ML20010C440
| ML20010C440 | |
| Person / Time | |
|---|---|
| Issue date: | 07/24/1981 |
| From: | NRC |
| To: | |
| Shared Package | |
| ML20010C401 | List: |
| References | |
| NUDOCS 8108200083 | |
| Download: ML20010C440 (40) | |
Text
-
s s
I 1
1 1 4 UNITED STATES OF AMERICA l
2 NUCLEAR REGULATORY COMMISSION 3,
PUBLIC MEETING 4"
j SAFETY GOAL WORKSHOP I
e.i SESSION 6 h
0 g
GENERAL DISCUSSION n
7]
s j
8 Cliffside Inn
'J d
9i Harper's Ferry, West Virginia i
O I
10 t Friday, July 24, 1981 z
i 11 The meeting was convened, pursuant to notice, at 3
i j
12 l2 : 07 p.m., George Sege, Chairman, presiding.
i E
E 13 i PRESENT:
E i
h 14,
D.
RATHBUN S.
DERBY w
2 i
E 15 '
L.
LAVE D.
MAC LEAN J
16 '
P.
SLOVIC C.
GIBBS en d
17 E.
O'DONNELL s
18 '
H.
KOUTS
{
E l
?
19 '
T.
COCHRAN h
20 3 H.
INGRAM i
I I
h 21 E.
NEISS li l
22 j M.
TEMME l
1 23 i C.
WHIPPLE I
24 D.
BRIDENBAUGH
[
25 V.
JOKSINOVIC y
k l
8108200083 810724 ALDERSON REPORTING COMPANY, INC.
PDR REVCP NRCIS9DS i
i PDR 1
l a
Y (12:07 PM).
g I
MR. RATHBURN:
Gentlemen, I would like to present Mr.
j 2
George Sege, who will conduct the general discussion in this last 3l session of the workshop.
4{
MR. SEGE:
Thank you, Dennis.
I g
5i To start off the general discussion, I would like to 9
I 3
6l call on the three panel chairmen in turn, and then various mem-E 1
7 bers of the -- other members of r_ne workshop to articolate what s
j 8
they feel are the most significant highlights of the judgements d
z, 9l that have emerged from this workshop -- the most significant things 2
Og 10l that you should be doing to the discussion paper that has been i_5 l
j 11j placed before you for discussion in order to convert this to a a
i j
12 sound and effective document for Commission consideration.
=
13 Let me call on Herb Kouts first.
14 MR. KOUTS:
Well, I have a short list of things which t:!
is,!
g' 16 s
d 17 w
?
i
- o 18 :
=
l I9 5
l 20l t
21!
22 {
23 24 j 4
25 i
ALDERSON REPORTING COMPANY. INC.
e i.
3 l
i!
1 sk 4 1 1 I think have struck me most as beina strenaths that we can F
~
2 draw from this meeting.
It's not surprising to me that the draft 3
i 3y which was produced and which was circulated for comment did not i
4' strike universal acceptance among all tne points of view that aref I
1 l
g 5l represented here -- probably not complete acceptance by anyone j
6 here.
Af ter all, this was a great extrapolarion from the Palo R
2 7
Alto meeting where we realJy did not even talk about the spe-g' t
I i
8 cific content of gaols, but only tried to talk about the char-l 0
l 9I acteristics that the goals should have.
zc g
10 l And I did find, for my part, tnat most of the char-x i
]
Il !
acteristics that Panel A at Palo Alto had asked che characteristic a
l 5
I2 of whatever document was generated; most of those features were 5
i 13 '
really incorporated in the draf t report -- perhaps not in a com-z 5
I4 '
pletely acceptable way, but they were addressed.
It's clear i
e i
.}
15 tnat a redraft of this document is going to have to be made from
=
l j
16 '
a number of points of view -- clarity is coing to have to be l
A d
17 I introduced, another look has to be made at the recommendations N
I
}
18 ;
from Palo Alto, and perhaps the comments that came in after that i
I9
- time, f
n 4
l 20 There will have to be introduced better argument, l
]J 21 better understanding, better basis for understanding of what is 22 )
said in the document; and even -- it seems to me -- an expansion !
.i 23 of quality of the goals.
Now, beyond what is done by way of 24 a number of qualitative goals -- additional qualitative goals --
25 I'm not at all sure what the cuidance of this aroup has been with!
I i
ALDERSON REPORTING COMPANY. INC.
I
e i
l l
d l
C
'l i
sk 5 I I respect to the specific goals which are in the document.
I think!
h 2]
that the document will either be very good to follow Helen Ing2mt's i
i 3"
suggestion that there be a sort of distant goal incorporated --
l 4
what is it we're really trying to accomplish by all of this 5
g objective?
H 6
I think the arguments for this were very powe -ful, e.
7>
=
S and the distinction between goals of that kind and the operative 3
8 devices which are recommended for Commission use will have to be G
a i
9' made clear in the document.
Some of the numbers in the cuanti-z c
1 g
10 jr tative statement will probably have to change; although this is E
5 II l
not the kind of forum where you decide what those numbers cught l
e j
12 ]
to be.
Attention certainly will have to be given to the spe-
=
J y
13):
cific feelings that certain numbers are too large or too small,
=
{
14 i
or something of this sort.
5 l
]
15 '
But the actual determination of what is reasonable
=
j and proper probably follows a systematic study which takes into 16 A
N I7 account the rationalle behind the choice of the numbers in the E
y 18,
first place.
The final thing that I think is clear is that the i
h n
g 19 1 discussion that the document makes with respect to the relation-n o
f 20 ship of this technology with other technologiee; that corres-a
]
pondence and the implications of that correspoidence with safety j 2I 22 l goals will have to be clarified.
These are the major insights 11 23 that I've received at this mee ting.
24 MR. SEGE:
Thank you very much, Herb.
25 Lester Lave?
l l
ALDERSON REPORTING COMPANY. INC.
e s
I 5
i i
s sk 6 I !!
MR. LAVE:
Yes sir.
2l Th..re are a small number that I draw from the panel 3,
-- my panel -- and from the discussion last time, and I want to l
4 summarize the ones from last time so that people think that I had I
5 g
a tin ear will let me-know; and at least I'll write in there that Ha 6l I have lasti: e animity in the things I have to say:
R 7
The first idea revolves around the notion of different 0;
goals for different people, and the very distince audiences we J
z.
9l have of needing some kind of -- almost piece of rhetoric -- to O
=
\\
h 10 communicate what the goals of the NRC, with respect to the popu-
=
i j
IIj lation in general -- some-5 ng more specific than that in comm-e 4
3-12 '
unicating from the workshop to the NRC.
And then we get progres-E 5
l3 l sively more specific for communicating to the industry; and
~
=
I4 l'
T then finally licensing individual reactors.
M i
g 15 :
And part of what I want to stress out of all of that E
t g
16 i is that these cannot be disparrate statements; they're statements A
I I7 in different languages with different levels of specificity, but
=
IOl they all have to integrate to the same thing -- you can't have 3
C i
s 19 g
q something that means something dif ferent.
There's no possibility
=
?
20 of hiding from one group what's said to another, and any contra-2I dictions are going to come out and be played upon.
22 )
The next notion is one that says that it doesn't make 23 to talk about goals -- qualitative or quantitative --
I any sense I
24 without talkina about the implementation, means or verification, 25 and so on.
This is all part of one package; they all have to I
O l
ALDERSON REPORTING COMPANY,INC.
o e
i F
?
6
?
Ek 7 I
be dealt with at one time.
They all affect the credibility of 1
3 24 what it is that one has to say.
That, then, leads to some state-a i
3 !
ment about integration -- I'm sorry; to just go back for one f
I 4'
second:
Il g
Those dif ferent goals should be not only phrased l
5 H
a j
6 l dif ferently, but should have different amounts of quantification' R
A i
=
7 q in them depending on which group you're talking about.
