ML20010C439
| ML20010C439 | |
| Person / Time | |
|---|---|
| Issue date: | 07/24/1981 |
| From: | NRC |
| To: | |
| Shared Package | |
| ML20010C401 | List: |
| References | |
| NUDOCS 8108200081 | |
| Download: ML20010C439 (55) | |
Text
~
l i
I i
UNITED STATES OF AMERICA I
l NUCLEAR REGULATORY COMMISSION 2
3 4
PUBLIC MEETING I
e 5
SAFETY GOAL WORKSHOP h
8 6l SESSION 5 m
7 IMPLEMENTATION j
8l, a
d 9
Cliffside Inn z
b 10 Harpers Ferry, West Virginia z
11 Friday. July 24, 1981 d
12 The meeting was convened, pursuant to notice, at z*
I E
10:35 a.m., L.mLave,nThe Brookings Institution, presiding.
5 13l1 E
14 l PRESENT:
a l
E 2
15f D.
RATHBUN E. O'DONNELL s
.T 16,
L. LAVE D.
BRIDENBAUGH s
i A
\\
1 g
17 MR. ERNST S. DERBY 18 !
H.
INGRAM G. SEGE 3
l t
19 !
D. MAC LEAN T.
PAGE 5
8 n
20 V. JOKSIMOVIC i
21 '
G. COREY 22 i T. COCHRAN I
23 E.
WEISS 24,
H.
KOUTs i
l 25 N. WALD i
8108200081 810724 ALDERSON REPORTING COMPANY. INC.
l 1
1 EEEEEEE1EEE I
j 2
MR. RATHBUN:
We are now in Session 5, in which we i
3 will discuss implementation issues.
Lester Lave will be the
-4 chairman for that particular discussion.
However, before I e
5 turn the floor over to Lester, Mel Ernst from the Nuclear E
i n
i, 8
6 Regulatory Commission Reactor Licensing staf f has a few minutes o
3 R
R.
7 worth of comments that he would like to make on the subject of Aj 8.
implementation.
I do 9
I know that this subject has been of great interest Y
10 to a number of you.
z=
5 11 Mel?
<3 6
12 i MR. ERNST:
Well, I did..a lot of thinking last night z
i 5
i d
13 af te r the session yesterday.
I thought the session yesterday aa i
i E
14 was good.
I think as intended and anticipated there was, in a
E 2
15,
Bob Bernero's words, I guess, very little arm waving about PRA s
j 16 l yesterday, an awful lot of hand wringing, and very appropriate
^
t d
17 '
hand wringing, I think.
E l
18 l I got some sense by 7:00 o' clock last night that
=
i b
19 !
many people in the group were coming to the conclusion that I
I 20 }
quantitative goals were premature because of the uncertainty on I
21 ;
PRA.
My personal view is to express caution,that perhaps such l
22 l a course could be in some ways permitting the emperor, so to i
23 speak, to ride the streets maybe partially clothed.
24 My perception is, as Dave said yesterday, that the f
25 staff is currently performing and reviewing PRA studies and l
t ALDERSON REPORTING COMP ANY, INC.
3
~
o
\\
1 l
I I I using the results of such studies as an aid in assisting the 1
2l decision making process as to whether certain fixes perhaps need I
3l to be applied to nuclear power plant systems to improve their i
i 4;
reliability.
5l So therefore unless we suddenly outlaw the use of the e
E 3
6 word " probability", say, in regulatory matters, the staff will E
i s
7*
really continue to do that kind of, and I perceive it to be l
~
M j
8 useful, exercise.
ao 9
It also appears to me this group supports the cautious Y
10 use of PRA in the licensing process or in the regulatory process, z=
i G
11 i therefore I would like to say that we did consider the uncer -
5 I
i d
12 tainties in PRA in trying to formulate a draft implementation z
,=
i i
E 13 l process, albeit there does need to be some flesh put on the E
l
{
14 l proposed implementation process, but the concept is still as we b
l 15 '
currently see it in the draft paper.
