick, accurate access 4 to and interpretation of appropriate procedures (a lll 5 10-page procedure buried in a file cabinet or thick 6 loose-leaf binder is inadequate)"? 7 A Yes, I see that. 8 Q What did you mean by that? 9 A Well, that an approach to management of , 10 operating transients, which was based on having access 11 to a large body of procedures, O.K., and I don't know 12 how large these are in the control' rooms -- that I () 13 believe that that needed to be looked at and see if 14 we could improve the ability of the operator to, you 15 know, diagnose his transient and improve his ability 16 to manage it. 17 Q Are'you familiar with the ATOG program? 18 A Yes. 19 Q Do you think that'that'is a program that 20 is directed at the objective that you have just 21 described? 22 A Yes. 23 Q Is it correct that one of the objectives , l lf'T 24 of the ATOG program is to direct the operator's t :
%d 25 attention to treating the symptoms which he observes
p 1 Roy 342
.v O)
L 2 rather than having to diagnose what has caused the 3 transient? 1
'I A In the early portion of the transient, i
4
]ll 5 that's right, it directs him to certain fundamental 6 Primary and secondary variables and he manages thu 7 transient on the basis of his observations on the 8 basis of those variables. l 9 Q By variables --
10 A Pressures, temperatures, this kind of thing. 11 Q And he does not have to diagnose in the 12 early s'tages of the transient what has caused those I (l U 13 variables to change as they haver is that right? 14 A Within the concept of the ATOG, that's 13 right. . 16 Q Is it your understanding that that is 17 intended to be a simpler approach for the operator, 18 a more direct way to treat a transient? 19' MR. FISKE: Than what? 20 Q Than the method of-accessing and 2[ i interpreting procedures that you were referring to in 9 22 item 3 of GPU Exhibit 463. r '23 A Yes,_it is a concept that we thought could 24 ' improve that management of transients, enhance that, I h~ "
' \ _)
- 25 fand the best I can recollect now, the'NRC requires at
,_a, y.
- p. -
1 1 Roy 343 y 2 least an evaluation of a symptomatic approach to 3 transient management. 4 Q You are referring to the ATOG approccr. lll 5 as being a symptomatic approach? 6 A Yes, during the early -- managing the early 7 portion of a transient. ' 8 Q Dunn's prescriptions f or operator operation 9 of high pressure injection formulated in February 1978 10 are a symptomatic approach to system management, 11 aren't they? 12 A I -- 73 ( ) 13 Q Let me break it up into smaller piuces. 14 Dunn's prescription -- and we can get it 15 out if you would like to have it in front of you } l_ t 16 before you answer -- says that if high pressure ' 17 injection has come on, leave it on until you deternine 18 that either of two criteria arc met; isn't that right? 19 A Yes. As I recall. 20 Q High pressure injection autonatically 21 actuating is one type of symptom that the operators O 22 would observe; isn't that right? 23 A That's not exactly a symptom. That is a , I 24 l consequence. That is equipment being initiated as a 25 consequence of certain variables, you know, following* ,; h m
\
1 s.
\ i l
.e
> .7.-~ ; - .
((. 4. w ~.- - ' l 1 Roy l ,, 344
',('
~ .,
2 or takington a certain value or state, combination of t %
'\
i - 3' f variablei -- in this case, it ic only one variable --
~~ .
a 4 . and that -- the reason I am hesitating, that doesn't ^
.l 5 match up. I have to think this through. That doesn't
! 'G match'up with a symptoms-oriented approach. 7 . Q Let me take it further.
. 8 i In order,for an operator to apply Dunn's
- i-- i
\
9 ' prescriptions, which B&W sent out after the Three k> ,b A' N 10 Mile Island accident, the operator is not directed to iit * 't 3 n.s S(' 1,U
=
analyze what casualty or upset condi. tion caused high w .. i \ ' [ I ,-- ' 12- pressure injection to actuate?
] 13 % That is true.
14 C; In that sense,-Dunn's instructions are
.s * ; 15 simplified instructions like the AToG instructions; I
i
. g - -
A-16 isn't that true? s 3 . 4 < W
,Q [~ sy b ,N
.e -
f r,; +- - MR. FISKE: Well, I am going to object
, ,. - w . .
,18 to the linking of the two.
3 If you want to ask 4 19 f .him whether or not they.are simplified 4 N , ( '20 g it.s t ru c ti o n's , I have no problem with that. g 521 ..
*N ~
MR. SELTZER: I will accept that suggeation. 8 .w. QQ ' 9_ . .
^
22 A Yes, I think they would qualify for the 4,, W. ~5 s ..
desbription for those-- '$for that preheription. (3 V ~23 -
- b. 7 - ' A; .