Quanti-E 3
!i A
0' fication to the population in general, I think, means nothing;
'm O
9'
?_.
we all know from the work that the social psychologists have done y
10 '
that people -- even experts -- have a very difficult time relating M
2 II
-3 y
j to numbers that are smaller than 10 And I think that those s
12 '
2 numbers are kind of meaningless to the putlic; what we need is 5
g 13 a qualitative description of what they mean.
=
T I4 But when you're talking about licensing a particular
-e I
h 15 ;
reactor, then... " ther the vendor nor the regulator has any idea
=
i l
y 16 of what you're talking about unless you put some numbers in therei z
17 Those numbers should be subject to chcnge as we gain more know-
=
3 18 '
ledge, as we resolve some uncertainty as our goals change.
And W
I9 3
so it should be easy to change them, but in the meantime, they e
c 20 $
ought to be crystal clear and verifiable.
21 Okay; with respect to implementation:
I think that 1
i 22 1 Ed Zebroski made an admirable statement about qualitative imple-23 mentation goals, stressing the need for on-going collection of 24 data for the analysis and integration of the analysis and the 25 data toward current reactor operations, current design, current i
,\\
ALDERSON REPORTING COMPANY. INC.
i
~
f 7
l Ek 8 1 ]
licensing -- a continual feed-back system the objective of l
1 2j which is to improve -- to enhance -- safety over time.
Some way E
I 33 or the other we all have this feelina that if the system is i
i d
l 4 j going in the right direction and going there at a non-trivial j
l e
5 speed, then even if you didn't start out exactly right to begin E'
4 6j with that it will all come right in the end.
G 7j To go a bit more afield:
There was this idea of how 6
l g
8' in the world do you wind up making comparrisions?
Is it reason-G 9
able to talk about nuclear being unique; how do you look at these i
i 10 !
things?
And it seemed to me the principle that we'd arrived at d
11 last time in Palo Altc was that you need to look at relevant j
e j
12 comparrisons.
The question however is what is a relevant com-E a
j 13 parrision for one person -- is not relevant for another; and
=
r 5
14 that's just a very dif ficult -- a very sensitive -- issue.
E I
j 15 ;
I had thought that there was at least near agreement I l
E 1
j 16 on the statement that when we're talking about a new installation l x
d 17 that what we ought to be doing is comparing the risks associated s
l E
18 ;.
with that new installation with our alternatives; namely:
The h
I 19 risks of conservation, the risks of alternative ways of generatind R
1 20 5 tha t energy, and that af ter adjusting for uncertainty and differ-i 21 ences in cost levels and so on, that we ought to be chosing the f
22' system that has the least risks.
23j And I understand that it is not an easy or content-l 24 less statement to say that since the dimensions of risk are dif-25 ferent and the dimentions of uncertainty are dif ferent for 4
ALDERSON REPORTING COMPANY, INC.
I 1
l i
8 i
k I
i Ek 9 1
different technologies, one can't in any simplistic way say l
l 2
clearly one technology dominates another.
I hink that there wasi 1
1 3ha lot of clarification in my mind about wha t quantitative sa fety f 4
goals meant, and what the role of PRA was -- that PRA was in one 5
g sense less than ability to get an overall look at the safety of H
j 6
a nuclear reactor.
6 7
On the other hand, everybody agreed that it was a
~
f.
8 very good tool for setting priorities for looking at relative 9
values.
Not perfect, btt a very good tool.
And certainly when
?.
10 '
one is talking at the operational end of this set of goals, then d
II l PRA clearly looks to be essential.
At the most general and com-
]t j
12
~
municating with the public, PRA is a disas ter.
And so the qu a-
~
13 tification ought to be at the more specific end of it; it ought n
14 to be a quantification that varies easily -- where the rules are
=
{
15 ll easily changed o"ar time t a ?; ones where there's either a 3 0 'b d
statute or a vast rule making that's required to make some of n
N 37 these changes.
3
{
18 I thi nk that I'm just going to end with the note l
i 19 g
that I found this workshop and the discussions of it immensely n
20 j more helpful than Palo Alto workshop; and that to be immensely 2I to be more helpful than literature tha t I've read.
That is, I 1
22 !
think tha t not only have we managed to wave away a lot of issues i
23 that have previously plagued us in de discourse, but we seem to i
24 have arrive not in agreement but at least at aarselling down of 25 the issues where we disagree -- and a lot of agreement on a set f
l ALDERSON REPORTING COMPANY. INC.
l
I 1
2
)
9
?
i 4
i sk 10 I j other issues.
And I personally found that very helpful.
l l
2l MR. SEGE:
Thank you very much, Lester.
4 l
3 Paul Slovic?
4 MR. SLOVIC:
I have a short list of points that high-i 5
light; the first of which has been mentioned by both the previous 2
6 !l speakers:
And that is the need to elaborate the goal structure, 7hl 5
5 and distinguish between the different types of goals and their 0
'i 8 ',
purposes and relevant means of implementing them -- ranging from
=s 0
9; the broad symbolic goal to something which is narrower and might
~.z t
i 10 +
j be construed even as a standard.
=f II l Second, I think that there is concern expressed e
I,' '11 2
throughout about the need for the document -- the goal statement i
I3
5_
as a whole to be logically sound and coherent throughout and n
I4 internally consistent; and this will take a certain amount of e
l
}
j care in checking to insure.
l 15
=
j
.f 16 '
Third, I'd like to comment on the notion that the l
z 17 I
3_
goal setting process involves sort of a trio of componenet which l
C 1
IO hase not been considered in balance so far; we heard on the firsti i
i, n
I9 i
day that the probabalistic assessment -- which is the first com-l n
- l l
20 d ponent, and then the one that's dominated the ciscussions today i
l I
-- has certain problems and a great deal of uncertainties asso-22 1 ciated with it.
We heard this morning from Neal Wald that the 23 consequence model which goes along with PRA has been relatively 24 '
eglected, and also has a lot of uncertainties.
25 And I would like to again emphasize th at there is ALDERSON REPORTING COMPANY. INC.
3 l
J 10 i
l i
I l
3k 11 a 3rd comoonent -- the social value model -- which is even less l
2 $
well-developed and in need of careful consideration.
The dis-3 cussion this morning in this social evaluation session -- just l
4 on one topic, risk aversion -- I think served to demonstrate the '
I e
5 complexity of the issue and the difficulties involved here.
~
6 o
And for those of us who feel that the current docu-7 l:l F.
l ment, third draft statement, did not adequately reflect the social b
u 1
F'a 8!
value issue -- I think we face a challenge of providing scme c
9 i
feed-back in the form that will be useful for the revision of j
E 10 t j
1 such a document.
That's all of my comments.
=
i n
11 g
l MR. SEGE:
Thank you very much, Paul.
[
12 '
d I would -- at this point -- welcome views of the
=
13 g
highlights of what needs to be done from other members of the 14 d
workshop.
And I see Ed O'Donnell's card turned up.
r 9
15 a O
MR. O ' DONNEL :
Yes.
1
=
16 i
7 I'd just like to summarize my own views of what's x
17 y
been discussed and also what's in the proposal:
I agree with E
18 l
what's been said here about the need tc e labora te the qualitative ;
m l
19 i j
goals, and to make them perhaps the fore-front of what is commun-;
20 4
?
icated to the public in terms of what the objective is in trying 21 i
to achieve adequate safe ty.
Uith them must go, I think, the i
22 ]
quan ti tz ve goals for use by the staff in decision making where 23 PRA is used.
I think the qualitative goals should include tne 24 idea that de fense in depth is some thing chat is going to be 25 retained; and regardless of what results of any PRA are.
i i
i ALDERSON REPORTING COMP ANY, INC.