=
y 16,
I think I would like to take just a minute to sor' of w
f 17 outline the proposed concepts in the paper to hopefully bring w
=
i 5
18 l the group towards further constructive criticism as to our E
19 l proposal, recognizing that PRA really is not a new art form, 5
20{
PRA really is an attempt to better quantify reliability:of j
i 21 systems and resultant risk to the public, to try and start 22,
putting some numbers around what we used to call good l
23 '
engineering judgment.
f 24]
It is not the intent to replace good engineering l
1 25 l judgment but only to augment.
l
}
ALDERSON REPORTING COMPANY, INC.
i
4 o
o
=
i 1
1 What is proposed in the paper in a draft sense is that i
2i NRC should specify the requirements for the use of PRA in both l
3l the analytical made and in the decision making mode.
The l
4l primaryilicensing basis would still be the existing' rules and l
l c
5 regulations.
It would not be some, as we shorthandedly use, some 2
6; CCDF curve of something like that, some risk curve.
It would e
R l
8 7l still be the basic rules and regulations.
Therefore, PRA would j
8l only be used as an adjunct to our current way of regulating.
'J l
9l What is proposed is, generally one could use the PRA
=
i l
c i
g 10 ;
results in a realistic way, realistic instead of very z
=
i j
11l conservative calculation, to augment our perception of the 3
d 12 ef fectiveness of our regulatory process, and where this E=
1 N
13 realistic or best estimate use of PRA indicates that some 5
5 14l improvements could be made in the reliability of systems, we would w
2 15,
so do.
a 1
7 16 However -- and we would also use PRA in this manner to s
A p
17 prioritize safety issues and research and issues of this sort.
E 18l However, what is proposed in the paper is if PRA is used to 19 l dismiss existing regulatory requirements, whether this be 5
i 20 l generically by changing the rule or dismissing an extant rule, 1
21 !
or whether done case by case to permit a plant not to comply; h
22 '
perhaps with a particular regulation, such use of PRA should be l
23 done in a conservative, an appropriately conservative manner.
24 1 So, the intent is to use PRA, which we think is a very j
i i
25 effective analytical tool, in such a way to augment where ALDERSON REPORTING COMPANY, INC.
g 5
4 f
y appropriate and improve the safety of nuclear power plants, but i
2j not to throw out the system as it currently exists.
I just wanted 3, to have a few minutes to come back and provide seme perspective I
?
j 4l on what we are trying to do.
Thank you.
l j
5!
MR. RATHBUN :
Thank you very much.
e 9
j N
6 Lester?
o i
l
)
7 MR. LAVE: I would propose to deal with two questions l
3 I
8 8,
during the hour and 20 minutes that'we have.
The first is to a
l i
i d
j o
9 spend some time on quantitative safety goals, since we have done 1
i i
I h
10 very little of that during this workshop, and if there is some z
i 4
=
2 11 l considered opinion, then we ought to sort of get to it.
The a
B I
d 12 l other one is to talk about implementation issues, and let me begin z
=
i d
13 l with the quantitative safety goals, a nd try and summarize what m
i j
E 14 !
I think that I've heard over the past day and a half.
l 6
2 15 -
The first, which came out of our panel, was the i
a.
2 j
16 l observation that quantitative safety goals don't mean anything l
l p
17 unless you have some accepted techniques of calculation and w
x
\\
18 l verification.
That is, it doesn't make any sense to simply put 4
i P
i E
19 ' up numbers and then wave at them without some way of getting at A
4 20 these things. I certainly think that is true.
l 21 !
The second observation had to do with a vast range of 22 uncertainty with respect to actually trying to calculate what 23 were the safety levels that had to do with any particular
)
24 reactor, and there I have been trying to ask people around about 25 j the relevant numbers, and a summary of my understanding of that, f
1 2
ti ALDERSON REPORTING COMPANY,INC.
i
6 I
1 subject to correction in the discussion which follows, is that i
i 2
if one sticks to the kinds of events that are being modeled, then 3
you are probably talking about a controversy which is no greater I
2 4!
than about 10 in both directions, so that an overall range is I
i 3
4 g
5l certainly not greater than 10 or 10 but that those numbers do N
i 3
6; not include and probably cannot include a set of events such as R
7j safeguard, sabotage, terrorism, society falling apart in some E
n j
8' sense, and if we are not worried about the United States society d
d 9!
falling apart, then there are reactors in other countries, or.
I 10 you could have a local society falling apart, and something going z
h 11 on in that situation, and that that probably is' inherent in any a
I
- j 12 '
aspect of technology or human activity,
- hat there are such
=
f 13 risks which are around and there is nothing that can be done
=
j 14 f about them.