, e .
ju j. 24 Q W, v.In other words, they are a more simplified
.' h.y M h) -s
/
?e . 25 ' approach to; procedures.than the procedures that you 9 ,. s . f y,. _s
,a p: es ~ .r --,e e- - y , - - - - e- wa --
1 Roy 345 r\ ' 1h 2 understood the operators otherwise had to refer to? . 3 A That time frame, I didn't know what 4 procedures were out there. lll. 5 g All I am trying to do is get a comparative. 6 You responded to Mr. Fiske's suggestion that they were 7 more simplified. Since you agreed, more simplified 8 than what, Dr. Roy? 9 A I don't recall the procedures out there. t !- 10 I think a better characterization of those particular l' 11 ones would be like -- well, obviously they would be 12 supplementary to what's there, but taken -- taken i [ .C) 13 separate from what's out there, o.K., just the 14 prescription itself for leaving high pressure injection 15 on,-in terms of them being guidance: if it comes on, 16 do so-and-so. That-is a symptomatic approach to 17 guiding the action that the operator night take with 18 respect to high pressure injection once it is ; 19 initiaced. That I can state. 20 Q And that symptomatic approach is cimpler 21 than asking the operator to diagnose what has caused O 22 the. transient and caused the automatic actuation of 23 high pressure injection; isn't that correct? I 24 MR. FISKE: I will object to that.
,/~') .
sm 25 _ Q well, let me ask you, based on your
i
=1 Roy 346
\ 2 familiarity with Dunn's prescription and other 3 Procedures that you were referring to in GPU Exhibit 4 463 --
A In all respect, I think we are mixing
. lll 5 6 apples and oranges.
7 MR. FISKE: I think the problem here is i 8 that Dr. Roy has said that he was not familiar 9 with the content of the procedures that were in 10 effect prior to the Three Mile Island accident, 11 so he_was familiar with what Mr. Dunn wanted 12 to do, but I think you_are asking him to compare 13 13 that with something he'doesn't know anything g
).
-14 about.
15 Q That is true, you have never looked at the 16 loss-of-coolant accident procedures that were in 17 effect at-the time of the Three Mile Island accident? 18 MR..FISKE: No, I said -- 19- MR. SELTZER: Let me ask him that. 1-20 MR. FISKE: Has he ever, outside the 1 21 presence of counsel? lh :22 MR. SELTZER: Yes. I l l 23 A No. j l
~
L 24 MR. SELTZER: I will.ask to have marked as
- y. .
)
~,
25 ' GPU~ Exhibit 464 a set'of materials relating to
, , , , -__ .a. _.- 2
1 Roy 347
's,j/
2 a speech that Don Roy was to give to the ASME 3 1979 winter meeting. The cover letter is dated 4 September 21, 1979. lll 5 (Letter dated September 21, 1979, to Dr. 6 Don H. Roy from Mr. Snyder, with attachment, was 7 marked GPU Exhibit 464 for identification, as of 8 this date.) 9 Q Did you give a speech before the American 10 Society of Mechanical Engineers winter meeting in 11 December 19797 12 A Yes, I did. J
)~ 13 Q Beginning at the fourth page there is 14 material that is headed " Trends in Improved Reactor 15 Safety Design."
16 A Page 47 This is 4. Yes, I have that now. 17 Q Is that a copy of or a draft of the speech 18 which you gave? 19 A No, it isn't. These notes were prepared 20 for me by Mr. Taylor. I had contacted him and gave l 21 him the general subject of the talk I was going to e 22 give at the ASME'and asked him to write down some of 23 his suggestions as sort of topics or rough skeleton r' 24 for the paper that I would present, so-these are -- ss
'25 these were his inputs to me as I put together my talk
F 1 Roy 348 2 before the ASME. 3 Q Did you rely on his notes in preparing 4 your speech?
. llh 5 A I don't know to what extent I factored 6 these in. A couple of things come to mind. I did 7 talk about the entire generic issue of closing the 8 loop which we talked about, I did talk about 9 availability and safety, that coupling was emphasized; 10 I recall that specifically in my talk. I did give a1 11 brief summary of ATOG in my talk, which wasn't, I 4
L 12 don't think, addressed here. They are'the recollection 13 that is in-my mind now. 14 Take a look at page 2 of the draft speech. Q 15 Did you make any comments about LER's 16 - frequently ending with words that tend to focus on J-j 17 the plant'ending up in a safe condition with no threat i 18 posed to the public health and safety? c 19 -A I don't recall using that input or that 20 being in my talk. 21 Q -Did you prepare a typed version.of the 22 talk that'you ultimately-delivered? 23 A. No, I' don't believe so. Well, as a 24 matter of fact -- no, there was no request for a- {['}-
\_--
' 25 paper and I don't recall preparing one.
