ML20028C174
| ML20028C174 | |
| Person / Time | |
|---|---|
| Issue date: | 01/05/1983 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1169, NUDOCS 8301070146 | |
| Download: ML20028C174 (67) | |
Text
w wl m
3-1.
L e
O OFFICIAL TRANSCRIPT
- O NUCLEAR REGULATORY COMMISSION DKT/ CASE NO.
]((
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON GENERIC ITEMS PLACE weenington, D.C.
O DATE aenuary s, 1983 PAGES 1 enru es gfp 8 gyr " ;;
AR o
newmcme (202) 628-9300 440 FIRST STREET, N.W.
WASHINGTON, D.C. 20001
1 O
UNITED ST4TES or AMERICA 2
NUCLEAR REGULATORY COMMISSION 3
ADil?CRI COMMITTEE ON REACTCR SAFEGUARDS 4
SUECOEMITTEE ON GENER1C ITEMS i
5 Room 1046 i
l 1717 H Street, N.W.
6 Washington, D.C.
7 Wednesday, Jan ua ry 5, 1983 i
8 The Subcommittee on Generic Items met, pursuant l
9 to notice, at 8402 a.m.,
Hyer Bender, Chairman, I
10 presiding.
11 W.
l l
13
- O i
15 t
l 16 1
17 18 19 20 21 22 1
23 O
24 25 O
ALDERSON REPORTING COMPANY,INC.
t 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345 m
. ~.
. -.- _ =_ - _
J l
2 i
1 4
Q 1
ACPS MEMnERS PRESENT:
i j
2 j
3 MYER BENDER O
4 WILLIAH KERR 5
DADE W.
MOELLER i
2 i
6 JEREMIAH J. RAY j
7 DAVID A. WARD 8
{
9 DAVID OFRENT C
10 I
11 DESIGNATED FEDERAL EMPLOYEES 12 13 SAM DURAISWAMY
!O 14 4
l 15 ALSO PRESENT:
l 5
16
,l l
17 JAN PRESTON 18 MALCOLM ERNST l
19 WARREN MINNERS t
20 l
21 l
l 22 I
23 l O 24 l
25 l
lO l
l l
ALDERSON REPORTING COMPANY,INC, 400 VmdiNit WE., S.W., WASH 8NGTON, D.C. 20024 (202) 554-2345
3 O
>aocrro1xcS 2
MR. DENDER:
The meeting will now come to 1
3 order.
This is a meeting of the Advisory Committee on 4
Reactor Safeguards Subcommittee on Generic Items.
5 I am Myer Bender, Subcommittee Chairman.
The 6
other members present today are Dr. Moeller, Mr. Ray, 7
Mr. Ward, Dr. Siess, Dr. Ok rent, D r. Kerr.
8 The purpose of this meeting is to discuss the 9
methodology used by the NRC staff in the prioritiza tion 10 of generic safety issues.
We will also discuss a report 11 by Ms. Jan Preseton, ACBS Fellow, on the related matter.
12 This me.eting is being conducted in accordance 13 with the provisiotis of the Federal Advisory Committee l
14 Act and the Government in the Sunshine Act.
15 Mr. Sam Duraiswamy, on my right, is the 16 Designated Federal Employee for the meeting.
17 The rules for participation in today's meeting 18 have been announced as part of the notice of this 19 meeting previously published in the Federal Register on 20 Wednesday, December 22, 1992.
l 21 A transcript of the meeting'is being kept and 1
22 will be made available as stated in the Federal Register 23 Notice.
It is requested that each speaker first
()
24 identif y himself or herself and speak with sufficient 25 clarity and volume so that he or she can be readily O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W WASHINGTON, D C. 20024 (202) 554-2345
4
(}
1 heard.
2 We have received no written statements from 3
members of the public.
()
4 We have received no requests for time to make 5
statenents fron members of the public.
6 Just a brief preamble to what is going to go 7
on this morning.
Hopefully, everybody has received a 8
cooy of this report tha t Jan put together.
Is there 9
anybody who did not get one?
10 I have asked her to discuss the content of it, 11 so I won't try to go through it and offer any 12 praliminary comments.
13 What we should be dealing with this morning is
()
14 whether the Committee can and wants to take a position 15 on this prioritization technique f or generic items.
16 The staff told us when we heard them the last 17 time that the approach was to use this technique as a 18 screening basis for eliminating the things which did not 19 seem important, and to allow them to concentrate their 20 resources whare they would do the most good in a safety 21 context.
22 They have already gone through the exercise of 23 prioritizing - setting the priorities - on a number of
()
24 items and they have been at tached to the report which 25 they provided to us.
ALDERSoN REPORTING COMPANY,INC, 400 VIRGIN!A AVE, S.W, WASHINGTON, D.C. 20024 (202) 554 2345
5
()
1 The Committee, I am sure, will recognize that 2
the list has become very, very long and wha t has 3
happened to it is what many of us were afrnid would
(^V) 4 happen, namely that it has gotten so long that you 5
cannot tell what is important from what is unimportant 6
any more.
So, this idea of assigning priorities may be 7
some way to define che importance of things.
8 What we have to decide is whether the 9
computational method is more usable than just good 10 judgment.
11 I think rather than go any f arther than that 12 right now, I will ask first for comments from the 13 Committee on what it would like to do if it is different O
14 than just hearings from Jan, and if there are no 15 comments, I will ask Jan to pick up.
Any comments?
16 Dade.
17 MR. MOELLEPs One item I am sure you vill be 18 a dd re sing is the time schedule.
19 MR. 2 ENDER:
Yes.
He is talking about the i
20 time schedule for getting things done.
21 MR. MOELLER:
Right, and when they want 22 responses.
When the staff wants responses to what they 23 have done thus far.
()
24 MR. SIESS:
I thought the important time 25 schedule was, this meeting has got to be over by 9 O
ALDERSON REPORTING COMPANY,INC, l
400 VIRGINIA AVE, S.W, WASHINGTON. D.C. 20024 (202) 554-2345
6
()
1 o' clock.
2 (Laughter.)
3 MR. WARD:
That is one point of view.
4 MR. ECELLER:
But in other words, the way I 5
have seen the proposed plan, to divvy up the items for 6
varicus Committee members to review in detail.
It would 7
take me several months at least to try to go through 8
that.
9 MR. BENDER 4 Well, we will have to discuss to tha t, I think, Dade.
It may turn out --
11 MR. SIESS4 On th e methodology, Mike, one 12 thing that occurred to me, it would be a little exercise 13 to see how sensitive this approach is.
14 That is, suppose I went in with a prejudice 15 that something the staff has considered as low and I 16 thought it was high, how little or how much would I have 17 to manipulate the numbers in order to get the answer I 18 wanted.
19 MR. PENDERs I think Jan will say something 20 about that.
I personally think that is probably the 21 central issue.
22 MB. SIESS:
Because I got a feeling when I 23 looked at something with as many terms in it as I see
()
24 here, you know, that you could say about this what 25 somebody said about statistics.
You use statistics like O
ALDERSoN REPORTING COMPANY. INC.
400 VIRGINIA AVE., S.W WASHINGTON. D.C. 20024 (202) 554-2345
7
()
1 a drunk uses a lamp post - for support, not for 2
illumina tio n.
3 I have a feeling I can get just about any 4
answer I want without t ryin g awfully hard.
5 MR. BENDER:
Well, I think anybody that has 6
gone through this exercise, even in a superficial way, 7
comes to that conclusion.
8 What the staff has said - and I think we have 9
to think about it some - is whether this is a good way 10 to screen.
11 Nothino is fixed in stone by this technique.
12 I think they have poin ted out a number of times that we 13 are not using this as a judgment basis that cannot be 14 changed.
That it is a way of deciding what to do with 15 our priorities over the next two yea rs and if we wanted 16 to change our minds after a month or two, there is 17 nothing in' the regulations that prevents it.
18 So, the real question is whether it will lead 19 them in the right direction.
20 MR. SIESS4 You know, Jan said sor.ething in 21 her report about a Delphi approach.
I am not all that i
22 enthusiastic about the anonymity tha t-goes along with 23 the Delphi because I do not think all experts are equal.
()
24 But what it brought to mind was the procedure l
25 thst the staff used several years ago in assigning i
i ALDERSo*t REPORTING COMPANY. INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345 l
8
(}
1 priorities to the research to improve safety.
Remember 2
that?
I forget what NUREG it was.
3 But they went through a process that was O
4 almost completely qualitative and arrived at a list of I 5
think it wa s five or six areas to im prove safety.
And 6
that process, I tought, worked very well.
7 Now, maybe that ic because I agree with the 8
answers that they came out with.
But it was a very 9
qua lita tive process and they dealt with, I would suspect 10 as I recall, only about 20 or 30 candidate itens.
11 I have not gotten much feel for how that 12 process would work on the laundry list they have here, 13 counting all the 0737 items.
But that comes to mind.
14 You know, it is not Delphi but this took into 15 sc ount a lot of things they do - the number of reactors 16 that would be affected.
They would complete the 17 qualitative on the perceived benefits and the perceived 18 costs.
19 HS. PRESTON:
Well really, Dalphi was only a 20 suggestion.
21 XR. SIESS:
Again, your thought was congealing 22 opinion in some way.
23 3R. BENDER:
Why don't we let Jan just give us
()
24 a quick overview of what she did, and then we can try to 25 come to grips with what we want to do about it.
O ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
T 9
()
1 MS. PRESTON:
If all of you really have looked 2
at the report I probably don't even need to use the i
3 viewgraphs, especially since the most important portion 4
of it is just the first three pages.