I E
15 '
So that it doesn't make sense in talking about any j
16 technology to try and vest more confidence in those estimates m
p 17 than you can with respect to a rather limited set of human x=
E 18 activities, h
}
19 l Okay, so that we do have a large range.
Let me compare n
20 '
that again with the endeavor going on in other parts of the 21,
government in looking at risks from carcinogens and other 22 '
tcxic substances, where the levels of uncertainty are greater l
I
-3 3
23 than 10
-- sorry, 10 or 10',that there is a much greater range 1 24j when you are trying to look at the consequences or saccarine, I
25 for example, or water borne carcinogens than these kinds of i
i ALDERSON REPORTING COMPANY. INC.
7 g
I l
numbe rs.
j t
2i And as you might guess, there is immense controversy l
3l that follows from attempts to regulate under those circumstances, 4l and yet at least in those cases I am one of the strong 5!
proponents of some sort of quantitative regulation.
e R
l n
i 3
6, Then George Sege raised a different issue, which was, o
R M
7; even if quantitative safety goals are not helpful at the moment I
8 I
in making decisions, then they may be dynamically helpful in i
d d
9 forcing us to confront problems, in forcing us to try and find i
h 10 l some better tcols for getting at these issues, for helping us E
l l
ij l systematize and make more consistent the approach in regulation, f
3 1
d 12 l and I must say I am drawn to all that, and finally, one of the 3
I j
13 '
things that seems self-evident when you look at regulation in the E
i 14 ;
past, is that when you have strictly qualitative criteria, then a
b it is very difficult for regulators in practice to make decisions 2
15 l x
i z
.]
16 and justify them, because the gounds seem to shift all the time, x
g 17 and from year to year you may get vastly different decisions,
=
18,
really vastly different decisions, unless you have some anchor r
i w
i
[
j9 for all that, all of which turn out, I think, to be arguments for A
20,
some attempt at quantitative safety goals.
2j !
Then, the final issue I want to raise here had to do 22,
with the suggestion that I made about an inductive approach to 23 quantitative goals of saying, well, why don' t we try a few l
I 24j cases and see what emerges from all that.
That was objected to j
i I
on the grounds that it is not clear that that approach is going 25 i i
f 1
ALDERSON REPORTING COMP ANY. INC.
8 1
to turn out to be better than a formal confrontation of the issue 2'
in rulemaking, and I think that that is a subject that we might 3,
want to talk about a bit.
Okay?
l 4!
So, to summarize, the issue that I am trying to get at o
5, here had to do with the usefulness of quantitative safety goals, 4
I n
2 6
how well we can -- not setting the goals themselves, but how well e
R g
7l can we specify a priori what are the probabilities and risk A
l 3
8!
levels that are associated with untoward events, are they then i
d
=
9!
going to be of help in refining the tools and refining the Y
E 10 l regulatory thinking, are they going to be of help in regularizing
?
I 11 l' the kinds of decisions and providing some anchors and so on.
<3 i
12 Okay, let me stop at that point.
Helen Ingram was
=
d 13 l first.
oa 2
14 '
MS. INGRAM:
I think perhaps I react positively to ds!
15 l all you have said, but want to add something.
I think that we w=
J 16 have construed the notion.of goals quite narrowly.
In many 2
d 17 political processes, the role of goals goes beyond simply the w
I'
=
E 18,
mechanical business o f implementing goals. Goals have a
="
}
19 political function of indicating a dedication to a value or an n
20,
ideal, such as minimizing risk, and one can express a value which 21 '
you can 't exactly test with the nation that it is a goal. we j
I 22 ] intend to achieve in the future.
I 23 For example, in water pollution po licy, it is zero l
t 24 l emission.
Something we don't know how to achieve, don't even i
25 i know how to measure, really, in some cases, but that is i
I 2
l I
ALDERSON REPORTING COMPANY. INC.
i i
l
l 9
i 1
i _
set out there as a goal, and that has the political function of l
2 !
telling the public at large, we don't like the idea of dumping 1
l 3
things in water.
That was an expression of a strongly held d
4 value position, and that value position may shift.
I 5!
One thing these goals, if they are to be successful, i
0 6,
must accomplish is to be a true expression of the..public's i
M 7
concern, ttlus 'public's value position.