3 1
11 s
N 1
sk 12 I
Also, I think an important consideration to be l
t 2 !1 included in the qualitative as well as the quantitative goals is l l
I 3
h the concept of resource allocation and some sort of " benefit 4
balancing" once you've met some primary goal.
Again, I think 5
g the -- if we're going to use PRA, which we are; in fact, both H
j 6,
the industry and the NRC are using PRA -- in a quantitative sense r R
7 one must have quantitative goals whereby we can reach decisions --
i 5
0 A
8[
or at least have input inte decisions.
G 0
9 i The goals, I think, should be considered interim; not z.
t b
10 I something that should be set down in a rule that's inflexible and i
/
i t
=
1 i
?
II Il not subject to change.
They should be reevaluated periodically a
il i
12 as we gain more knowledge of their use and application.
So what-5 g
13 ever goals -- I think the qualitative goals should be set down
=
z 5
I4 '
with more conficence than any quantitative goals.
Quantitative 1
j 15 i goals obviously have to be considered in' erim and for trial use.
=
j 16 With respect to the specific proposal, I agree with A
$[
17 the statements made that there's got to be a much better linkage a
be tween the qualitative goals and the quantitative proposals; and!
y 18 l
g I9 ]
there's got to be a much better explanation of the rationale and l l
20 the underlying philosophy or thinking behind this entire structure.
21 And the concepts that we've discussed in this workshop and at l
22 1 Palo Alto have to be discussed; and if we're not capable of deal !
a 23 ing with certain subjects, let's admit that -- that we don' t know !
24 how, at this time, to quantify a particular element The numbers 25 themselves in the proposal look to be reasonable; I don't think I
9 i
ALDERSON REPORTING COMP ANY, INC.
l
e i
12 l
)k 13 I
we have to -- again, if we' re propc sing these as interin goals 1
2 be concerned whether or not they re a f actor of too high or low.
j 3 k We have to, I chink, start using some thing; and if these valuas 4
look to be reasonable as a starting point, well, we could use l
I e
5 these.
l P.
g 6
I think with respect to what is missing in the quan-I.
4 U
^l 7
tita tive sense, again, is an explicit expression of the ALAPA
{
U d
7.
8 principle -- which is kind of in here, but on a very qualitative E
I 9i basis.
And if that's going to be part of the decision process, l
d_
g 10 :.
then in determing the need for change, it should be in here d
i
=
II 4
explicitly in a quantitative sense.
I think the most important l
e j
12 goal in the whole set is really the individual goal because that 5
~
13 sets the tne limit at which no individual in the public is going
=
r 5
I4 to be exposed beyond -- and it addresses, I think, the issues of 1
=j 15 jl equity.
So, it's very important tha t tha t goal be in there --
l l
16 g
and as we said, properly.
i n
f 17 The large scale fuel melt goal I think should not be
{
18 on the same basis as the individual goal; that is, it's really a
~
r 19 3
subordinate con sidera tion.
That is whether or not the probablity j l
20 ]
of a
-4
-5 larae scale f uel me l t is 10 10 or some other number, i
I 2I is an engineering consideratior. and is not directly related to 22 heath and sa f e ty.
If you do a complete risk assessment, you'll 23 take that into account, as well as the mitigating feature.
So, 24 I think it's important from the concept of having accident pre-25 vention but it should not be accorded the same importance of the i
ALDERSON REPORTING COMPANY. INC.
i
13 5
i sk 14 1
goals on individual risk.
I 2,
MR. SEGE:
Thank you, Ed.
l j
3j Tom Chchran?
j 1
4)
MR. COCHRAN:
Well, first I would say I think that i
{
5l the stcff must address this issue of distinguishing the goals c
j 6 'i from the standards; and that wa s the first issue *. hat I raised E
i 5
7 at the previous meeting and it still is a problem.
I also like E
I g
8' the idea of the distant goal's of Helen's, but the staff must con-O 9
sider very carefully how that's framed in order that one doesn' t i 3,
l y
10 '
give the same sort of perception that one has in this proposed d
f Il time limits for corrective action where one would have intended j
a I
j 12 the appearance that one never has to arrive at these things --
E g
13 that it's okay to operate in an unsafe manner as long as we have
=
n 5
14 this vision of the next coming out there.
t p
15 (General laughter) o j
16 With regard to comparrisons of other technologies, I r
I d
17 am sorry that we didn' t get to Footnote Sl on page 1 which I s
18 i think should have been the subject of discussion.
Not only have
=
19,
the nunbers that we've reen in some sense been derrived, I believe A
i 20]
from the historical date. of other tec5nology rather than the real i l
21 )
competition.
But one is confounding that problem by taking one 1
22 !
little aspect of this technology and making that comparrison --
l 23 aamely, safety and not the whole fuel cycle and so forth -- and 24 which even exacerbates it.
l 25 As I menticned this morning, I think the staff is l
ALDERSON REPORTING COMPANY. INC.
i
)
)
]d 4
I Ek 15
- i still miles away from what -- at least my pe rception -- of j
2 1 what's desirable in perhaps the general public perception in i
I 3l terms of the -- as the way I framed it -- that your 21,000 cancers j
I a 1' i
versus the insurance that we're not going to have any sort of I
5 g
melt-downs.
I think that the staf f ought to go back and re-read 7.'
j 6l the Kemeney report statements about the fact that the Commission G
I 4
- 3 7
1 can't insure the oublic o
that these things are cafe, and that t'
1 8'
6 there has to be fundamental changes in attitude.
And I think this u
9 approach that the staff is taking is going to exacerbate that j
10 >
j problem and not correct it.
=
II I think if you want to correct it and get more safetyj s
12 l
4 1
you have to demonstrate that when you see a problem you go in there
-7 f
13 and rectify it, cand not that you hide it in this -- or set it S
14 aside in yet another sort of methodology called probability risk,
l 7
0 15 b
assessment to justify not doina something about the hydrogen
=
s problem.
As an aside, I want to stand ccrrected; Mr. Kouts x
l 17 j
i informed me that I've made an errregious technical error in j
=
f 18 assuming that stainless steel is better than circaloy; that I'm l
i 19j in formed i t's actually worse.
i" t
But I think that that doesn't set aside the basic l
20 21 l hydrogen problem issue of how one responds to -- how one mitigates 22 I
that p rob lem -- it's just that you can't do it with stainless 3
steel.
The -- if there's going to be some sort of quantitative goals -- first, I would say that there have been two choict s 25 offered; the s ta f f 's choi ce wi th th e q ua r. ;. tative goals and so ALDERSON REPORTING COMP ANY, INC.
15 9
t sk 16 I [
forth.
And some people like them.
And there was.this alternative l
2 h that some of us proposed of getting away from that and sticking l
0 i
3j with qualitative goals and having in the qualitative goals a l
i I
4l statement about the big accidents; and I think that's the pre-i i
5l ferred choice and just back away from quantified goals for the 3
6 present.
I don't think you' re ready for that.
re
=
i y
7[
I think if, in fact, you gv ahead with that, you t'
ii g
8' need something in terms of the implemen tation to insure that u
O 9!
there aren' t abuses of this methodoigy PRA.
In my view, the role 3.
t 10 0 of the PRA has always been that it's a useful tool becnuse the s.
E II i people that build the models and exercise them generally know 4a j
12 more about the sys tem -- because they' ve had to focus in on it E
i
-f13
-- and know more about uncertainties and how they're propogated.
x I4 And if you could just convince them to do that -- and never show l
E 4
{
15 I
the view graphs -- then I think you would be in great shape.
=
j 16 But they want to ;how the view graphs; and then they x
I7 want to establish policy on the basis of the view graphs.
And 3
IO ;
tha t's whe ra I think -- and the view graphs with the numbers down; C
i 19 3
the side.
Perhaps you need to --
i 1
l 20l MR. SEGE:
You will please to notice that the view I
i 21 graph machine has been removed.