If there is any 5
problem - referring to staf f - I will go ahead and use 6
them.
7 Mr. Bender asked me to look at the report, 8
specifically at the methodology and some example to see 9
whether I felt like it would do the job well.
10 The staff, as you can see on the first page of 11 the report, has developed a value-impact score which is 12 safety benefit over cost, The variables are the number 13 of reactors affected and the average remaining life of
()
14 the affected reactors, which gives you an idea of the 15 magnitude of the problem.
16 In other words, a problem that would just 17 involve the oldest five reactors would not be near as 18 important as a problem that involved the newest five 19 because the remaining life would be larger.
20 And the other factors, variables in the 21 enumerator of the equation, are accident frequency 22 reduction.
That is, if you solved this safety issue, 23 what is your reduction in accid ent frequency.
Some
()
24 safety issues may affect more than one accident but that 25 is the varia ble th a t is used.
()
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
10
()
1 The other variable in the enumerator, the 2
equation, is public dose from the radioactive material i
3 released from containment for the events associated with f-)s
\\~
4 the safety issue.
5 This variable also includes directly the dose 6
rate due to an implementing of the change -- I am sorry.
7 Just a second.
It involves, directly involves 8
either the --
9 MR KERR:
Population dose?
10 MS. PRESTON:
Well, no it does not involve the 11 population dose.
One of the things was involved 12 directly and one of them was considered but not 13 directly.
OK, occupational doses associated with 14 resolving the safety issue were taken into account but 15 separated, or stated se pa ra t ely.
1S Clean-up was included in this dose variable.
17 And then, the denominator of the equation is 18 the costs of the effort, including NRC and industry 19 costs.
20 MR. BENDER:
These are dolla r costs.
21 MS. PRESTON:
Dollar costs.
22 The staff approach was to use this equation as 23 a screening, and then they were going to adjust it using
()
24 engineering judgment.
25 One of the thines that they felt, like if it O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
11 0
1 es ree111 e ver7 1eree Proh1e - e 1eroe aose vetenti 1, 2
then they would give it a higher priority.
There is a 3
list of things in the report, in the attachment on pace 4
15, the first attschment, that shows the risk thresholds 5
tha t ware used to change the ranking that oriainally 6
came out of this score.
7 And then they say they use other 8
considerations to adjust it.
9 MR. BENDER:
Jan, excuse me.
The costs, 10 dollar costs that they use, are they direct costs or 11 costs that include interest and amortization, and that 12 sort of thing?
13 MS. PRESTON:
They did ure 19 8 2 d ol1a rs.
They O
14 did use future costs and they were broucht into 1982.
15 MR. BENDER:
That could account for inflation 16 but does it account for interest and all those other 17 things?
18 MS. PRESTON:
I did not see any.
19 MR. BENDER:
As you know, the elapsed time may I
20 be a very important aspect of the cost.
21 MS. PRESTON:
They did consider whether 22 elapsed time would increase the severity of the issue.
23 But I did not see th at ela psed time was figured into the O
24 cost.
25 MR. BENDERS Well, the discount rate only ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WAS.-ilNGTON, D.C. 20024 (202) 554 2345
12
({}
1 accounts for at most inflation.
2 MS. PRESTON:
But the cost numbers included 3
NRC costs in finding a solution or trying to develop O
4 resolution and regulation, and also then the 5
implementation by the utilities of the solution.
6 MR. PENDER:
And the costs include staff time.
7 MS. PRESTON:
Yes, based on about a 8
hundred-thousand per staf f year, which is typical.
9 MR. BENDEP And they also include the labor 10 of the licensee?
11 MS. PRESTON:
Correct.
12 MR. SIESS:
If the assessment was that 13 research might be needed to resolve it, was that 14 included in the cost?
15 MS. PRESTON:
Pardon?
16 MR. SIESS:
Was research included in the cost?
17 MS. PR ESTO N:
Besearch, NRC-projected work was 18 included.
Let me get down to the cost.
i 19 I had a lot of problems with the cost variable l
l 20 and the reason was, it seemed to me that some sort of a 21 resolution had to be identified before you really could 22 apply this.
l 23 There was just no cost figured in of, "Well, l ()
24 we don't know what the answer is going to be, so we 25 figure it is going to take us two years of NRC research l
ALDERSoN REPORTING COMPANY,INC, 490 VIRGINIA AVE., S.W, WASHINGTON, D.C. 20024 (202) 554-2345
13 1
()
1 a t a level of such-and-such to find an answer." Anybody 2
who has been involved in research knows it is very 3
difficult to project how much research it will take to 4
find an answer.
5 TR. SIESS:
Or if there is an answer.
6 MS. PRESTON:
Or if there is an answer.
That 7
is one of the things that concerned me very much about 8
the cost.
I think in my report I called it a " chicken 9
or the eng" problem.
10 They are asking the people using the formula 11 to know the answer before they decide whether t,he 12 problem is worth solving.
I think we get into that 13 problem quite a lot.
14 Some other problems I had with the equation i
15 were that there is a lot of averaging that goes into it, I
16 dose rates.
If you are developing an equation it should 17 be used over the whole affected plants, all of the l
18 affected plants.
19 You are going to have a lot of variances 20 between BWRs, PWFs.
Take turbine missiles.
That is 21 ver y highly dependant on whether you have vital things 22 close to your turbine.
If you have your control room 23 richt next to your turbine then that is a lot more
()
24 serious than if it is separate.
25 So, I felt that thir equation fell down, the
(
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
14
()
1 basic equation fell down if you had a problem that had 2
extreme variance fon one plant to the next.
3 In the analysis they do take that into 4
consideration sometimes, but I was not always convinced 5
tha t it was done adequately, that things were combined 6
to reflect
-- in fact, none of the things that I am 7
saying here is anything new.
8 In the document the staff notes all of these 9
uncertainties.
What they are doing is, they a re sa yi ng,
10 "Well, we ar e going to take a shot with this equation to 11 screen and then we are going to use our engineering 12 judgment.
We know that there are uncertain ties.
We 13 know that there are differences in plants.
What we are
(
14 going to do then, once we have it, is juggle things 15 around when we think they are right."
16 Well, my major problem with this is, if you 17 are going to use engineering judgment to modify an 18 equation that has basically been derived by making an 19 awful lot of sssumptions, then you a re lending sort of a 20 pseudo-scientific basis to the result.
21 I have a lot more confidence in structured 22 engineerins judgment techniques such as Delphi or -- I 23 am not familiar with the methodology that D r. Siess was
()
24 referrino to that has been used by the NRC in the past.
25 If you sre using an equation that purports to O
ALDERSON REPoRTINu COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
15
()
1 be quantitative and then you adjust the num bers, then 2
sometimes biares or proconceivad notions can get 7
3 filtered in.
And it is not really above board.
4 It is not something that is devious, it is 5
just a natural tendency that if you believe that 6
something is important, when you come up with a score 7
that is low then you say, "Well, what can I think of to 8
make this thing have a higher score?"
9 The report says a t all points, you have 10 uncertainties.
We are using engineering judgment.
But 11 they still relate the ranking, the high, medium, low and 12 drop, to the Com. mission's safety goal which are hard 13 numbers.
Maybe that would be a good viewgraph to put up.
O
\\'
14 This figure comes in after you talk about, 15 after they talk about assigning the score based on this 16 equation.
Then th e y talk about other considerations and 17 en71neering judgment.
But this still stands out.
It is 18 not really, strictly stated, the initial score is not 19 nec essa rily what the final priority or ranking of high, t
20 low, medium still reflects.
i 21 I felt lik e the equation would be a good 22 reprecentation in severs 1 casec if you had data.
We do 23 not have good dose data.
We do not have good accident l ()
24 f requency based on a specific safety issue data.
25 And unless you really have good data in the l
($)
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGioN, D.C. 20024 (202) 554-2345
16
()
1 first place, I feel like an equation that sort of tends 2
to relate things to numbers in the final ranking is 3
dangerous in some cases.
O 4
I am not really sure that you can develop a 5
methodology that you could feed into a computer and get 6
a scientific, a quantitative ranking.
7 But I have confidence thst they could take the 8
data tha t they have gathered and, using sene structurerd 9
consensus judgment technique, come up with a very 10 reasonable priority ranking - as Dr. Siess says that 11 they have done in the past for other items.
12 I looked at two or three secific rankings that 13 they did and I think that my uncertainties and my 14 discomfort was borne out.
On seismic scram, for 15 example, the initial assumptions they make are, you 16 cannot benefit any way from a seismic scram.
And in 17 fact, you are going to introduce spurious scrams so you 18 get a negative score.
19 Well, why even go through the exercise of 20 developiong a score if your initial assumption is that 21 it is not goino to help you?
22 I have talked to several people about seismic 23 scrams and there do seem to be a lot of good reasons why
()
24 you might not want a seismic scram.
25 But none of those cocd reasons that I heard, O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
17
()
1 such as non-de tectibility, you cannot detect one early 2
enough to really do you any good.
And things like, 3
well, the plant has a certair. momentum and really, if it Os 4
is not going to trip because of the ea rthquake it is 5
safer to jut let it keep going because if you trip it 6
you have introduced a transient.
7 None of those things were listed in the 8
assumptions, the initial asst 5Mption that they made.
9 They basically sai.1 tha t it cannot he]p you because to having a plant having had a longer operating time does 11 not save you a nything.
12 So, I was very worried about the way it could 7
13 be applied.