S o here I think the 3
8 8
distinction between goals and standards may be quite useful, n
i d
l
=;
9 One can establish a goal which we don't know how to verify.
That i
?
10 l seems to me best to do in qualitative terms, because that most E
i i
5 11 !
easily connects with a value position, that we shouldn't have i
3 i
I j
d 12 excessive risks.
All right?
That is stated in words, not in z
5 1
j j
13 4 numbers.
)
m E
14 Now, it may be helpful to state that in a number, but a
i E
15 that number may hot be verifiable or achievable.
It is something 16 out there that we are aiming towards.
Then, as we talk about a 3
1
^
\\
f 17 l standard that one implements, that of course has got to be i
j 5
18 '
evaluated and it has to be measured.
It seems to me then you need:
E" 3
19,
to say, this isn't the goal, this number'is just a standard, and n
20 we are setting this out with a notion that it is part of a process 21 i to work towards the achievement of that goal.
l I
22 !
MR. LAVE :
Okay.
I am sing to get out my needle, for i
j 23 some of you people, and not let you get away with something, so l
t 24,
Helen, let me get out my needle.
j t
25 ]
If I look specifically to the Water Act, I might I
1 ALDERSON REPORTING COMP ANY. INC.
l i
i
,~m-,-.-
y
n-n.-
,m,w
-,w e -
,,,n.-e-
---r-...
.--,--,w
10 l
i l conjecture that the Congress did the opposite of what it 3
i that it was saying intended by saying zero discharge by 1985, 2
! something that any informed member of the public knew was 3;
i and probably led to an increase in cynicism in the 4l impossible, i
5 public, because somebody who --
e f
Perhaps because it was quantified. Had it MS. INGRAM:
d 6;
e g
7l remained a value position, that in general -- what happened was, R
l
! Congress went ahead, and was driven by the notion of needing to 8 i d
9j implement by quantifying a number.
I would prefer we didn't make d
i 10 !
that mistake.
What I am trying to say, though, is that a goal l
E 11 ;\\ has a purpose, all right, and that purpose is to express a d
12 value position, and if we get so concerned with verifiability d
a that we fail to do that, all we do is put in concrete what we 5
l d
13 am now know we can achieve.
14 d
MR. LAVE:
Give me a kind of a statement of,what the x
2 15 a
l x
' water goal might have been, instead of zero discharge by 1985 16 s
MS. INGRAM:
That we should sach year work toward a A
d 17
! situation where there is less pollution, which we haven 't 5
18 E
I, 19 achieved, incidentally. I mean, if you look at what we have achieved, we haven't done that.
Every year the water doesn't a
20 21 get cleaner.
That every year we should have cleaner water, i
f 22 working toward a situatior there even dirty, really dirty places
.d 23 eventually get clean.
Because that is the kind of la. guage 24 Congress might just as well -- and~it is just as easy to sc:
25 standards on that idea as what Congress did do.
ALDERSON REPORTING COMP ANY,INC.
l
l 11
- f l
l 1
MR. LAVE: Okay, Doug, do you want to withdraw?
l 2
MR. MAC LEAN:
No, this is a quick point, that if, as
)
3 I believe the public us aally thinks of their goals and ideals i
4!
in qualitative terms, and if as you, Lester, pointed out, and as I
e 5i Ed Zebroski mentioned before, which is probably true, that for a 0
j 6
regulator these qualitative terms don't give much guidance, and i
a" 7
they need to transfer these into something thatucan be j
8 measurable, well, then, that is fine, but it also suggests the d
=
9 real need to relate the quantitati.
numbers to the qualitative 6
10 l standards, and this, I think, as we mentioned in our discussion i
E i
=
I 2
11 paper, was one of the real drawbacks of the proposal as it was 3
I d
12 stated, and also I think that in light of the discussion of the z
E l
d 13 I numbers that have gone on, it seems to me that there isn't very
=
\\
{
14 I much consensus or clarity exactly of what the general common E
15,
sense basis for deriving these numbers has been.
So that is a s
y 16 big concern.
A G
17 MR. LAVE:
Dr. Joksimovic?
w
=.
E 18 :
MR. JOKSI.v0VIC:
I am forming an impression in this H
{
19 ; workshop that the recurrence of Lewis Report impact might be A
20 ; around the corner, and we all know that his report was grossly 21 I misinterpreted upon issuance.