f J g 4 - ( '7 -* 3 y vena- 4. & .e w , * - - - we g 9m mv+-a v,y
1 Roy 349 7%,) 2 'Q Did you ever-discuss the problem of closure 3 on Bert Dunn's prescription with John MacMillan? 4 A No, I recall no such-conversation. l 5 MR. SELTZER: Let's end with the one-6 paragraph memo from you to Plant Design Unit 7 Managers, subject: Telephone Calls, August 8, 29, 1977, GPU Exhibit 465. 9 (Memo dated August 29, 1977, to Plant-10 Design Unit Managers from D. H. Roy, was marked 11 GPU Exhibit 465 for identification, as of this 12 date.) ' I)
\_/
13 Q Is GPU 465 a copy of a memorandum which 14 you sent to all of your unit managers in the plant 15 design section in 1977 to make sure that telephone 16 calls were answered? 17 A Yes, it is. l 18 MR. SELTZER: I don't have any further 10 questions.
.20 MR. FISKE: We have no questions.
21 '(Time noted: 6:21 p.m.)- 22 DONALD HENRY ROY 23 l-_ g 24 Subscribed and sworn to before me f V
- _ 25 this. day of 1982.
1 y_ , , , _ . , , ._- , ~ + - - - -n>, e
i 350 f 3 ! CERTIFICATE 2, STATE OF NEW YORK ) 3 ,; : 33,: : COUNTY OF NEW YORK ) , 4 7 CHARLES SHAPIRO
, a Notary o.
s Public of the State of New York, do hereby 6 l , certify that the continued deposition of f 3 DONALD HENRY ROY was taken before j 8 February 3, 's me on 1982 consisting . 4
,. 9 of pages 171 through 352 ;
i* I further certify that the witness had
, been previously sworn and that the within i
transcript is a true record of said testimony; ! That I am not connected by blood o" -{ 14 i i
, . marriage with any of the said parties nor 10 ;
j interested directly or indirectly in the matter 16 i i in controversy, nor an I in the employ of any I of'the counsel. ; 16 l , l IN WITNESS WHEREOF, I have hereunto set my ' 10 hand this day or FE dA c AG./ i
, log t
l-21 f , L
~ t
) !
23l I
> A 41e4 Charles Shapiro, CS
; 24
, 2.s < ; i t !~ l s : :
l 351
- /'
I ND EX i l WITNESS PAGE Donald Henry Roy, resumed. 173 9 E XH I B I T S I GPU FOR IDENTIFICATION 4 449 Memo from Dr. Roy to G. D. Quale, dated October 8, 1979 181 1 450 Memorandum from D. H. Roy to i Distribution "c," dated August 10, 1979 217 451 Memo from George G. Zipf to J. H. MacMillan, dated May 23, 1979 218 f 452 Memo from D. H. Roy to L. J. 1 Stanek, dated September 11, 1978 232 i' = -453 Memorandum of September 11, 1978,
.! .from D. H. Roy to L. P. King, with attachments 244
, 454 Memo dated May 30, 1978, frca l D. H. Roy to Districution 255 I
4 455 Memo from D. H. Roy to Distribution, 4 dated June 8, 1978, with attachment 256 , 1-456 Memorandum from D. H. Roy to F. R.
'Fahland, dated September 11, 1978, with attachments 264 457 Memo to J. H. Taylor from D. H.
i (%
\ ,/ Roy, dated January 20, 1978, with
' ~
s- attachment 268 l 4 i i
.' .__ .. _ 'r.
352 (A_) INDEX TO EXHIBITS (continued) GPU FOR IDENTIFICATION PAGE
'E 458 Memo from Fahland to D. E. Lee, dated April 15, 1979 313 459 Memorandum from D. E. Lee to D. H. Roy, dated April 15, 1979 313 460 Memo to D. H. Roy from F. W.
Sullivan, dated July 22, 1977, with attachment 322 461 Memo from Liz Brooks to Fred Sullivan, dated August 29, 1977, with attachment 322 462 Document headed " Internal Use () Eyes Only - Performance Engineering Concept" 326 463 Memorandum from D. H. Roy to J. H. MacMillan, dated June 8, 1979 340 464 Letter dated September 21, 1979, to Dr. Don H. Roy from Mr. Snyder, with attadhment 347 465 Memo dated August 29, 1977, to Plant Design Unit Managers from D. H. Roy 349 i
\s' 9}}