22 1 (General Laughter)
)
l 23 MR. COCHRAN:
Perhaps you need to -- maybe you want l
24 d to consider licensing the code; I've suagested that before with i
25 respect to some of the other codes that the NRC has used.
If 1
ALDERSON REPORTING COMPANY. INC.
l
I 16 i
d t
4 1
I sk 17 you're going 7,
use this in a regulatory manner, license the i
2
.I q
code and force some some sort of verification process so that i
1 31 people can't willy-nilly use _hese things tha t really haven ' t -- :
4 that have errors and are used beyond tiie range where you have j
5 m
g some sort of licensing.
If you validate it, they can be used for 7
6 certain types of operations and that's all.
E'_
i a
7i i
I mean, that may not work; it's just a thought.
I 5
8' 9
strongly disagree with Ed's statement that if PRA is used, one a
9i
.j must have PRA goals.
I think if P RA is used for this technology 5
10 i t
in a way that necessitates PRA goals, I think it's more likely g
5 11 :
I y
l that it's being misused.
I agree that there -- with the state-l 12 l
p ments that there needs to be some sort of ALRAR concept; and I l
13 E
would add to that that there needs to be some assurance -- some 5
14 g
assurance, and not just these arm-waving arguments that this is l
5 l
15 r
3 y
not supplementing defense in depth.
But you need to incorporate :
16 l
j some "ery -- something that's going to control that, and insure H
17 l
that in fact this PRA isn't sort of the bull in the china shop.
E 18 '
g And it doesn't supplant supplant the defense in depth C
19]
I approach; and I've seen ncthing except some sort of little state ~
20 i
ments that -- you know, in the back, that you -- you know better 21 3 not do it.
You know, that we won ' t do it -- when I know full 22 1 well it will be; and in fact, I know cases where it is being 23 i
done.
Tha t's all.
24 1
1 MR. SEGE:
Thank you, Tom.
25 Mark Temme.
l ALDERSON REPORTING COMP ANY. INC.
l
3 l
i 17 l
h i
l sk 18 I,!
DR. TEMME:
Well, first I'll continue the chain or l
i 2]
unanimity that's begun here that suggests the principle need to.
l Y
I 3J fix th e document is to make a clear distinction in it between the!
4+
goals and the process by which they're implemented.
And I'd like j
5!
to talk a little bit about what I think about when I'm talking 3
6 7
about goals and process.
e*
7 3
d I think a desirable thing about the goals that are U
8h
=n stated is that they should be empirically measurable; or at least G
d 9i one should be able to determine empirically that they have not j
5 10 '
j been met at some point in time.
An example is zero acute fatal-2 11 l
ities as a result of a nuclear plant accident; to me, it is a 12 '
desirable goal -- and one, in fact, which I think can be met.
13 I
E_
Now, hecause every year or every five minutes -- however fre-I 5
14 E
quently you want -- you can look at the history of tne operation
=
F 15 2
of the plants a:
ask yourself if that has ever happened.
=
l 16 3
And when it does, ;ou have failed to meet your goal.
n 17 d
I think some of these things that probably ought to be included I
=
l E
18 q
as goals are less immineval to clear objective measurement -- buti j
19j they probably ought to still be goals.
That means tha' merely I
l 20)1 compliance with them is subject to some argument.
But I do make !
21 l a sharp distinction between the goals of the sort that I just J
22 l mentioned -- absolute types of goals; like not fatalities -- and i 23 i
i coals which involve words like minimize the risk of, or reduce l
l 24 the probability of because the demonstration of compliance with 25 those two types of goals that are vas tly dif ferent.
l i
i ALDERSON REPORTING COMPANY. INC.
o
i 18 e
i sk 19 I
In the one case, the demonstration -in principally l
2I by observation; in the second case, the demonstration is through 4-g 3
the application of some calculation -- that's what it means when 1
i l
i 4!
it says keep the risk of something below some level, or minimize !
S 5,
the risk of.
And I'm coming to the point, not in small measure 9
h 4
g 6y as a result of these discussions that have been going around, R
4 7
that it's not wise to have as a goal that kind of statement s
]
j 8
because it brings into a very high level the use of PRA; and I J
j
}
9 don't think it belongs there.
E 10 MR. SEGE:
Mark, I'm not sure that I understood that:
E i
l II What kind of a goal shouldn't there be?
9 I
{
12 DR. TEMME:
There shouldn't be a goal that says s
=
i 3
i g
13 !
minimize or reduce to some number the risk of having' fatalities,
=
i l
'A I4 l 2
5 l
or *he probability of having fatalities.
{
15 1
MR. SEGE:
There shouldn't be a goal passed for that
=
j 16,
without speci fying a number; is that what you mean?
i
{
d 17 DR. TEMME:
Because it's extremely arguable whether l
w 1
2 i
w 18 or not you can't wait ten years and look back and what happened a
s i
9, and decide very well whether you met that goal or not.
fou can i
g l
i 20 ;
intent some hypothetical -- rather, pathological -- events that l
j\\
- =
i 21 l clearly demonstrate that you didn' t -- but there's a lot of room Il i
l 22 )
for argument whether or not you did.
And a more desirable kind 4
I 23 '
of goal -- in fact, the most desirable kind of goal to state is j
i 24 ll one that allows you to determine clearly after the fact that i
ti i
25j it's been met.
l I
l-
)-
ALDERSON REPORTING COMPANY, INC.
i 19 j
I 1
I sk 20 In something like zero fatalities or no accidents l
2o that injur large numbers of people lend themselves to that kind 4
l 34 l
a of observation.
But coals that sav low risk, low probability --
4 l
even if they're stated quantitatively -- are highly arguable 5
even after the fact.
After you've said that you've implemented 3
6 1
3 this goal for a time, it's still highly arguable whether or not i
n a
s 7) j you've met i t.
n i
8' l
n So I think it's importarit to distinguish be tween L
9 those two kinds of goals even if you view them both as goals.
.j E
10 j
The process part of this paper should then define what the NRC
=
7 11,
is going to do to meet the goals; and this could include these j
c 12 of quantitative decision rules -- in my opinion, it should at 6
d 13 I
i some level; and it should involve decision rules that use the 14 d
probabilistic analysis; there's not complete unanimity on that i.
=
9 15 i
point, certainly.
l g
~
t 16
?
But there also should be at least an explanation of I
I H.
17 y
the relationship between the goals and the process by which you l
E 18 1 intend to meet the goals; and if the process is explained in seme;
]
=
19 l i
j hierarchy, there should be a clear link from top to bottom.
I l
4 l
20.I think that connection -- that is, the connection between the I
I 21 goals that are stated and the description of the process you 22 i
intend to use is by its very nature going to contain some sub-23 jectivity and some judgement.
And therefore, it is going to be 24 arguable.
25 So, I think it's probably a losing proposition to r
ALDERSON REPORTING COMP ANY. INC.
20 h
n hk 21 I
try to build a tight objective non-arguable link between what i
g I
i 2 ]
you say are your goals and what you say is your implementation l
j l
!I 3 i I
process.
And just to illustrate what I'm thinking about the way l 4
this document shtuld be put together, I would even suggest an 5
g i
outline for you to try to fcilow in which Section I would be about H
1 l
6 l
goals;Section II would be about policies -- and this would include s
a j
7]1
=
statements like we 9111, on a regular basis, reevaluate how we're u
1 8'
x i
s performing against our goals.
We will fix the process if we G
9i
?.
de te rmine tha t we're not performing against our goals -- that type 10 '
of statement shoulel be part of your policy.
k II i The next major heading T would put it decisions:
And e
2" 12 ';
I haven't had time to think through carefully all of what you E.
i f
13 might say there; it could get very big.
But what it ought to do I4 I 4
as a minimum is clarify where you intend to use these quantitative I
j 15 i
decision rules in decisions; and it probably ought to go beyond
=
r I0 j
that.