Cautious use or PRA is perrectly 1
22 appropriate.
As a long-term practitioner in the rield, I am the l
l 23 first one to admit that high level skill is needed, andPRAisnotj 24 ] a panacea, as some people would like it to be.
l 25 l However, there is a very popular misconception in this l
1 ALDERSON REPORTING COMPANY, INC.
I
I 12 o
I I
1 room and elsewhere about PRA and uncertainties.
Uncertainties, i
2i at the risk of repeating myself several times, exist in our i
3 knowledge, and PRA simply reflects that, and when you, Lester, 4!
talk about uncertainties in carcinogens, that is because the I
g 5
knowledge there is smaller, significantly smaller.than the know9-R 3
6l ledge that we have in the nuclear fi61d, - and as you know, through i
I g
7-our association, the API risk assessment task force definitely j
8, formed the impression that they kncw less about carginogens than d
9l we know about fission products.
a i
Oc 10 And hence PRA is a tool really to reflect that, and the E=
reasor. why we call it probabilistic risk assessment and not 2
11 l d
12 statistical risk assessment is for that very reason, because if z
I d
13 '
it was statistical, we would call it statistical, actuarial, and E
E 14 :
we call it prcbabilistic, and La Place in 1781, to the best of d=
2 15 !
my recollection, defined..what a probability is, and he said 16 simply that probability theory is nothing else but common sense i
A g
17 reduced to a calculation, and that is exactly what we are trying
=
5 18 l to do to the best of our ability.
5 i
19 +
So, to treat uncertainties as a deterrent, I think, is 5
n 20 !
an absolute misconception.
That should be an inducement.
1 21 '
Whenever there are uncertainties, PRA is the only way to go.
l 1
22 And the other point that you.were making about 23 usefulness of the goals, there is no question, I think, in my l
24 { mind that there is no other way to go from where we are here, and l i
25 l that is not only in the nuclear industry, but in other industries.!
i 1
ALDERSON REPORTING COMPANY,INC.
l
~
l 13 i
1!
I think that President Reagan has issued an Executive Order where i
2 he wants new regulations to be on a cost benefit basis as i.
I 3
opposed to a capricious and arbitrary manner, and PRA is:a too.1 4
i 4
to do it not in the nuclear but in the other fields.
I think i
t g
5 the whole trend in this country and elsewhere is going to be in N
i 6l this direction, so it is not an issue of whether we are going to i
R i
7' use the goals or not.
The issue is what kind of goals, and what
)
j 8,
kind of goals are the. most productive, and I think further i
e i
o 9
discussion in this field should be in that direction.
io j:
g 10 MR. LAVE:
Let me use that to ask the question -- let t
z l
=
j j
11 me point the finger at Gordon Corey, since we did not get a chance l
3 J
g 12 on the panel to hear you out on this, but you seem to have i
E I
}
g 13 j strong views.
What is the objection to trying to state m
i 1
w j
g 14 quantitative safety goals as a current reflection of our s
m i
2 15 I thinking about qualitative safety goals, of saying that at the E_
I 1
i j
16 moment we translate these qualitative safety goals into the A
d 17,
following numbers --
a
=
j 5
18 MR. COREY:
I don't have any objection to attemptino to a
=
g i
l g
19 l establish quantitative etandards.
I have -- I stated earlier that 1
a l
20 l I think it is -- I have a gut feeling that it is is premature to i
1 21l establish quantitative standards in such a form that,in anything I
f 22 f other than a tentative form today, simply because -- I. guess I
I 23 for reasons similar to what Tom Cochran might express, but for l
24 l a different reason, and that is that I think we would, if we --
)
25 I think if we were really to establish firm: quantitativ6 goals' 1
i ALDERSON REPORTING COMPANY. INC.
14 1
I today, we would necessarily establish them in a forum which would 2
be so strongly reflective of public concerns that we might i
j 3
end up with something that to me was completely unrealistic 4
in the way of a goal, and public concerns have a way of changing.
5 I was a murderer:in 1970 because of sulfur dioxide, and I trust e
3 i
a j
j 6
that if I were still working for the Edison Company in 1990, I R
i 7
wouldLbe a murderer for some other reason.