I was sort of concerned, as Dr. Siess was, 14 that you could go in with a preconceived notion and come 15 up with pretty much whatever answer you want.
16 MR. BENDER:
You have looked at the turbine 17 missile question?
18 MS. PRESTON:
Yes.
t 19 MR. BENDER:
Did that display anything 20 dif ferent in character f rom a seismic scram ?
21 MS. PRESTON:
Well, they were able to come up 22 with some good frequency numbers, not necessarily 23 nuclear but they came up with tur.bine missile
! ()
24 frequencies because there have been a lot of turbines 25 op e ra ti ng f or a long time.
l l
l ALDERSON REPORTING COMPANY,INC, l
400 VIRGINIA AVE., S.W., WASHINGTON, D C. 20024 (202) 554-2345
18
()
1 This is one of the problems that I had with 2
the averaging problem.
Since turbine missiles is such a 3
highly individual, design-specific problem, I was not 4
comfortable with throwing all the BWRs in one pile and 5
all the PWRs in one pile.
That was one of the problems 6
I had with it.
7 MR. WARD.
Let me just comment, Jan.
You said 8
they came up with good numbers on tne f requency of 9
turbine failures.
10 MS. PRESTON They had some --
11 MR. WARD:
Wait a minute.
But the frequency 12 of turbine failure leading to the missile large enough 13 to penetrate the containment and then damage some safety 14 system which could lead to a serious reactor accident, 15 that is just pure guess.
l 16 MS. PRESTONs Right.
l 17 MR. WARD:
I think they say it is only about l
l 18 ten to the minus fifth - well, it may be ten to the 19 minus second; maybe it is ten to the minus eighth.
20 I think that gets back to the point that the 21 structure is needed and very useful.
But I think 22 p re tending that this equation is telling you something 23 with the sort of input you have to have gives an
()
24 illusion of P. ore detail to the structure than really 25 exists.
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE, S.W., WASHINGTON D.C. 20024 (202) 554-2345
19
()
1 MR. SIESS:
It is very helpful if it supports 2
wha t you think.
3 XR. WARDS Yes.
You know, Y can see a
- S V
4 structure in even using this sort of equation to help 5
define the limits on hins you might have in a structure 6
and to use it for' particualr cases where you can 7
generate credible numbers as an aid to judgment.
8 But I do not think it is particularly useful 9
to subject all of these generic issues to this detailed 10 analysis.
11 MS. PRESTON:
I handed out today another 12 analysis which is on the human factors issue simply 13 because the two that I had attached were technical and 14 you would feel like, well, you might be able to make 15 some reasonable assumptions on technical issues.
16 Vell, the assumptions made and the numbers 17 gotten for the human factors analysis are even less 13 based on reality.
I know because I was trying to come 19 up with human fsetors prioritizations myself at one time.
20 As Dr. Ward says, if you have an equation that 21 you can base on good numbers it is good as a screening 22 before you start applyinc your engineering judgment.
23 But when you are having to use engineering
()
24 judgment to coca up with the numbers f o r your screening 25 and then use core engineering judgments to marsage it, (A/
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
20
(])
1 and then try to correla te it to the Commisson's safety 2
goal, I am just not confortable with it.
3 MR. SIESS:
What priority do they end up with O,
4 on turbine misciler?
5 MS. PRESTON:
Drop.
6 MR. SIESS:
Drop.
You know, seismic scrams 7
are lousy examples.
That is not a technical issue, that 8
is an emotional issue.
9 (Laughter.)
10 MR. EENDER:
Well, maybe turbine missiles are, 11 too.
I don't know.
12 MR. SIESSa I requested they do research on 13 seiemic scram, but most of the research was on how to do
)
14 it and not wnether it made any difference, you know.
15 There are all sorts of differences, depending 16 at what level you set the scram.
Half the plants are 17 coing to scram on earthquake anyway because there is a j
18 turbine vibration that causes the scram, rather than the 19 alarm.
And the other half it just alarms.
20 You know, you can get any answer you want on 21 it.
22 Incidentally, correction.
On NUREG-0438, the 23 pla n for research to improve the saf ety of light-water
()
24 reactors, as I can tell, they screen about 200 item s.
t 25 The items having various decress of spercificity, l O ALDERSON REPORTING COMPANY,INC, 4:10 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345 m
21
()
1 canging from one lump to some rather specific things.
2 They came out with five top items and another ten for 3
further review.
4 We agreed wi th the five items, which in itself 5
is remarkable.
That was almost completely qualitative.
6 MR. BENDER 4 About half, if you look a t what 7
people are making noises about and you decide that the 8
important judgments are based on those people that have 9
the loudest voices, then you can come up with an answer 10 that matches the loud voices.
11 And when you look at this set of numbers up 12 here that Jan has put up, I would have to conclude that 13 corrections to deal with TMI might not make the grade 1
k/
14 unless I were wlling to concede that the containments 15 were vey unreliabis.
16 The numbers themselves are troublesome as a 17 basis for making judgments.
(
18 MR. SIESS Well, there is nothing like l
19 numbers to give you a basis for an argument, you see.
20 MR. BENDER:
Yes, tha t is righ t.
l 21 MR. CIEES:
That is the trouble with a 22 quantitative spproach that it is easy to argue wi th.
23 ME. BENDER:
But there is only two orders of
()
24 magnitude bevaen high and low.
And the uncertainties in 25 the probabilities are, I am sure, a great deal higher ALDERSON REPORTING COMPANY,INC, 400 VIRGINTA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
22
()
I than that range.
2 So, that kind of judoment well, it just 3
lea ves me cold, I guess.
4 MR. POELLEPs I find somethiung intersting 5
here, too.
Appendix I for routine releases says that if 6
by spending a thousand dollars you can save a 7
person-rem, you must do it - which is in the median 8
category, a thousand person-rem per million dollars 9
savings.
10 The safety goal says, if by spending a 11 thousand dollars you can save a person-rem for an 12 accidental release, you must do it.
13 Both of those are mandatory - or they could 14 be.
The one is and the other could be.
And yet, they 15 only rank medium here.
16 I realize I am comparing a little bit of 17 spples and oranges but I find it interesting.
18 MS. PRESTON:
Well, I believe that they did 19 recommend that medium things be fixed.
20 MR. MOELLER:
OK.
21 MS. PRESTON:
But high thines were those 22 things that should be done righ t a way because they have 23 a high potential for problems.
()
24 MR. SIESS:
The distinction between high and 25 medium is a very fuzzy one, neither one is going to be ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
23 O
1 ouehed ofi the 11st.
2 MR. BENDE94 A lo t has to do with how f ast 3
something is gcing to be fixed.
If it is going to take 4
20 years to fix it then the pay-off is going to be small 5
because by the time you get it done 8
MR. SIESS:
You mean 20 years to physically do 7
it, or 20 years to decide to do it?
8 MR. BENDER 4 Well, some of both.
But I just 9
mix it up.
10 MS. PRESTON Well, that would go in the 11 denominator which, if you say it ir going to take you 20 12 yea rs a t a hund red-thousand per staff year, then you 13 would not likely get in the high category.
14 MR. BENDER:
Well, let's take a favorite 15 cubject, turbine orientation.
If it is not oriented in 16 a certain direction, nothing you can do to correct that 17 o rien ta tion will help.
18 MR. SIESS:
No.
You can put in barriers.
19 MR. BENDER 4 You can put in barriers, but 20 whether it will help or not, I think, is a matter of
(
21 j udgmen t.
22 But on a new plant or turbine orientation 23 now-a-days I think you could not tell the difference O
24 detween one orientation and another.
25 I do not know how you deal with something like i
i l
ALDERSoN REPORTING COMPANY,lNC, 400 VIRGINTA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
24
(}
1 that up there.
But if you took their judgment you would 2
say, "Well, all the fuss about turbine orientation was 3
for naught."
I think I would have to question that O
4 argument.
5 MR. SIESS:
I would not.
I am not so sure 6
that turning the turbine around -- I know it reduces the 7
probability.
But the original analysis had a 8
probability of one that if you penetra ted the 9
containa,ent you caused -- I think we talked about LOCAs to in these days; about core melts I don't know.
11 That figure is of f by anywhere from one to 12 three orders of magnitude, in my opinion.
Peneration 13 figures were within one order of magnitude.
We got much
()
14 better data on penetration.
15 MR. BENDER 4 I will not debate with you that 16 point, I fully agree with it.
The real point is that if 17 it does not cost any more to do it one way than another l
18 which I believe to be the case --
19 MR. SIESS:
But we do not know whether the l
20 cost was mainly longer steam lines and some cost of 21 plant a rran gem en t.
Nobody ever evaluated whether longer 22 staam lines made a steam line break accident more 23 probable, and that is a transient that we now worry
()
24 about.
We used to not worry about them, things of that 25 sort.
O l
ALOFit3ON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
s 25
()
1 But I am not so sure that it was ever a major 2
issue except it looked like it had an easy fix.
So we 3
did it.
It might have cone out on that basis.
If you 4
say the cost of it was negligible and the gain, even S
though it eent from ten to the minus seven to ten to the 6
minus nine was a gain, then it would probably have come
)
7 out higher than it did.
8 MR. BENDERS Well, at the time when it was 9
discussed it was very costly.
Now tha t we ha ve designs 10 tha t are based on peninsular arrangements it looks like 11 trying to back away from it would be very diff,1 cult.
12 5R. SIESS:
But it is low now or drop, is it 13 not?
O N/
14 MS. PRESTON:
Yes.
15 MR. PENDER:
I guess I do not understand what 16 it means by " drop it."