And then the final section would be the decision rules 17 l
d themselves -- clear down at the bottom of the -- of the last
=
l chapter.
h"!
.l Now, if you go ahead and develop and write decision f
n
?
l 20 j rules that involve quantitative probabalistic risk analysis, I i
2I would add some further cautions and suggestions:
First of all, n
22 l use standard terminology, and use it correctly.
For example, 23 although it's not well understood by a lot of people in t"is 24 room, complimentary cumulative probability distibution function t
25 is a clear singularly understood concept to anyone who studied ALDERSON REPORTING COMPANY, INC.
i
21 I
h sk 22 I
j probability and/or statistics.
l i
l 2]
Mean probability, on the other hand, is not such a j
d i
concept; it means different things to dif ferent people -- and as !
3 6
4 far as I know, it means virtually nothing to an expert in probab-a g
59 ility or statistics.
So, I think you need to get that kind of S
0 stuf f straight; and in all of this I refer to decision rules, and E
i7j what I think is very important there is to spell out well the n
i 8'
n rules for applying the PRA.
Tom Cochran referred to something I
9-
}d like licensing the code; well, that 's virtually what I suggest 10 !
3 here.
3 E
4 i
II When you lay out quantitative decision rules invol-I ving probabilities, you should come pretty close -- I don't think 4
g 13 you have to license the code -- but you have to confine the user 5
14 1' g
in such a way that he can't play a lot of games with the numbers.
e0 15 b
Steve Derby referred to this as writing an engineering standard;l l
16 i
to.:.e, that's what it means.
l n
17 Part of the discussion of the use of ':.hese decision 2
d
]e l
c f
18 l,,
rules should be some clarification of uncertainty and what we
)
i i
i 19 j mean by it and what we do about it.
And the page in this docu-l n
a 20 !
ment -- or two -- diat discusses uncertaintie s seems to use j
s Y
21 l several different definitions of tne werd at dif ferent parts of I
l 22 the page.
Vojin spoke to one kind of uncertainty in his 23
{
de fense o f the concept of probabalistic risk analysis; and I 24 agree with that.
We do have uncertain knowledge of physical 25 p are me te rs, the future behavior of component systemsandsoforthh 1
1 ALDERSON REPORTING COMPANY. INC.
i
l e
22 i
Dk 23 I I And P RA is a legitimate framework in which to put 2l those uncertainties and see what they do to us.
I think there 3li is another kind of uncertaintv that's in the background of all t
4I of our discussions here ; and it's kind o f an uncertainty about i
g 5l how good the PRA calculation is.
And I would try to separate N
j 6!
'hcse two kinds of concepts -- talk about uncertainty when you 1
-n s
7 refer to lack of knowledge about the wo rld; talk about the qual-M j
8l ity of the P RA when that's what you mean.
And you need to G
9 9
devise some system for determining the quality of th e P RA th at ' s
?
E 10 l submitted in compliance with the decision rules.
z E
11 l I had another note here that I made during this 43 j
12 I discussion that was just going on.
We we re talking a lot about E
I f
13 i relevant comparrisons.
And I th ink in the context of probabal-
=
z 5
14 '
istic decision rules we might conside r th at a use ful relevant t
15 {
a comparrision would be an existing nuclear plant because if you r
=
y 16 l can get done on an existing plant the P RA that's done following i
d 17 the rules that you intend to use, then you have a measure of a=
5 18 '
whether your decision rules are in tended to achieve equal safety,'
19,!
=
I
~
less safety or greater safety th an tha t plant.
I A
l 20 Now, this is not the kind of comparrision that l
i 21 '
justifies goals, but it is a re levant -- pos s ib ly -- a re le vant 4
22 )
and useful comparrision for devising decision rules which are l
23 in tended to implement rules.
I guess that that -- with regard l
24 to what might be done -- at least in my mind -- to fix the l
l 25 document, that covers it.
I have one more s tatement that I would i
d 1
ALDERSON REPORTING COMPANY. INC.
l 23 kk 24 1
like to make.
I 2
I think we need a serious commi tment also to fix 3,I prchabalistic risk analysis.
I think a lot of the concerns about ;
i I
4 it are legitimate; but that doe s not mean that we should throw I
i 5:
it away.
I think it's a good tool, and what we really need to e
n n
6; do is devise clear rules about how we use it, when we use it, h ow
~
o
-3 S
7l to use it well -- well, I guess that's really essentially all I
~
l
~n I
i 8
can say about that one.
n d
d 9i But I think -- Dave Okrent raised a question yester-Y 10 day; do we need a lot of research to get -- make -- P RA be t te r?
z=
t E
11 ;
I think we need some method of development -- methodology and l
B d
12 !
method of development.
What we need more is agreement; I think E
i E
E 13 that's the dominant problem.
And I think you ge t it by laying E
y '14 1
out a way to do it and trying to get people to go along with it.
E 15,
MR, SEGE:
Thank you, Mark.
s I
g' 16 l Helen Ingram.
i l
d' 17 MS. INGRAM:
It's bad luck to follow someone with E
18 such care fully f ormulated commen ts.
It seems to me that the staff l
l s
?
19 '
could feel good about the extent to which we agree about what's 20 the matter.
21 !
(Gene ral Laughter. )
0 l
22 j MR. SEGE:
Thanks a b un ch.
i i
23
.\\E : INGRAM:
I mean, what I have to say sort of l
24 j folicws what other people have had to say -- and I' ve tried to
{
25 ;
look at the documant and see places in particular to be fi xe d.
1 ALDERSON REPORTING COMPANY,INC.
l
Il i
24 l
3k 25 1
And then Mr. Temme comes along and suggests a new outline.
I 2!
don ' t have page numbe rs anymore, but let me suggest that if this i
i 3!
is the document that you move f rom towards something else -- I' m !
4 a teacher and I know once you' ve committed your ideas to paper g
5 li
-- st uden ts neve r comple tely change -- you know, if that 's the H
j 6!
case he re.
R l
h 7
Then in this business as a purpose of policy statement j
8 !
and you say that a goal -- and that's -- I think what you mean by a
9l a policy s tatement -- should serve the following cbjectives ; it O
,zo e
10 :!
seems to me tha t we ' ve talked about your adding considerably to z
't i
j II those objectives and making them much better de fined and discreet 3
i N
12 with the notior. that dif ferent goals serve di f ferent audiences,
=
13 i o r dif fe ren t purposes.
And there seems to be some of these t
=
l m
5 14,
that one function of a goal is to be an aspi ra tion ; some thing I
E 15,
we should move towards.
j 16 I An o the r f un ct ion -- an d th a t ' s really very important z
17 for the public at large to get the sense th at that's what our u
f g
18 va lue s are; that's the bette r world we ' re trying to move towards.
t I
n 19 I f another function of a goal is to be a standard to be used in n
a 20 )
the regulato ry p rocess, that's dif fe rent.
I think another goal 21 could be the understanding of the public; I don' t think we have 1
i 22 ]
a consensus on it, but it's my idea that we should add in purposes 23)l that these goals ohould represent and articulate public concern i
24 ;
-- as well as educate the p ublic.
I 25 I think you' re going to get in trouble with that word '
1 i
ALDERSON REPORTING COMPANY. INC.
i
5 25 i
s sh 26 1
unde rst anding or education -- because it sounds patronizing 2
and people won ' t like it.
So my advice, as a politically sen-i 3
sitive pe rson to you, is that you should use the word represent, f I
4l articulate -- and that needs to then be connected with the process 5l that you really do connect with public attitude.
g O
j 6
All right -- so if you worked hard on that part of G
7 page 3; I think the rest of your job gets easier.
And the next I
j 8
p lace where things get pretty rough is on page 7 where you talk a
9!
about the deve lopment of a policy statement, and here it seems to z
O i
y 10 !
me that you connet. most of the social and political and equity 5_
j 11 p roblems that you j ust can' t deal with that way ; you' re going to j
3 j
12 ;
get caught.