Right now, nobody
-f8 has yet accused me of being a murderer with nuclear power, but J
dd 9
there are all kinds of swings in public concerns, and I just am i
1 i
E 10 reluctant at the moment, two years after TMI, to start 3
3_
t j
11 establishing quantitative goals and going through the necessary B
i g
12 process of heightened debate on how many people you want to kill q
g 13 l and things like that.
But!'I have no objection at all to setting i
m i
i h
14 l up some tentative quantitative standards.
I think we have taken 2
15 a pretty good first crack at that in this paper, despite all of 4
1 w
I j
j 16 ;
the coraments and criticisms that have been raised in this
^
d 17 workshop.
I don't think they are all that bad, but as I.said, I i
M 18 l just have a general reluctance to adopt something today that we 1
5 h
I 19 !
will regret having adopted.
g 5
i 20 l MR. LAVE:
Let me get out my pin again, and I argue j
21 '
sometimes either side of this question, but since you have 4
22 argued one side of it, I will take the other.
I guess that as i
23 '
soon as Mel Ernst and his people put out some new licensing t
)
24,l procedures, that Tom Cochran is going to wind up saying, look, s
4 25 ) you guys, implicit in your judgment is the probability of a core a
i ALDERSON REPORTING COMPANY. INC.
I 1
~ -
15 1!
melt, meltdown of 10 is accpetable, and so on.
That is, in
-4 i
I l
2' any set of judgments, there is going to -- qualitative i
j 3
judgments, there is going to wind up being some implicit I
J 4i quantitative goals.
I i
5' MR. COREY:
Of course.
g l
9 3
6, MR. LAVE:
And given that that is all going to be E
7l 5
g publicized, why not start out with the quantitative goal, save 1
5
?,
8 Tom the problem of doing the calculation in.the:first place, and c
i 9
9!
let's talk about it.
That is, if later on when Tom; announces those--(
l 4
z0 G
10ll and says those are unacceptable and we get some general notion j
I j
11 l that they are unacceptabic, wouldn't we have been better off by j
j 12{l 1
3 going into a hearing in the first place and saying, these are i
2i i
g 13 our proposals, are they acceptable, and if it turns out they are m
I h
14 not acceptable, let's take some other ones rather than sort of 4
15 the back door route, j
16 :
MR. COREY:
No, I don't think so, for the reasons I A
\\
d' 17 l stated.
5 5
18 l MR.. LAVE:
Tom?
=
1 I
19 MR. COCHRAN:
I just wanted to use the pin on you --
20 !
(General laughter.)
21 MR. COCHRAN:
-- and I don't think you have made the 1
22;i case that your analogy with these other technologies is useful 23 '
here and correct.
The case that you chose was one where there l
i l
24 j are large uncertainties, and then the debate is really over --
i 25 it is very clear, and people understand what the issue is, and t
ALDERSON REPORTING COMPANY. INC.
I
l 16 i
i 1,
it is over whether you want to count the bodies later or be l
2l risk averse now and set a proper standard, set a tougher standard 3,
or maybe have to count the bodies.
4 That.:is not the situation you have in what is going on a
5 here, and I think you are being a little naive about the under-N.
3 6'
lying aims of the people that are pushing for quantitative goals I
g 7
and the use of PRA.
You've got to ask yourself, why is the 8f industry coming in here and arguing so hard for probabilistic d
c 9j risk assessments?
Why are they doing it here, and why are they t
i l
c l
y 10 saying, we don't want to touch it in the waste menagement section?
z
=
i E
11 ;
And'I would submit it is because they think they can l
j 12 l do precisely what everybody is saying and what even they are 5
l 13 l saying is inappropriate, and that is using to undercut the j
14 l defense in depth approach.
I mean, there is all this talk about, l
2 15 well, we are just going to talk about priorities and so forth.
s 16 Dut that is not what is going on, and you ought to recognize that, j
n 17 Yesterday evening, we saw a presentation using the
=
I 18 l WAsil 1400 model to give us all this data so that we could make t
=
C 19 f some sort of regulatory judgments, rulemaking about where to set 7
A 20 l the quantitative goal, and the Commission has just gone through --
21,
Congress or the Commission has been handed the Lewis Report, said 0
22 " the model was inscrutible, contains statistical errors in it, 23 an a consequence this model shouldn't be used again until it is improved.
l 24j I
25 l The Commission essentially endorsed those comments and j
i 1
ALDERSON REPORTING COMP ANY. INC.
i.