If it means not puttino barriers 17 in, I think that is one thing.
18 MS. PRESTON:
One of the things that had been 19 suggested to me when I was talking to people was not 20 just havin7 a fraquency, accident frequency reduction 21 but having a reduction in frecuency over the original 22 frequency, which shows you that you really have improved 23 something.
()
24 Which is why if you have a very low frequency 25 but high consequence you nicht treat it a little bit OO ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
s N
26 T
()
1 more importantly.
2 YR. 3 ENDEP What was the fix they postulated ?
4
'3 MS. PRESTONs It was to have a -- you mean for O
44 tha f re quenc y?
5 MR". BENDER:
No, for the existing plante that 6
made them stop paying attention to the problem.
7 MR. SIESS:
The fix has always been missile 8
barriers that has been talked about.
b've ry time you 9
talk about the terrible missiles.
10 MS. PRESTON4 Improved disc materials.
11 Basically it was try'ing t'o spend some more money 12 deciding wha t a better frequency was.
13 MR. SIESS:
Oh.
(
14 MS.'?3ESTON:
Whether improved disc materials 15' would h'elp, again, determining actual rejection 16 probabilities and energies, and the lik elihood of damage.
17 Whether routine in-service inspections would l
18 help, turbine orientation.
And it was just assuned that 19 it would take NRC a certain amount.
20 ER. SIESS:
Over-speed protection must be in 21 there comewhere, that is one thing.
22 MR. : ENDER:
You know, it is strange when you 23 look at this and it says turbine missiles goes away but
()
24 for turbine disc cracking is a generic problem.
25 If turbine missiles are not a problem, who ALDERSoN REFoRTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
27
()
1 cares about turbine disc cracking?
Sc, there are lots 2
of anomalies in the discussion you have to think about.
3 Is turbine missile going away because we are 4
showing that turbine disc cracking will not lead to 5
turbino rupture beca use of this high-class inspection 6
program se have going on?
7 MR. SIESS4 Well, the approach the staff has 8
been taking on the SEP plants-and some of the others is 9
looking at over-speed protection and in-service 10 inspection on turbine discs to control the probability 11 of a disc failure.
12 Now, we hsve a lot more information on what is 13 to go outside the casing from the upward test.
We have Ci 14 a lot better informa tion on penetration.
But the 15 current approach has been to reduce the probability of a 16 disc f ailure by reducing the probability of an 17 over-speed, a normal speed failure, and inspection.
18 But barriers have always been. considered in 19 the re a nd I do not know why they are not included.
All 20 the other things are protecting the investment, the 21 barrier will protect the public.
Now, maybe it is not 22 practical, I don't know.
23 T3, PRESTON:
I guess my point is that just by (N
/' /
24 looking at this example, the seismic scram example, you 25 can see that it ic very difficult to have all the O
ALDERSoN REPORTING COMPANY,.NC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
28
(])
1 information to really determine a priority.
You take 2
the best information you got and hopefully all the 3
in f orma tion you got and decide whether bef ore you a pply O
4 the formula or after how important you think it is.
5 What concerns me is that even though there are 6
statements to the effect of, "This table in here is 7
based on just the score and then it is adjusted," one 8
might look back at the report and go, "Well, the final 9
score was drop, and that means that these numbers still 10 apply."
11 That is what concerns me is, once you 12 initially assign a numerical value to the different 13 rankings and then you play around with it, you don't O
s/
14 have a quantitative basis any more.
I am not really 15 sure you ever could, there are too many uncertainties.
16 MR. "0ELLEEa Back to one of your comments, 17 Jan, about knowing the solution, you know, before you 18 start.
19 I think*if would help immensely if the staff 20 could tell us in each case, list the full range of i
21 solutions apd then say, "We did not evaluate these for 22 the following reasons.
We did evaluate these."
23 For example, I took air cleaning which was the
()
24 first item, and we were talking earlier about changing 25 the numbers.
Well, se changed those numbers every way O
ALDERSoN REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W, WASHINGTON. D.C. 20024 (202) 554-2345
29
()
1 we could and we could not make that a medium or a high 2
rank.
It was in the drop zone.
3 However, the only corrective action they 4
considered was improved raintenance and testing.
They 5
did not consider any research or; you know, installing 6
new "somethings or others."
So, you a re limited.
7 If you only look at maintenance and testing as 8
the only solution to the problem then, yes indeed, it is 9
drop.
But that says just drop maintenance and testing 10 as the prime methods for correcting this problem.
11 MR, SIE23:
Was it drop because it did not 12 have enough benefit?
13 MR. MOELLER:
It did not have enough benefit.
14 ER. SIESS:
Not because it cost too much.
15 MR. WARD:
Was it the ratio?
16 MR. F0ELLER4 It was thes ratio.
17 MR. SIESS:
Well, but you see,-looking at all 18 the solutions is not the answer because from the 19 standpoint of the cost the cheapest solution is the one 20 you want to look at.
From the standpoint of the 21 consequences it is the most effective solution you want.
22 ME. WAPD4 Was not Dade suggesting that 23 research might lead the way to a very cheap solution?
()
24 MR. MOELLER.
Richt.
25 MS. PRESTON:
And ;4t, the research is O
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
)
1 30
()
1 expensive, too.
2 MR. SIESS:
When you do th a t, the re is another 3
factor that has to be brought in.
You have to try to O
I 4
erstimate the cost of the re sea rch.
Then you have to 5
make some kind of an estimate of the probability that 6
the research will come up with an r.o. s w e r.
7 Now, EPRI just went through, had somebody go 8
through a cost-benefit approach', value-impact approach 9
prioritizing research.
10 They got into this question of, how do you 11 assign the probability of success to the various stages 12 of rersearch?
Of course, the whole thing blev up when 13 they did tha t because nobody knew how to do that.
14 MR. BENDER:
We need to try to come to some 15 position on this thing.
If you go through the list 16 which I did
, and I suspect some others ha ve done at 17 least in a casual way - the most obvious thing is that
'18 most of the issues tha t we had thought were important 19 either come to the category of unresolved safety issues 20 or TMI Action Plan, or they are to be prioritized later.
21 MR. SIESS:
Are these only the NRR issues?
22 MR. BENDER:
Yes, tha t is richt, it is only 23 NRR issues.
But that is about 85 percent of the generic
' ()
24 issues that have to be dealt with, I guess.
You know, 25 it is a number out of the air.
ALDERSoN PEPoRTING COMPANY. INC.
400 VIRG;NiA AVE, S.W., WASHINGTON, D.C. 20024 (202) 554 2345
i 31 h
I XR. SIESS:
Let's see, the TMI issues were 2
just postponed, is thtt it?
j 3
MR. EENDER:
Well, they are being worked on, i
4 and they have already been 5
MS. PRESTON:
Mandated.
8 j
MR. BENDER:
-- legally mandated.
So, putting 7
them into this list does not make much diff erence.
8 MR. SIESS:
We are not going to_second-guess a 9
legal mandate.
1 j
10 11 i
12 1
13 O
1.
l I
15 f
18 17 l
18 19
(
20 21 22 23 0
24 i
25
- o o
ALDERSON REPORTING COMPANY,INC, 400 VIRGINTA AVE., S.W., WASHINGTON D.C. 20024 (202) 554-2345
32
()
1 MR. BE3 DER:
Once the lawyers make up their 2
mind nothing can change it, even facts.
But that seems 3
to be just the way in which the staff operates.
O 4
MR. SIESS:
But you see, this is a factor that 5
I do not know whether staff took into account.
Somebody 6
has said tha t cur job is not just to protect' people from 7
things that will hurt tham, but to protect people from 8
things that scare them.
And an awful lot of the TMI 9
items were an emotionally-perceived fear of public 10 reaction rather than a real one.
11 I do not kow whether that is being factored 12 into some of the other items.
13 MR. WARD:
Well, you know, I think there is O
(J 14 some political reality there.
But I don't take part in 15 sort of institutionalizing superstition and ignorance, 16 to put it bluntly.
17 MR. BENDER:
Let me put a few words into the 18 Committee's mouth on its position.
We do not think the 19 numerical prioritization approach is a very good or 20 useful one.
21 MR. SIESS:
If it gives an answer I agree 22 with, I would say it is useful.
23 (Laughter.)
()
24 R. EENDER:
That is a subjective judgment 25 that I agree with.
Ab ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
33
()
1 MR, WARD:
I would say a structured approach 2
is usaful.
But this is more detailed than the data 3
suggested.
4 MR. KEBR:
Why don't we suggest that this 5
approach is better than any we know of and it is 6
probably OK if we don't tak e it too seriously?
7 MR. SIESS:
Refresh my memory.
I think there 6
is something in there that said tha t after they did all 9
the numerical, then they applied judgment.
10 MR. MOELLER:
Yes.
11 MR. SIESEs Now, what I cannot ge t a feel fer 12 is how much of the decision is judgment and how much is 13 numerical.
If a fair proportion of it is judgment and 14 it is good judgment, then I don't really care too much 15 about the numericsl.
16 MR. BENDER:
Well, let's see if we can get 17 some reaction from the staff.
18
'R.
ERNST:
Halcolm Ernst, NRR.
I have a few 19 points of clarification and I will try to answer your 20 questions, also.
21 As far as TMI Action Pla n items a re concer:ted,
22 I think we are only prioritizing those which really are 23 mandated at the present time.
So, there is a chance to
()
24 change one's mind if one wa nts to.