Much of what Paul says here is re le van t.
~
j 13 I don't think you can say -- as I ' ve s -id be fo re --
=
z 5
14 that these tbings are self-evident.
I think that you should b
R 15 instead say that these things are not self evident; the re are E
3 matte rs of grave controve rsy.
And there's a continuum of atti-16 A
d 17 tudes that range from here to the re, and then you want to say a=
18 l that we come down here fo r these reasons -- p robably you should l
=
19 l do so.
But you have to tell people that you know that you could a
20 come down dif ferently.
j
- l 21 !l All right, then, what are the issues?
Risk ave rsion:
ll 22 1 I think you have to address the q uestion of risk ave rsion, and l
4 23 i perhaps you have to recognize that the public believes th a t i
I l
24 !
nuclear ene rgy is di f fe ren t ; that it is a large low-probability 25 eeent -- ve ry la rge consequences -- with great un ce rt ai nty.
And i
i ALDERSON REPORTING COMPANY, INC.
l
y s
26 i
i ek 27 1
if I understood Toby correctly, it's not e ven irrational for 2
,l people to particularly be risk ave rse under ce rtain circum-f 3J I
stances.
So, I think you could make a fine rationale for risk i
?
4.l; ave rs ion -- not s imp ly s ay well, the public's crazy; but in fact, 5l
.n; i
this is the way they feel, and this is part of our goal.
?
3 6
2 All right; then the question of bench marks :
You u
n 7i j
can ' t j ust throw in the notion that it's nuclear energy and it S
8 9
ought to be as good as anything else.
At the ve ry least, I think u
d 9!
j you have to say -- take Lester Lave 's position; and that is that E
10 !
E it's got to be as good as any other competing means presently
=
1 11 l l
avail ab le.
I t ' s go t to be new coal plants, not existing coal 1
'i s
12 ;
p lan ts.
That's not good enough for me personally; I think that i
13 5
it ought to be much better -- because people feel the way they S
14 d
I do about it -- and perhaps this idea o f Mr. Temme's that you 9
15 g
use as a bench mark existing nuclear plants, and the degree of l
i
~
16 d
safety available the re is something that you ought to at least I'
h 17 0
consi de r.
=
i 18 g
j The equity issue we talked about ve ry little, be causei I
19 3
we just didn't ge t there.
I think it's not that we ' re not con-20 !
ce rned ; it's j ust that we didn't cet there on the agenda.
The re I 21 is a reasoning there that ecuity disappears if the risk is small ;
.I 22 ]
enough; but we have no agreement here that the risk we have 23 i
talked about really is small enough.
It mi gh t be small enough i
24 4
If our aspirations, in fact, were re ali z e d -- i f we had this goal.
t 25 that the re be a ve ry low risk.
And i f the ve ry low risk we re, int ALDERSON REPORTING COMPANY. INC.
g
?
27 Ok 28 1,
fact, the real world -- then we do r. ' t need to spend a lot of 2,'
time and effort.
t 3!
And perhaps you could say that -- maybe you need to j
i 4
say as an interim goal where we have some operational standards; s
S we ' re, in fact, we ' re working unde r at p res en t, the risk may be en 3
6' so maldistributed that with some people the re are real equity G
3 7
j questions.
At any rate, you' ve got to do something other than s
i j
8l put that little footnote the re which says that if the risk is' 4
9 9l small enough you don't need to bother with it.
I think that's
?
10 l only going to arounse critics.
z=
j 11 l All right; another thing is about the scope of nuclear 3
i f
12 j e ne rgy.
Todd LaPorte has made a strong argument that as nuclear 5
y 13 -
e ne rgy, should it be come a larger part of the market, the nature
=
14 and the magnitude of the risk change.
That 's nowhere in this s
i 15 do cume n t.
It is such a common argument that e ven I,
outside this 2
y 16 '
c l ub, have heard of it.
So maybe it needs to be addre s se d.
A d
17 All right; then the business about the goals them-G 18 se lve s :
We ' ve talked a lot of quantitative and qualitative
=
l 6,
p i
19 l goa ls, and I'm not sure that's a useful way to think about it; A
20 '
instead, different types of goals for dif ferent purposes is 21 ;
p robably bet ter.
And ce rtainly, for an aspiration of qual; tative j l
22]
goal is de fini tely it, you know.
i 23]
On th e o th e r h an d, if you' re going to have a standard!
f 24 l maybe you move more toward -- or if you' re goinc to h ave an l
25,
ope rat ion al goa l, maybe that gets closer to being quantified.
Bui i
l i
1 ALDERSON REPORTING COMPANY, INC.
I
l I
1 l
28
)
i 1
sk 29 I '
f that division has dominated our debate ; and I' m not even sure f
2 j
very use fully.
So maybe you ought to avoid talking about qual-d
. 3 's itative and quantitative in that kind of adversary-sort of a 4I fashion.
Maybe you should articulate the goals in line with the y
5l p urposes and the di f ferent kinds of goals.
m j
6 !
And then implementation -- which now really concen-E I
7l trates only on PRA -- ought to cover all of these things.
If you a
7.
I i
8 re have an aspiration, how do you move toward the aspiration?
If 4
t 9,
~.
you have a goal of education, what is the process by which this
?
\\
C 10 l g
actually se rves?
If you' re going to use this document, then, as
=
f the basis for what comes next, it seems to me what we've said say:s 12 E
that those are the places in the document you need to concentrate I
13 '
your e f forts.
5 14 j
MR. SEGE:
Thank you, He le n.
0 15 i
b Ellen Weiss.
i
=
i
~
16 s
MS. WEISS:
I just have one brie f comment :
I put my n
W 17 j
card up af ter Mr. Ko uts gave his summary because I was s urp rise d I0 I
that it seemed to be ass umed in what you said that the product c
1 19 s
of these meetings would be a redraf t that assumed quantitative 20 1 1
goals and varification of those by PRA to perhaps some change in i
I 21 the n umbers.
We debated that a lot, and I don't ha ve any in te n-1 22 :
tions of opening up the issue again; but I think that it 's clear t
23 that there was a strong feeling around the table -- quite l
24 i
4 i
p os s ib ly, the maj ority feelinc -- that cuantitative goals a ren t 25 app rop riate right now.
And I would hope that that would be i
ALDERSON REPORTING COMPANY. INC.
q e
a 1
29 sk 30 1
re flected in the formal report.
I 2l MR. SEGE:
Thank you, E lle n.
I 3f I think Mr. Whipple has had his card up fo r s ome 4h time.
I 5
MR. WHIPPLE:
All right.
g S
5 6,
Maybe it's a re flection o f the to which our viewc G
7' have conve rged somewhat, but I ' ve che cked of f a lot of issues j
8l as these other people have discussed their comments.
So I have 4
9 9l just a few le f t.
x0y 10 cne is that I think we agree within certain bounds 11 j
l that P RA has a lot of value for understanding the technology, u
j 12 and particularly in a dynamic sense for making us2 of the inform-3j 13 i ation that comes in from the operating experience towards knowing
=
Z i
14 i what the risk le ve ls are and certainly more impo rt antly, from 3
_b j
15 knowing what the important risk sequences are and identifying j
E I
j 16 opportunities to reduce risks.
So for that reason alone, we ' re A
(
17 going to use P RA; it would be foolish not to -- the question is e
Cw 18 i how we do it.
=
?
I 19,
An d I can't see how one goes about using PRA without R
20 '
ultimately coming down to a quantitative goal either explicitly 2I stated -- as in this do c umen t -- o r imp li cit from the results of l 1
f 22 i individual decisions mace in specific cases.
Again, that 's a t
I i
23 di f fe ren t approach from one startinc with a soecific goal and I
i 24]
thinking of it as a licensing basis ; thinking that we have to I
l 25 demonstrate that the plants meet that le vel in order to be i
i ALDERSON REPORTING COMPANY. INC.
l
i l
30
- 31 1.i acce pt ab le.