17 1
said, you shouldn't use the model where it is inappropriate, yet i
2 the staff is churning this stuff out using a bad model, an i
i 3
it scrutible model to set new safety requirements, and to supp.' ant j
4 a very different approach for setting safety standards on the e
5 Lasis of fictitious, you know, the fiction that comes out of these 9
3 6
models, and one of the major criticisms with the Rasmussen model e
7l Eg was that the uncertainties were too narrowly defined, and that s
j 8
the thing -- you know, things are really potentially much worse do 9
than what that model suggested, and now we hear best estimates, Y
l 10,
best estimates in the quantitative goals of the staff, best z
i
=
G 11 estimates over here.
He just wants to talk about best t
<2 d
12 estimates.
z l
3 I
13 l MR. JOKSIMOVIC:
I didn't.
l E
I 14 !
MR. COCHRAN:
I wrote it down when you eaid it.
d I
u 15 !
MR. JOKSIMOVIC:
I didn't say best estimate, s
16 i MR. COCHRAN:
But I don't think the two are analogous 3
i g
17 i at all, and I would oppose putting in the quantitative goals,
=
18 I because I can see the abuses that are coming down the road.
We 5
l l
{
19 !
are already neeing them.
And I don't think you ought to rely 5
20,
on the Union <of Concerned Scientists to step in some rulemaking 21 to sort of clean up tl is mess that -- by throwing something out 22 and seeing how it washes.
I think if you are going to throw 23 something out, that you ought to throw something out that is 1
24 pretty good to begin with.
25 '
MR. LAVE:
Okay.
Let me see if I can get my scal;el I
ALDERSON REPORTING COMPANY, INC.
18
{
l l
1 and resolve what the nature of the differences are.
It seems to 2I me that insofar as -- if you had a technique, if you could l
3!
improve PRA to the point where you agreed with the nature of i
4; the calculations --
I e
51 MR. COCHRAN:
You can't do that.
You can't do that
.s 6l for this technology.
R 7
MR. LAVE:
Let me give you a hypothetical.
Okay.
As a 8;
hypothetical, if you had a means of calculating risks which you d
9 thought was fine, then would you not use that instead of other z
e i
y 10 ;
rules which by their nature had been arbitrary to cegin with.
z Ey II MR. COCHRAN:
Sure.
3 i
N I2 !
MR. LAVE:
Okay.
So now, the nature of the quarrel is
=
m j
13 '
the extent to which PRA, since that is the case in point,
=
x 5
I4 accurately or inaccurately reflects the underlying probabilities, b
_j 15 !
MR. COCHRAN:
It is primarily that, but there are other
=
l j
16 issues.
I mean, if it is inscrutible and the public doesn 't A
d 17 '
understand what is going on, you've got another problem as well, w
i
=
}
18 '
but for the sake of argument, I will take you this far.
Ib I9 g
MR. LAVE:
That's great.
Okay.
Now, at that point n
20 Mel Ernst had something to say, so I assume along these general 21f.
lines.
22 MR. ERNST:
Yes.
Somewhat repetitive, I guess, but I 23 do want to again re-emphasize that the proposal in the draft 24 is not undercutting the present regulatory structure.
In our l
25 l view it is to enhance the safety of nuclear power plants and l
i
?
i ALDERSON REPORTING COMPANY. INC.
l
l 19 4
leaving the regulatory structure essentially as is.
I MR. COCHRAN:
.It savs it is not.
j 2
MR. ERNSTr. I"would like to agree with you also, agree i
3 4i with you, Lester, that I don't thihk the staff in general 1
5!
prefers ad hoc decisions, yet there is PRA and there are PRA o
i N
i i
6l studies being perforced, and the staff does use PRA techniques I
t i
j 7
in analyzing issues and uses the results as an assist in making 8,
decisions on safety matters.
I d
i s
n 9'
I think for that reason there is a preference to Y
5 10 l taking these ad hoc sometimes individual standards and coming up l
E with some kind of a more codified institutional standard setting 1
I 11
)
3 i
i d
12 ;
as a benchmark for making individual decisions.
Z 5
j j
13 i I would also like to agree with Mr. Corey that my i
i
=
14,
perception is that the proposed core melt goal perhaps is mis-W j
2 15 l advertised.