25 Also, saf ety goal is not mandated.
In fact, O
~/
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., *.W., WASHINGTON, D.C. 20024 (2C2) 554-2345
34
()
I we do not even really know what the safety goal is.
We 2
used the proposed safety goal as sort of a benchmark.
3 It is not mandated.
4 While the chart up there indicates a matrix 5
that really is cost and benefit, I would wa n t to clarify 6
that was only one pa ra m e te r.
The other was the pure l
7 safety benefit in the matrix.
So, it is net just cost 8
ben efit, it is also safety benefit.
9 MR KERR:
Excuse me, what do you mean by to saf ety benefit?
11 YR. ERNST:
Well, in the bottom of the matrix 12
- the matric is man-rem versus dollars.
But also to 13 qualify as a high, medium or low, one can just use the 14 bottom one, namely the pure reduction in core melt or 15 the pure reduction in man-rem, irrespective of dollars 16 as a second way of categorizing it.
17 MS. PRESTON:
let me put that one up.
18 MR KERR:
How do you relate core melt 19 frequency to, say 20 MR. SIESS:
Jan, this would be useful.
21 MS. PRESTCN:
Yes, I don 't have that among the 1
22 vie wa ra phs.
23 MR. EPNST:
Yes, that is the one that would be
()
24 more helpful.
25 ES. PRESTON:
It is in the attachment page 12,
(~T U
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
35
()
1 Attachment A.
Page 13, I am sorry.
2 MR. ERNST:
In fact, as a matter of interest I 3
think in most cases the high or medium, or low ranking 4
was principally ordained by the safety benefit and not 5
so much by the cost benefit.
That is my recollection 6
and I may well be wrong.
7 MR. BENDERS What is the diff erence, again, 8
between safety benefit --
9 MR. SIESS:
One is a benefit and the other is 10 an absolute, Mike.
11 MR. BENDER:
What?
12 MR. SIESS:
One o f them is a ratio, it is the 13 value-impact score, the benefit provided by cost, and 14 the other one is essentially the enumerator without the 15 16 MR. ERNST:
I don't have the matrix in front 17 of me.
18 MR. EENDER:
I think you are missing the point 19 of the question.
I understand the mechanics of getting 20 the value.
I think it has to do with the question of, 21 is there really an understandable difference.
22 We are trying to protect people against 23 radiation exposure and whether you do it on the basis of
()
24 dollars per unit of exposure or just rate of exposure 25 does not seen to be more than how you are establishing i
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
36
(])
1 the significance of the numbers.
2 MR. ERNST:
I think my point was only one of a
clarification.
If we are hung up on the kinds of fixes O
4 and the costs of fixes, then in most cases I am saying 5
tha t was not a determinant factor because in most cases 6
you got your ranking because the reduction in core-melt 7
frequency and the reduction --
8 (Simultaneous conversation.)
9 MR. ERNST:
Certainly, most of your highs and 10 mediums would have wound up being high or medium 11 regardless of the cost, regardless of the assumption on 12 the fix.
13 MR. BENDER:
Well, unfortunately the
(
14 p ra cticality of things is lost in that kind of 15 judgment.
Cn the one side you said, "Well, we are not 16 going to look at things and select them just because 17 they do not cost very much."
You made that argument on 18 one side.
l 19 But on the other side you are not saying, "We l
20 are going to eliainate things because the cost of 21 correcting them is so exhorbitsnt that it does not make l
22 any sense to fix it, even if it were an important item."
23 MR. ERNST Yes, I think this is a good one to l ()
24 show the plot.
It shows that our judgment on safety 25 issues is pretty well borne out by the analysis and that O
V ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
~y
q-e
37
( ))
1 is that it is, if it ic a high-risk item more than 2
lik ely you can fix it.
I think that is what the study 3
shows.
-)
r) 4 I! R. S E.v D ER :
I wish I thought I understood 5
that well enough to agaree with you, but I don't.
6 MR. MINNERSs Let me try again because I think 7
it is being misrepresentet wha t we did.
8 Jan talked just about a cost-benefit ratio -
9 what we call a value-impact - which is the safety 10 benefit which it the value, end the impact is the cost, 11 NRC cost.
That is the only one that she ta lk,e d about.
12 But the other part of the matrix is the l
13 absolute safety benefit - whether it is measured in
/']
l
\\/
14 man-rem or core-melt f requency which one can generally l
15 use.
So, you have to look at both of thoce factors.
l l
16 If you are above a certain sbsolute safety l
l 17 benefit it gets a high priority.
If you are below a 18 certain safety benefit it gets a low priority.
And in l
l 19 between is where the value-impact score changes your 20 priority.
l 21 If you have a high impact, value-impact score 22 in this intermediate total risk, then you get a high; 23 and vice-versa, you can get a low.
So, you cannot just l ()
24 look at one factor.
That gives you the wrong answer.
i i
25 MR. BENDER:
Well, the problem really has to l
l O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W WASHINGTON. D.C. 20024 (202) 554 2345
38
()
1 do wi th whether you believe the numbers at the right 2
hand line because that will determine whether you are 3
going to have to pay any attention to cost or not.
\\
4 MR. MINNERS:
If you believe -- I am sorry, j
5 you said the right-hand line?
j 6
Mk. BENDER:
I am talking about to my richt.
7 MR. MINNERS:
Well, your question was up l
6 here.
We do not have any issues which have very large 9
costs and la rge risk s.
There is nothing in this here.
10 MR. BENDER:
You don't have anything in this I
11 risk.
12 MB, MINNERS:
We don't have any generic issues 13 that are currently proposed which have a high risk and a
}
14 high cost.
l 15 MR. ERNST:
The part there, I think, is 18 showing the value impact against the safety risk 17 reduction, and I think the plot shows that generally in 18 areas where you have a f airly high safety, important l
19 safety problem, you have reasonably economical fixes.
20 MR. SIESS:
If you ignored costs, all those 21 heavy lines would be vertical lines.
22 MR. ERNST:
That's correct.
23 MR. SIESS:
All you have done with your cost
()
24 figure is jogged those lines over at certain spots, and 25 you jogged them over one square.
O ALDERSON REPORTING COMPANY,INC, 400 VIRCiNIA AVE., S.W WASHINGTON, D.C. 20024 (202) 554 2345
39
()
1 MR. MINNEES:
Yes, that's right.
2 MR. ERNST:
That is exactly right.
3 MR. WARD:
Warren, could you tell us a little 4
bit about the points you ave on there?
5 MR. MINNERS:
These are a plot of all of the 6
calculated value-impact scores and risk of all the 7
issues tha t a re in the report that you have, the current 8
version.
9 If there are some issues that we said that 10 wore resolved, or licensing improvement or 11 environmental, they are not on here.
But of the safety 12 issues that were analyzed and a value-impac t score was 13 calculated and necessarily a risk number was calculated,
)
14 that is part of it here.
i 15 Just to how you how things f,all.
16 MR. SIESS:
The vertical spread is due to 17 bringing in the cost.
18 MR. MINNERS:
And this is the cost-benefit 19 ratio.
If you have a high cost-benefit ratio you get a 20 lot of safety for your buck, if we made it high.
If you 21 don't get much safety for your buck, we made it a drop 22 or a low.
23 MR. SIESS:
But if you did not include costs,
()
24 that would essentially be a single flagonal line.
25 MR. MINNERS:
That would be essentially a
()
I
/
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON D.C. 20024 (202) 554-2345
40
()
1 single diagonal line and then you just take the jogs out 2
here, I guess.
3 ER. SIESS:
Right.
O 4
MR. BENDER:
Jan, you had a question?
5 MS. PRESTON:
The problem is that you are 6
saying that, "Well, we figured in the thir,qs that were 7
high risk also could overshadow the fact they are a bad 8
value-impact score."
9 My point was that you do not really have any 10 good data to tell you the change in risk any more than 11 you have for the value impact.
12 The main difference between thos? is the cost 13 data which we all admit you havo to know the solution or 14 there are a lot of uncertainties.
But there are an 15 awful lot of uncertainties that I was not comfortable 16 with in the change in risk.
17 So, I am not really sure tha t -- my main point 18 was that I don 't really think you have enough data to 19 come up with a numerical score.
20 MR. SIESS:
All those little dots ought to be 21 g ra y circle s.
22 MR. MINNERS:
That's right.
I agree with your 23 comment tha t ycu should not take these too seriously.
< ()
24 You have to be very skeptical of these numbers.
25 But I would be a lot more skeptical of a bunch O
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
41
()
1 of guesses by people sitting around a roon and it is 2
glorified by calling it the "Delphi" method.
It is not 3
a systematic 7.ethod.
You have no. set of criteria by 4
which everybody is judging it the same way.
You do not 5
have the detailed analysis which you can criticize.
6 Ea ch one of these issues on the turbine 7
missile, you can go in and you can look at each one of 8
our componen ts of the numbers and agree or diragree and 9
pick a different number.
You cannot do that with the 10 Delphi method.
11 MR KERR Did the staff find this approach 12 helpful?
13 MR. MINNEFSa I found it helpful.
14 MR KERR And did it give different results
(
15 than one would have gotten previously, in your view?
16 MR. FINNERS:
Yes.
17 MR KERR:
Did it illuminate some things?
18 MR. FINNERS:
Yes, and I had several instances 19 of that, and I told the analyst -- he came up and he had 20 calculated a low score and he thought it was a high 21 issue.
I said, " Fine.
Your judgment is wrong or the 22 numbers are wrong, go fix one of them."