I 2
But none the le s s, I think on that more limited basis i
E i
3 the value is pretty great.
Of the issues that I would like to i
f 4
see included in greater detail in the document itself, one issue i
I e
5 Helen has brought oct repeatedly and quite clearly -- and that nN 6'
ie the issue of whether one uses a normative decision analysis --
U i
5 7
.i cost-benefit, cost-e f fectiveness based -- approach in logic; or 1
l f8 I whether used as one uses a more empirical approach based on G
l 9!
attempts to include public values, perceptions and political z
y 10 t re alit ies in the app roach.
z r
=_
i 11 I
p l
And I have litt le do ubt that if goals come out,
B j
12 f they will re flect some sort of a balance between these two con-13 !
j ce rns.
But I'd like to see the dif ficult job of describing how A
5 14 each of those factors was considered included explicitly in the R
15,
text.
I think that's an important issue.
e
}
f 16 On a related point, I think the rationale of the i
d 17 comp a ritive risk statement needs to be expanded and made cle ar; 5
18 l and I think that is the point at which ':he ec uity iss ue can come ;
,i
-s i
19 in -- I think that relates to the comparative lo ci c.
I do have l
9 l
20 a concern with a p remise that I' ve heard throughout the two days i
21
]
of this workshop that the goals will be approximately around the l
1 22[
reactor risk le ve ls -- give n the -- you know, t o the e xtent that f
t
.i i
23 the un ce rtainties pe rmit.
i 24 1 And I think thac re fle cts insuf ficient attention to 25 the importance of an ALARA criteria.
I think that without 4
1 ALDERSON REPORTING COMPANY, INC.
l Ji 1
32 I
knowing what the trade-of f curves are between cost and risk --
i 2'
and without having a propose' ALARA crite ria -- we can't have any li 3
idea how these goals will affect reactor safety. And I guess my l
4l hope is that the ALARA will be the driving force in safety deci- !
l g
5l s ions mo re than some overall ceiling on risk based only on equity !
W j
6!
That is, I hope we meet equity at a minimum; but I also hope th at R
i
=
S 7!
actual reactor risk levels are well below that simply because U
g 8l cost e f fective systems can be found to bring it to that point.
U
~.
9l On the issue of bringing values -- public values --
3 10 into the goals, I think it's impo rtan t to do so; but I also z
i h
II l recognize both the complexity of representing public values an d a
12 '
2 the difficulty of expressing them in an easily comprehensible way, E
r j
13 '
Now with some specific issues :
On the question of intertemporal
=
z 5
I4,
equity, I ' ve found that the approach in rhe docurent it's 15.
simply somming early and latent f atalities -- seemed to satisfy
=
j 16 '
e veryone here.
l N
I7 MR. JOKSIMOVIC:
Not me.
w I
I 3
IO i
MR. WHIPPLE:
Not you -- but it satisfied most people!
i s
19 and I think that may be a worthwhile simplification.
The is s ue
=
20 of risk aversion I think needs a more detailed tre atment, but I 4
]
am ve ry much enamoured with the approach Paul mentioned of using 21 22 the unce rtainty as the basis for dealing with aversion.
Doug i
23 l MacLean made the comment that risks have dimensionally dif ferent i i
24j q ualit ie s that are important to people; andI agree with that.
I I
25 i And I think the re 's a historical recocnition of 1
i ALDERSON REPORTING COMPANY. INC.
l i
3~9 4
1 33 1 f fact in the relatively high levels of cost used to reduce risk 2[
in the nuclear area; and in the NRC's emphasis on public inputs Il 3 !l and in the fact that nuclear risks are, in fact, managed dif-f 4
ferently in this society than most risks.
They are managed with i
e 5
more public focus on them.
i I
m.
i i
j 6i At the p resent, I don't see how to go much beyond L'
I 5
7l those kind of simple platitudes in evaluation; I don't think that i
j 8'
the goals proposed in the document necessarily igno re va lue s.
O i
9l But I would like to see more attention given to a discussion of
?
10 !
the values themselves -- that is, given the uncertainties that are d_
I i
j 11 l present in all of these factors, I 'm not s ure I wan t to put in 3
[
12 '
complicated f actors including alphas in them; but I think it's E
i g
13 important nonetheless to talk about risk aversion in some detail.
=
11 5
14 ;.
Finally, the question of learning to use the goals:
_=
i j
15 I don't see that we have any choice other than attempting to i
I I
/
j 16 i start wi th the existiag system and try to roll the' use of quan-i d
17 4 titative goals in on top o f it.
And while th at may tend to a
M 18 l enshrine perhaps past practices, I don't see the alternative of l
=
C 19 '
going with an abrupt ch an ge in approach as a possibility.
And 6
}
1 20 that in that sense, Mark Temme 's point about usi ng existing l
h plaques as existing bench marks makes a nice co nne ction.
't 2i 4
l 22 3 And I think it 's e nice connection with Helen Incram's l
~
23 j
earl 4_e r point when she was speaking of a qualitative order i
24j quality goal o f pe rhaps we ' d like to do a litt le be tter each l
25 ;
year.
I think that's something that would app ly to nuclear power i
ALDERSON REPORTING COMPANY, INC.
i 33 I
I sk 34 I
plant safety as well.
2l I ' ll q uit there.
I 3l MR. SEGE:
Thank you, Chris.
I 4i' We have approximately 5 minute s le f t; I see 4 cards i
5 j
up.
If each of the 4 people who have their cards up will limit themselves to 1 minute, I would appreciate it.
An d in th at way,
n
- D 7
l there will be room for one more speaker for one minute af ter R
S 8
5 that -- so le t ' s just go around the table starting with Dale Y
l
~.
9l M
B ride nba ugh.
5 0
j 10 l MR. B RIDENBAUGH:
Okay.
- j. IIl!
E I' ll go as fast as I can:
I certainly would endorse 12 Helen Ingram's idea that a goal should be considered to be used 9
I as a level of aspiration.
And I think along -- if you choose 3
14 i that, ycu definitely need some specific language in the goals e
9 15 0
on ALARA -- and of course, we've talke d about that.
Another
=
~
16,
possible way of working with that system would be to -- picking B*
u 17 '
d up on what Tom Cochran says -- perhaps license the standards 5
18ll
_n that derrive out of the qualitative goals.
I 19 l!
s I
5 I think that, as we've stated, we ought to hnld on n
20 ',
t o the concept of a goal that essentially avoids the pcssibility
'I 21 i i
o r the lik lihood, at any rate, of a large accident.
And a la rge j 16 22 J l]
accide nt should be defined as not only including the loss of I
23 I I
l 1
life, but also the prcperty and natural resources as well.
I t
24 3 q
- Finally, I think that as Helen said if you can t
25 get much mo re s pe cific on p age 3 as to what is the purpose of i
i ALDERSON REPORTING COMPANY. INC.
l
I W
34 l
sk 35 1
this whole goal -- I think it. ould help tremendously; and to 2i get rid of the general statements tha t say a policy statement is O
3 l intended, and say this policy statement shall recognize that this i
4' is a working paper.
I think that's needed for the next ve rs io n.
s 5
MR. SEGE:
Thank you, Dale.
n I
ei h
6 Vojin Joksimovic.
E i
5 7
MR. JCKSIMOVIC :
In view o f this 60-second limit, 7.
8 I'll be issuing statements in the forms of telegrams.
d k
9 (General Laughter) z O
i y
10 i The issue of integrals ve rs us limit lines has to be l
5 11 addressed in the document.
As you know, in mv cpinion that's a M
N 12 ve ry fundamental issue -- publication of the goals.
P reven tion E
i j
13 mitigation philosophy, in my opinion, has to be incorporate d
=
w i
g 14 !
into the goals.