23 He went back and he looked ha rder, and he i ()
24 found out if you are digging through the lER file and l
. et cetera, he found 25 talking to the vendor, et cetera, ALDERSoN REPORTING COMPANY, INC, 400 VIRGINI# AVE., S.W., WASHINGTON, C C. 20024 (202) 554-2345
42
()
1 out that the frequency of the event was a lot higher 2
than be had first estimated and it was justified as ligt,..
3
~
n I think that is the way you oucht to do it.
V 4
If your judgment does not five with the numbers, you 5
have to fix one or the other.
6 3R. SIESS:
But that is one man's judgment.
7 MR. MINNERS4 No, it is not one man's 8
judgment.
If we are doing our job right and people take 9
these seriously in the peer review precess - which I 10 don't think they are doinc - then you should get 11 comments from people who would say, " Hey, this is a dumb 12 number and you guys don't know what you are talking 13 about."
14 I agree, you ough t to have more than one 15 person doing it, and that is a procedural problem.
It 16 is difficult to get people to pay attention to these 17 rankings.
18 Now that I think we have some credibility that 19 resources are going to be alleca ted on the basis of 20 these, we seem to have a little more interest in people 21 giving substantial criticisms.
22 MR. SIESS:
Well, now, let's go back to the 23 research to improve safety where a purely qualitative -
r(
24 not Delphi certainly - but they had a review panel that 25 had about 30 people on it, representing a fairly wide O
ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
43
()
1 spectrum of both knowledge and inte res;t and, I will add, 2
inprejudica.
3 Do you think they came up with bad answers?
/
4 MR. MINNERS4 Based on the research program we 5
are doing now, I think we are doing a lot of dumb thinos.
6 MR. SIESSs All right, but do you happen to 7
remember what the fivo items were?
8 Let me just read them to you.
Alternate 9
containment concept - still a pretty hot subject, 10 e sp ecially after T I.
11 Alternste decay heat removal concept - even 12 hotter after we look at transients to shut down heat 13 removal systems and aux. feedwater.
14 Alternate emercency cooling concepts - that is 15 the cid LOCA problem which has not gone awa y yet.
16 Improved in-plant accident response - that is 17 close to the top of our post-TMI list.
18 MR. MINNERSs I think the one on ECCS.
19 MR. SIESS:
In that time range it was pretty 20 darned hot.
21 Advanced seismic design.
We have different 22 opinions but it is s til l a pretty hot subject.
23 Those are all items, that is pre-TMI.
O) 24 MR. MI:iNERS:
Let me comment on that.
g_
25 MR. SIESS:
TMI made some of them look pretty O
ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (201; 554 2345
un
()
1 damned good.
2 MR. MINNERS:
What that research board does 3
- r. o t do well, Dr. Seiss, and what I think this method 4
does better is throw out the bad ones.
Picking high 5
ones is easy.
6 The report in fact that Janet did does not 7
even look at the high ones, it looks at the low.
8 3R. SIESS:
Yes.
9 MR. MINNERS:
And that is the hard part of to wha t we do is throwing out the bad ones.
Picking the 11 high ones, that is the easy one.
We do that every day.
12 MR. SIESS:
Unfortunately, throwing out the 13 low ones, there is always somebody vbo thinks that
(
14 should be in high.
15 MR. BENDER:
I think if we a re re ally trying 16 to make an issue over whether we should throw out the 17 low ones or not, this is a good wa y to do it.
It seems 18 to me that the emphasis is in the wrong place.
19 If the purpose is not to see which of the 20 important ones to concentate the resources on, then --
21 MR. MIN!!E R S:
That is the other side of the 22 coin.
What you throw away leaves you resources to work 23 on what you want to do.
That is half the job.
()
24 MR. BENDER:
If you are putting a lot of your 25 resources into the very low ones, I guess I micht O
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON D.C. 20024 (202) 554-23.'5
45
()
1 concede that that is worthwhile.
But when I look at the 2
rate at which the high ones are being accomplished, I 3
have a hard time believing that you are really putting r~s O
4 any resources on the low ones.
5 So, I have to ask mycelf, "Well, can I use 6
this method to decide how to use the resources I have?"
7 Tha t is what is bothering me right now.
I do not think 8
it will help you very much.
9 I am like Dr. Kerr, I do not care whether you 10 do it or not.
As a matter of fact, if it makes you feel 11 good to do it, I would say, go ahead.
That is a 12 personal judgment.
13 But if I am going to lay this out in front of 14 the public as something which is really helping to deal 15 with the way in which we are using our resources and 16 explain to Congress that that is the basis for it, I 17 think it is a lot of garbage.
I wonder whether there is 18 any way of showing that it is not.
19 MR. MINNERS:
Well, the problem with that is 20 that you have to come up with another way that is not 21 garbago.
22 MR. EENDER:
Yes, that is a good point.
But 23 you know, to substitute one kind of garbage for another
()
24 is not accomplishing very m uch.
25 MR. ERNST I would like to make one final ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W. WASHINGTON, D.C. 20024 (202) 554-2345
46
()
1 point, if I me y.
The Delphi versus what we are doing 2
now, versus --
3
.Y R. BENDER:
I am not campaigning for the 4
Delphi.
5 MR. ERNST:
Yes.
Versus other techniques.
We 6
are talkin7 methodology.
No methodology is going to be 7
perfect.
You are never going to get rid of the 8
uncertainties that are inherent in the calculation of 9
these things.
to We really are using what you migh t call a 11 modified Delphi procers, except instead of considering 12 the worthwhileness of an issue in talking in qualitative 13 g en e rali tie s, people can focus on specific numbers and 14 argue them.
We have had that.
15 I would like to emphasize Warren's point.
If 16 somebody has a better number or argument, or an 17 assumption that we misssed, fine, we will accept that if i
18 tha t is a general consensus, that was missed.
19 MR. BENDER.
Again you are emphasizing the 20 wrong thing.
The problem really is that we are not sure 21 that we have numbers.
Everybody else says, "look, if 22 you have better facts, we will use your facts," of 23 course.
()
24 The problem really is that in trying to make 25 the judgmerits en the basis of some numbers tha t really ft
'J ALDERSoN REPORTING COMPANY,INC, 405 VIRGINIA AVE., S.W., WASHINGTON D.C. 20024 (202) 554-2345
47
()
1 we do not know - there hava been several orders of 2
magnitude - to make judgments involving a range of 3
circumstances that are about two magnitude wide, we are f
4 offering the opportunity to make gross misjudgments just 5
on the basis of not understanding the data very well.
6 MR. ERNST:
That is one reason we went to I
7 three orders of magnitude on the saf ety goal before we E
had a drop because of consideration of uncertainties in 9
the calculations.
10 ME, BENDERS Well, but that does not help us 11 much, does it?
The safety goals may even be too low.
I 12 think there are a lot of people who say they are too low.
13 So, really, you are not dropping things out at 14 the level at which they ought to be dropped out.
15 MR. ERNST:
That may be, but we did wind up 16 with over half the issues in the 1cwer drop category 17 where they have been on the books for a long time.
18 MR. BEFDER:
I will agree with that, But 19 hell, I could have started without the exercise.
20 MR. SIESS:
Mike, I guess I am not sure why we 21 would not worry about the nethodology.
I assume that we 22 will eventually go look at the classification of these 23 items and try to decide which ones we agree with and O
24 which ones we do not.
25 Now, I guess having discussed the methodology O
ALDERSON REPORTING COMPANY, INC, 400 VIF< GIN!A AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
l u8
(}
1 we will feel battar about disagreeing with something 2
becauce we know the methodology by which it was arrived 3
at.
O 4
But is not our main problem to decide whether 5
these things are properly classified in our opinion?
6 Our methodology may be different.
7 MR. B"NDER:
Well, I think that just looking 8
at it in a personal way I think I would agree with you.
9 But if the staff wants the Committee to agree or l
10 disagree with this thing, then I think we a re being put 11 on record where it invites the opportunity for somebody 12 to say, "Well, look, we offered'you the method and you 13 agreed with it.
And now you are objecting to the 14 judgments made on the basis of the method."
15 That is what bothers me.
If the Committee 16 vants to write a letter that says, "If you want to do 17 it, fine, but the judgments ought to be used without 18 much emphasis if they are derived in this way," then I 19 think that is a legitimate position.
20 MR. SIESS:
Well, it seems to me there are two 21 things the Committee could do.
One is it could say, 22
" Gee, the methodology is great, therefore the answers 23 are OK."
()
24 And the other is simply to ignore the 25 methodology and look at the results and see whether we ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., SM., WASHINGTON. D.C. 20024 (202) 554-2345
49
(])
1 agree with them.
2 Now, I would suspect as we look down the list, 3
we are going to agree with 90 percent of the highs and 4
about 90 percent of the drops as a Committee.
There 5
will be individuals looking at drops saying, "No, I
6 think you ought to do that."
7 And the others that are in between do not make 8
that much difference anyway.
The mediums are still 9
going to be looked at.
The lows nobody is going to get 10 around to in our lifetime.
11 But I would suspect from just look,ing at the 12 thing on whatever basis we look at things, we will agree 13 with 80, 90 percent of them, and then we probably would
(
14 not agree with the others no matter what the methodology.
15 Sut I do not think we are going to say, "Yes, 16 the methodology is great and therefore we accept the 17 whole kit a nd ca boodle."
18 "R. DENDEF s Let me try again to establish the 19 position which is, if the staff wants to use the 20 methodology we have no objection to it.
But as for 21 final judgments, we think it would have to be done on 22 some other basis.
23 MR. SIESS:
We are going to render an opinion
( })
24 on the results at some point in time, and we are not 25 going to tell them what our methodology is.
O ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
9 50
()
1 MR. BENDERa Well, I am not going to say that.
2 (Laughter.)
3 MR. EENDER If the first part of the gg U
4 sta tement is all right, I will draft th e letter on the 5
basis of th a t.
Does anybody object to it?
O MR. MOELLER.
I think you may have been a 7
little harsh on the staff.
I believe and I know they 8
are conscien tiously trying to come up with a 9
quantitative formula f or reaching these decisions.
10 I see this as an iterative process in shich 11 they come to us with the plan.
We react.
Then they 12 come back, and so forth.
13 Now, that was what I was sort of referring to 14 ini tially.
If the schedule is such that this all has to 15 be decided in the next three weeks and we have to write 16 a letter saying, "Yes, we fully agree --
17 MR. BENDER:
They want a method sanction.
18 MR. MOELLEB:
Well, then I am not ready to 19 sanction the method.
But I am ready to encourage then 20 to keep moving f orwa rd.
Perhaps - I do not want to be 21 discouraging to them - but perhaps to " massage" this a 22 little more or consider it a little more, a little 23 longer time and come back to us with some other pla n.
()
24 But I vsnt to compliment them on what they 25 have done.
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINI A AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
.~ -
1 51
()
1 MR KERR.
I do not disagree with what you have 2
said.
But do you have suagestions as to how they might 3
e-)
" massage" it or that it migh t be i.. proved significantly k/
4 by additional massacino?
5 I guess I tend to think that not much 6
additional massaging is needed.
I think thay a re using 7
it as an aid to decision-making and not to make the 8
decisions.
9 MR. MOELLER:
That is good.
I think if you 10 asked me to make a couple points on it, one would be to 11 list a complete itemization of corrective measures that 12 should be taken and then say, "We have only considered 13 the following ones because we do not know enouch about 14 these others," or something like that.
That would be 15 one improvement.
16 The second improvement would be to somehow put 17 into the formula the error band or uncertainty in some 18 of the nu'2 b e rs.
We would not argue with them.
They 19 would just say, "This is th e number we took, it could be 20 higher, it could be lower."
Then we would look*at it in 21 that light.
22 But I have been trying to do this same thino 23 on other problems and I know how - and we all do - we
()
24 know how difficult it is.
I give them a lot of Brownie 25 points for doing what they have done.
O k_/
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
52
(}
1 MR. BENDEP:
Chet?
2 MR. RIESS:
I am not going to disagree a 3
hundred percent with Dade.
But I do not think that the O
4 data base warrants any significant attempt to define the 5
quantitative approach.
6 I do not think they are going to come up with 7
any better answers than they hve now.
I prefer to look 8
at the answers and then argue, like Warren said, if I 9
disagree with it try to figure out why and take them up 10 almost case by casa in the areas of disagreement.
There 11 are not going to be that many areas of disagreement.
12 MR. BENDER:
It seems to me Dade made a fairly 13 important observation in his suggesting that whatever
(
14 the method is, that it take account of the 15 uncertainties.
It seems to me we could say that.
It 16 might eliminate about 90 percent of the reservations 17 thst most of us have about it.
18 Unfortunately, it may turn out that in most 19 cases the uncertainties are so large that you have to be 20 very careful shout using the method.
But I think that 21 is a fairly straight-forward way of qualifying the 22 result.
23 MR. MINNERS:
The uncertainties a re included,
()
24 are supposed to be included in the issues.
We got a 25 little sloppy at the end and I notice that the ALDERSON REPORTING COMPANY, INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 2CO24 (202) 554-2345
53
()
1 calculations of the uncertainties sort of faded away in 2
some of the later analyses.
But that is supposed to be 3
gg one of the features of it.
U 4
MR. PENDER:
Jan?
5 MS. PEESTON:
I have to agree tha t when I 6
first looked at the score I falt lik e it was a good 7
score.
The only reservation, really, I have about the 8
methodology - other than the fact that the data base is 9
inadequate - was the fact that there is the averaging 10 tha t oces on.
11 I think for those problems that there is an 12 extreme variance in severity that perhaps they should be 13 a little bit more careful about applying the methodology
()
14 because if you have an extreme problem for just a few l
15 plants, that still is important.
I l
18 MR. BENDER:
Dave?
17 MP. WARD:
I would like to go back just a 18 minute and ask a clarification f rom Warren on the 19 calculations that have been done.
20 Do all of the high priority items come from 21 involved release categories that include core melt?
22 MR. MINNERS:
Yes.
l 23 MR. WAED:
They do.
Well, then I am a little l ()
24 bit puzzled, or I question a little bit how useful LERs l
25 have been in providing frequencies there.
I mean, I am l
ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W, WASHINGTON, D.C. 20024 (202) 554-2345
54
()
1 recalling the kind of agonizing effort that ORNL and SAI 2
did in trying to make some sense out of ten years of LER 3
data, precursors to core melt opposite the results of 4
paas.
5 So, is it really valid to say you get 6
something useful from the LER data base?
7 MR. MINNERS:
The cethod we use is what 8
everybody else usas, I think, a synthesis in which you 9
use empirical data and you combine it with some kind of 10 a PRA event tree, fault tree methodology.
11 The numbers that you cannot get out of the 12 data you get some other way.
You use Swain's Handbook, 13 all righ t?
Or you use the Delphi method, even, and you
()
14 come up.with a number, 15 You have to do that.
I am extremely disturbed 16 by the discussion of this method in an absolute. way.
j 17 You keep saying this method is not very good and I agree 18 with tha t.
19 dowever, comparing it to the other available 20 methods which basically come down to a group of people 21 sitting down, writing it down in a qualitative way is 22 what the argument he re is.
23 MR KERR I want you to know that Dr. Moeller
()
24 said that it is a good method and he gives the staff 25 large Brownie points.
So, don't go away with a negative l
(:)
l ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
55
()
1 feeline about the nethod.
2 MR. MINNERS:
I am trying to make the peint 3
that it has to be compared to other methods.
It cannot 4
be just looked at by itself.
There is no absolute.
5 MR. BENDER:
Jerry had a question.
6 MR. MINNERS:
So, all the questions that you 7
have about a bad data base apply to any of the methods 8
that you would do, the ranking sitting around a research 9
table, the same thing.
10 MR. RAY:
I tealize you are over time, but I 11 wondered if the staff could take just a couple of 12 minutes to respond to a point that Jan makes on page 2 13 under the caption, " Conclusions."
O>
N-14 After she makes the point that only very 15 limited use of the formula was appropriate she says, 16 "This may be substantiated by the fact that it applies 17 to less than 20 percent of the safety issues identified i
18 in the NUREG."
19 If that is true, I wonder why.
'as it because 4
l 20 of the data base?
21 ME. MINNEPS:
I do not understand that 22 statement.
I do not know what that statement means.
23 MR. BENDER:
I think if you counted up what is
()
24 in there you would see that if you lock at everything l
l 25 that is in the unresolved safety issue, in the l
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
56
()
1 uncatecorized list -- and what is the other one?
2 MS. PRESTON:
TMI.
3 MR. BENDER:
And TMI, there just are not very 4
many tha t a re covered.
5 MR. MINNEES:
We do do TMI, we do all of the 6
TMI issues.
They are all being done.
7 MR. RAY:
So, your point.is that it is more 8
than 20 percent.
9 MR. BENDERS I do not think it is.
10 MR. MINNERS:
All tha t the rest of that 11 sta tement sayc is th.m.t the list of generic iscues that 12 we had was a very long list of everything, including th e 13 kitchen sink.
That includes licensino approving issues, 14 environmental issues, issues which are already being 15 implemented.
We just happen to have a comprehensive 16 list which is a procedural thing to be sure that we 17 capture eve rything.
18 HR. RAY:
Well, if there are only 20 percent 19 of them that can be prioritized this way, we have not 20 accomplished very nuch in the way of prioritization.
21
,MR. BENDER:
Well, they really just have not 22 done that much.
23 MR. RAYS I realize this is not your statement.
()
24 MS. PRESTON:
Maybe I can clarify it.
25 MR. MINNERS:
A method was applied to 99 O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345 I
57
()
1 percent of all of the generic safety issues that are 2
open issues.
If the issue is closed and we already have 3
a requirement, we did not prioritize it.
Why prioritize b
Ns 4
it?
5 If somebody wants to bring up the question of 6
a de-ratchet on the TMI issue or a cule or regulation 7
snd wants to volunteer the generic issue, we will 8
evaluate that too.
But that was not the basic purpose 9
of the study.
10 The basic purpose of the study was, " Hey, dn I 11 have some ha zard to the public which I possibly need 12 more requirements on?"
We had 99 percent of those by 13 this formula.
(
14 MR. RAYa Do you want to tell us how you got 15 the 20 percent?
16 MS. PRESTONa OK.
The 74 items received a 17 high, medium, low or drop priority.
There are a total 18 of 460 isi;aes of which 33 were covered in othe-issues.
19 Basically, they combined them, turbine missiles was 20 one.
There were several tha t were thrown into the 21 turbine missile issue.
22 I used 460 minus 33 and said, "Vell, you cot 23 74 of those, it is like 16, 17 percent.
Th at may have
()
24 been a little bit of an overstatement though I did not 23 feel it was because I felt you might as well go back and O
l ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
58
(')
1 prioritizo the 0737 issues.
2 Let me mak e this point, though --
3 MR. FINNERS4 I am sorry, sa y tha t again, I 4
did *ot hear what you said about 0737.
5 MS. PRESTON:
I felt like, when I started out, 6
that the 0737 issues that were mandated, it would not 7
have been a bad idea to prioritize them.
But that has 8
nothing to do with this.
9 Let me say one thing, though, 74 have been 10 prioritized.
In the table on page 41 it gives a listing 11 of how things were resolved 74 are high, medium, low or 12 drop.
Fucure prioritization meaning - that was Note 4 -
13 meaning they have not been done yet sss 156.
So, you
)
14 can at least, say, change the number 20 percent to 30 15 percent in my report and I do not really think you can 16 get any disagreement with that from Warren that less 17 tha n 30 percent have been.
18 MR. ERNST All of the Note u items have not 19 been prioritized.
20 MS. PRESTON:
They have been where they were 21 not in the document.
22 MR. ERNST:
That was the purpose of Note 4 to 23 say they are in the process of being prioritized.
()
24 MS. PRESTON:
OK, all I had to work with was --
25 MR. SIESS:
Where is tha t, in another NUREG7 O
ALDERSON REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
59
()
1 MR. ERNST ro, it is out for office review 2
richt now.
These are other office itens.
As soon a s we 3
get comments back --
O 4
YR. BENDER 4 look, there is no point in 5
debating whether there are 20 percent or 80 percent.
6 The real question is, do we want to support the method
~
7 as being an effective way of judging generic safety 8
issue prioritization.
9 As nearly as I can tell, the most that I have 10 heard in support of it is that we do not have any any 11 objection to using it if it will help the staff in its 12 judgment.
13 At the same time, we are not of the opinion 14 that by itself it represents very much of a basis for 15 making such judgments.
Is that incorrect?
16 MR. SIESS:
I think I would say, as far as I 17 am concerned, I have no objectons to the method.
But I 18 do not support the method to the extent that I accept 19 all of the answers.
I reserve the right to disaaree on 20 an answer.
I do not want to say that accepting the 21 method means I accept the answer.
l 22 MR. BENDER:
'4 e ll, that is true of anything.
23 f That is not much of a judq=ent.
()
24 MR. OKRENT:
Did your discussion include the l
25 process of using the method as part of the method, or is f
i ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345 1
60
(])
I tha t e onsidered sorething sepa ra te ?
2 MR. BENDEPs We spent a fair amount cf time 3
discussing how the method would be used.
I think I
(
4 wculd have to judge right now as the staff has just 5
said, "Well, this is a way of dealing with things that 6
we do no t al ready have a position on."
7 MR. OKRENT:
No, but what I have in mind is 8
really, a method is sometimes only defined by the mode 9
in which $t is applied bec?use prior to that it is a i
10 little too general or whatever.
11 By " mode" I mean, for example, the degree to 12 which the results are given critical review.
That, to 13 ome, is part of method.
It is not part of results, it n/
s-14 is not part of data.
15 MR. SIESS:
They have done peer review.
16 XR. BENDER:
I think only to the extent that 17 they involve resources have I heard tha t they have used 18 it in that way.
However, I guess that a lot depends on 19 how much effort went into the prioritization part of the 20 effort.
21 MR. SIESS:
How much peer review have you had?
22 ME. ERNSTs We ha ve ha d peer review f rom, 23 clearly, within NRR and outside, other members of the
(])
24 stsff, ACSS, and it is a public document.
We do 25 consider the document to be a living document.
O 1
ALDERSON REPORTING COMPANY,INC, 400 VAGINI A AVE., S.W., WASHINGTON, D.C. 20024 (2C2) 554 2345
.. _ _ - ~
-- ~
61
()
1 MR. OKRENT:
No, I am talking about peer 2
review within the staff prior to your submitting it to 3
the ACES for comment.
4
R. E R N S T s Thst was done with th e cognizant 5
branches involved, and in most instances when they had a 6
different judgment on frequency or sequences, or i
7 consequences we accepted their judgment in coming up 8
with the final numbers.
9 MR. OKRENT:
Well, as tl.e Committee is well 10 a vs re, we did little peer review on the backup document d
11 on seismic scrsm, and the choice of informa, tion there 12 that was used in it as f ar as I can tell is very poor.
13 If +he stuif did peer review on that, that 14 information should have, in my opinion, he changed 15 before if was presented to the ACRS as the reason for 16 decision, whatever the decision was.
17 This is what I mean by process, and if you are 18 telling me in fact you reviewed it and the information 19 in there was good from your point of view, I have a 20 problem with the process and t h ere f o re with the method.
21 MR. EENDER:
We are still coing to look at the 1
22 priorities when they are set.
That is the basis on 23 which the Committee is going to make a judgment.
()
24 It seems to me un til we have seen the product l
25 it is hard to make an overall ascessment.
Now, the best l ()
l l
ALDERSON REPORTING COMPANY,INC, 400 VIRGINTA AVE S.W.. WASHINGTON D.C. 20024 (202) 554 2345 l
62
()
1 we can do is make some comment about whether the method 2
should be used at leert for'whatever value you can get 3
out of it.
4 My inclination right now is to write something 5
positive in that direction tut put a fair amount of 6
cautionary language in it.
If that does not bother 7
anybody that is about what I want to propose to the full 8
Committee.
9 MR. OKRENT.
Well, I have one other question -
10 I am sorry for coming in late.
11 They have something called cost in the formula 12 and again I was wondering whether the staff has done 13 peer review or whatever it used on the cost part of that 14 analysis.
15 MR. MINNERS:
We tried, Dr. Okrent, but it is 16 a difficult thing to do, to get costs.
We ha ve done the 17 best we can on the information.
18 As I tried to discuss in ny presentation the 19 last time is, there has to be some cut-off point on how l
20 much time you spend on prioritizing an issue.
It is not 21 supposed to be a value-impact analysis to make a final 22 decision.
23 So, that is one of the problems.
So, I think
()
24 the costs havo generally suffered from that because to 25 get qCod costs you would probably have to go out to do O
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D C. 20024 (202) 554 2345
63 1
some kind of a cost analysis.
2
- 93. FRNST I would like to make a comment 3
since Dr. Okrent was not here earlier.
O.
4 In the matrix we looked at, I think, on only 5
about four or five issues was the ranking of the issue 6
changed because of the cost-benefit analysis.
I think 7
the ranking is essentially, most issues remain the same 8
based on the net benefit that you would achieve compared 9
to something like the safety goal as the categorization 10 that you get by considering value impact.
11 MR. MINNERS:
As Dr. Siess pointed out, when 12 you look at the criteria ma trix, it has jogs in it 13 because of the value impact.
So, it does not change 14 things very much.
15 MR. GKRENT:
Well, I have been re-reading the 16 latest staff position on ATWS which is an important 17 issue, and cost there is a factor and sometimes you have i
18 a value impact of
.5 where in the cost more than half of i
(
19 the cost is due to down time.
20 MR. MINNERSa Right.
21 3R. OKRE9T:
But it is conceivable that the 22 things could be dona with no down time in view of the 23 lenoth of time required for some of these things.
Do I O
24 mexe the==1nte 25 MR. MINNERS:
I agree.
But once you get to O
4 ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
, _. - -. -. _, _. - =
6u 1
the question, down-time overrides the other cost.
Then
)
2 you have to figure out what kind of a fix am I coing to 3
do.
Would that actually require down time.
That is a 4
hard judoment to make about a future event.
1 5
MB. EENDER:
We are about 30 minutes over the 6
allocated time for this.
7
.1R. RAY:
Mike, as a member of your 8
Subcommittee I would be pleased to see a draft along th e 9
lines which you mentioned earlier.
I suggest you do it 10 that wa y.
11 MR. BENDERS We will put together something.
12 I think based on the discussion this morning that 13 somebody from the staff should come in and plan on j
()
14 giving about a 10 or 15-minute overview of how the i
15 method would be used.
Not so much th e me th od, I think 16 the Committee understands it pretty well.
17 But what is its real value and what are the i
18 risks in using it as a basis for judgment.
19 MR KEPR And give some examples of cases, 20 perha ps.
We would find that helpful.
21 ME. BENDEE4 Are there other comments?
If 22 not, let's adjourn this meeting on schedule.
23 (Lauchter.)
{}
24 MR. BENDERS Within the crder of uncertainty 25 tha t we work with, it is on schedule.
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W WASHINGTON. D.C. 20024 (202) 554 2345
. -. ~,
-. -. _ ~..
.. _. -. _. - _... - - ~..
.. - -. - -.... _.. - - - -.. =. _ -. _ -
I
}
65 i
t
]
1 (Whereupon, at 9a25 a.n. the Subcommittee j
2 ad.1ourned, to proceed with other business.)
3 i O 5
l l
6 l
7 8
i 9
10 1
11 i
12 i
13 l h 14 l
15 16 1
17 I
L i
18 i
19 20 k
21 l
22 l
23
[
l t
24 I
25 i
O k
l i
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 564-2345 I
MUCLEAR REGULATORY CO.'EISSION m
This is to certify that the attached ;receedings befors'the ADVI.SORY COm1ITTEE ON REACTOR SAFEGUARDS in the natter cf:
SUBCOMMITTEE ON GENERIC ITEMS Date of Proceeding:
Januarv s. 1991 Dccket !! umber:
Place of Proceeding:
Washington, D.C.
were held as herein appears, and tha: this is the original transcript therecf for the file cf :he Cccmission.,
M.
E.
Hansen Official Reporter (Typed)
!f-A444Als Official Repceter (Signature)
I i
_.s 8.
f l
--,.--wwr,-,....
c_-,.
_ -. _ _ _