External property damage goal ought to be con-3 z
15 '
J:
s ide red ; an d I though t the re was a consensus on that in the Palo j
16 '
Alto meeting, and I haven' t seen anything in the document on that!
k i
l p
17 With regard to the quality of the PRA studies :
Since !
$}
18 that aapears to be a concern, I think the document has to address;
~
"_ 19,
the i eue of acceptable methods for pee r reviews.
Aspirations in!
g f
20 P I
the f ut ute : I think ought to be addressed by virtue of distin-21 l guishing the goals for future and current generation of power l
1 22l plants.
Otherwise, when the document is fixed, I urge the 23 Commission t o o urs ue its acgressive course in de veloping quan-i l
24 i titative goals on a trial basis.
25 MR. SEGE:
Thank you, Dale.
I a
ALDERSON REPORTING COMPANY. INC.
l
l 4
V 30 ek 36 1 '
Steve Derby?
\\
I want to emphasize the notion of a 3
hierarchy in the goals in the standard.
I think that it has to 1
4l be made very specific in the report and be very systematically e
5l unfolded in general view of the operation of this technology in n
+
n I
3 6
this country as an industry and a said plant to what that means 6
5 7li to a specific plant in a specific local population.
I also think 7.
8!
that there has to be some description or acknowledgment of the d
1 0
9!
hazard that is trying to be cor.t ro lle d ; from what I s aw S un day z,
i O
y 10 i
-- yesterday evening -- that that hazarc -- once a population z
=
j 11 ce nte r is e xpose d o r is in the p a th o f a p lure, that at least 2
g 12 f the consequences are comparrable to commercial airline crashes.
5, i
y 13 we do not have accidents that kill one or two people ;
z g
14,
we have accidents that either do not kill anybody immediately, or
{
15 kill several hundred people.
I think that there has to be that i
.f 16 j acknowledgment if that is true -- to be spe cific about what you z
17 are trying to avoid or control.
3=
w 18,
MR. SEGE:
Thank you, Steve.
i E
i h
19 f Doug MacLean?
n 20 l MR. MAC IEAN:
I can do this very q uick ly :
21 l
One; agai n, this is for my hopes for what the revision i
i 22 '
of the ?.eport will contain -- and one is much more attention on 23 the way the quantitative goals -- assuming that they ' ll be con-i 24) tained in this -- are related to the qualitative goals.
I think I t
25 that this will be impo rt an t for public discussion.
And se condly, I 1
i
\\
ALDERSON REPORTING COMPANY. INC.
t il 36 l
Ik 37 4
1 also important for the public discussion -- 1 would hope th at 2,
whatever quantitative goals are listed are de fended explicitly e
3 against the objections to having them listed at all; both the i
4l objections to how they might be imp lemented or misimplimer.ted 5l if the conside ration is that they ' re p rema ture -- that all of e
~
e.
8 6
these are somehow included in the statement be fo re it goes out a
t 7
for public comment.
j 8'
I guess finally that -- somewhat more gene rally ;
Ud 9i I see that all sorts of pessibilities are opening up as a res ult
?.
I 10 of thinking in these kinds of terms.
I ' m thinking of Paul Slovic't 6
l 5
ij ;
comme nt th at goals should be thought of in a constitutional 5
i s
i 12 l sense; and maybe when nuclear issues are ove r, we can all x=
h 13 l reconvene to consdier rew riting the constitution -- and wonder E
S 14 i if whethe r li fe, liberty and property are, if not inalienable, at least i
e E
l
-6 l
2 15 be p ro te cte d t o 10 j
w s
x 13 (Gene ral Laugh te r) n p
17 MR. SEGE:
Thank you, Doug.
x 18 '
Are the re any othe r concluding rema rk s ?
I'm sorry, I
=r il 0
19 Cla rk ; your card was hidden behind Vojin's and I di dn ' t notice itj 1
?
o 1
l 20,
MR. GIDBS:
I just h a ve th ree rathe r simp le sugges-t tions; o"e, I feel that the documen t could be improved by including 21 t
t 3
22 ]
norre kind of discussion of both the co nsequen ce s of action and 23 '
inaction.
We know what the co nseq uen ces of not having had a 24 !
sa fety goal has been to date by our present state of af f airs.
25 To what e xtent does this rep re sen t a continued dif ficulty ; to I
i I
J ALDERSON REPORTING COMPANY, INC.
I
l
~
37 Ek 38 I
what extent can it be imp roved upon?
l 2
I also think that this s ame type of assessment co uld 3I be made fo r e ach o f the goals ;
i.e.,
i-appe ars to me that ce rtain i
l 4'
of the goals are more important th an othe rs.
And the consequences e
5j of either having or not having a goal on each of -- for each of A
i j
6; those items -- I think could be articulated and would contributa R
S 7
to the content of the document.
j E
j 8l Se cond ly, I would suggest that the document be ex-0 1
t 9;
5 panded to include some additional perspective.
Specifically, what 10 are the competing risks that are being evaluated and used as the z=
i j
11 basis for the establishment of the goal; and to what extent do a
N 12 f those competing risks rep resent a genuine competing risk --
i j
13 s pe ci fi cally, the old volentary versus involuntary exposure type
=
- n 3
14 ;.
argument.
If the goals fit into one category or the other, to b
E 15 what e xtent is the placement of the goals in those categories E
\\
j 16 s upporte d?
A
{
17 Fin a lly, I would like to suggest that the document be i
E 18 explicitly constrained to apply to light water reactors ; and my I
g i
19 re as on for that is that within the resource base in yellow cake 5
i 20l in this coun t ry, there is a fixed number of light water reactors l 21 that we can build.
The re fore, once that resource base has been i
i l
i 22 i dep le te d, any other tyne of machine will have to be some thina l
23 di f fe rent.
Knowledge of that domes tic resource base, and its 24
,i translation into a number of actual operable re actors, can in 25 turn shed light on the question of to what extent the goals as l
1 I
I ALDERSON REPORTING COMPANY, INC.
I
~
i 38 i
sk 39 1
established will represent a total risk throughout the usage 2!
of the concept.
t 4
3p I would expect that a different stancard would 4'
probably apply to a dif fe rent type of re a cto r.
[
i i
g 5l MR. SEGE:
Th ank you, Cla rk.
O j
6l I guess this concludes our general discussion; thank U
I 6
7l you ve ry much. It has been a ve ry help ful session indeed ; and I
~
U i
j 8'
appreciate all the ve ry good contributions that have been made d
?,
9l to this Honor Court of discussion.
I guess I will turn the Chair 0
10 '
back to our Gene ral Chairman.
t E
11 l Thank you.
j d
j 12 i MR. RATHB URN :
Thank you ve ry much, George.
l 5
I g
13 '
And I would like to thank each of the panel chairme n z
5 14
-- Tester and l aul and He rb -- for their fine work here -- and I
~~=
i j
15 l each of you for your comments and suggestions.
We ' ll go back l
^=
16 and do an awful lot mo re work.
.j w
d 17 Thank you.
x 5w 18.
(Whe re upon, the hearing was adjourned as described above. )
1
=
19 :1
_5 1
i 20 I
i 21 22,
i i
23.
i I
i 24 25 ALDERSON REPORTING COMPANY. INC.
~
s NUCLEAR REGULATORY CO!dMISSICN This is to certify tha: the at: ached proceedings befora :he k
in the ma :er o f
- SAFETY GOAL WORKSHOP - SESSION 6 - GENERAL DISCUSSION
- Da e of ?roceeding:
July 24, 1981 Docket !!umb er:
? lace of ?r0ceeding: Harpers Ferry, West Virginia were held as herein appears, and tha this is the original transcrip thereof for the file of the Oc==ission.,
Rossie Sutttn Of ficial Reporter (Typec)
+
f 2A2 Official Reporter (Signature)
' b.
-