We have heard a lot of discussion of the subject w
=
l J
16 '
yesterday and today, and I think I am in complete agreement that I
E i
d 17 it is not really a safety goal, it is really an interim E
i I
E 18 !
engineering standard that one strives for.
2 l
i I
19 '
I think that would help from a perspective standpoint 4
l 20 l maybe. understanding why at least at the present time there is no j
i 21 l containment tentative engineering standard.
The perception was
+
1 1
atf.least atuthisitime that we didn't quite understand how to set j
22 l 4
l 23 such a standardy,and we might better understand in the future
.4 i how to set such a standard, and I think that is the rationale.
i I
25 So, we do hava a subset of engineering standards or l
I i
l ALDERSON REPORTING COMPANY. INC.
l
. m
,_.s
._m._
_ - = _
20 4
i l
I ;
tentative standards that usefully implement a safety goal, and l
2 perhaps' stating a, core melt as a safety goal is confused and j
i I
3
- improper, i
l 4i MR. LAVE:
Okay.
I think Ellen Weiss was next.
g 5
MS. WEISS:
I wanted to respond to comments made at N
1 j
6l the beginning.
I have read the Lewis Report several times.
And R
A 7
I think the point for our purposes, the major point made in s
j j
8 that report and made by the Commission's endorsement of it, is d
1 m
9 that the absolute probability figures presented in NASH 1400 I
i 2
C l
g 10 are not reliable figures, and what that means is that PRA does z
=
j j
11 not reliably estimate the level of risk or conversely level of 1
a j
j 12 safety of any individual reactor, and of course that is precisely 5
i j
y 13 l the use to which it is proposed to be put, at least in many of i
=
14 the applications that are proposed in this paper.
e 1
l E
15 :
And the objection to PRA is that it implies a level of 1
w 1
j j
16 j precision by use of mean probabilities or point estimates which i
'^
i d
17 e is not technically supportable and which cannot be accepted.
'z
\\
M 18 '
It is not non-controversial.
And which is subject to what I
=
4 H
E 19 believe are rather easily foreseeable abuses.
Many of them have 3
A 20 j been discussed over the last couple of days.
I think they r
21l generally fall under the rubrick of resistance to particularly 22 l moves towards mitigation matters, like hydrogen control, for 1
I j
23 example.
4 24 I want to respond finally to the suggestion that there 25 are two ways in which one sets rules.
One of those is arbitrary j
s l
ALDERSON REPORTING COMPANY. INC.
J
21 l
l 1l and capricious and the other is using cost benefit analysis as the I
i 2'
Reagan Administration proposes.
Cost benefit analysis as i
3 established by this Administration is not a purely scientific
~
4 endeavor.
It is not an effort simply to make setting of j
5 standards a more rational one, it is a political decision, and i
N j
6! it is a decision to sort of shift the presumpions in this society l
R j
{
7 away from more protection to spending less money for protection 4
s
)
8 in favor of allowing business to use it the way it wants to use it,
d i
d 9
So I don't consider that an improvement rationally.
i i
l o
I
{
y 10 -
I consider that a political decision.
3 4
11 7
MR. LAVE:
That does not mean that one could not j
3 j
12 establish benefit cost analysis, putting the burden of proof in 5
s 13 1 the opposite direction and get more systematic analysis in a way E
l l
14 '
that you would agree with, i
j E
15 ;
MS. WEISS:
What it means is, there are lots of ways j
j to make these decisions, all of which have some political 16 l
=
j d
17 content, and none of which I don't think one can either dismiss a
1 3
i M
18 or accept on purely scientific grounds.
=
r w
j 19 j MR. LAVE:
Okay.
Now, just for reference purposes, n
i 20 l let me point out that when the Food and Drug Administration
-6 1
21 I uses probability of 10 as being " negligible" what they mean by I
l
-6 1
22 l 10 is less than one cancer per million lifetimes, and when there i
.l i
23 '
is uncertainty, what they mean in. practice is that their best
-10
-11 24,
estimates have to be scmewhere on the order of 10 or 10 so l
f l
25 that the bound of tae estimate is surely going to be less than i
)
(
i ALDERSON REPORTING COMP ANY. INC.
I
a 22 i
l I
-6 1'
10~10
-- sorry, 10 all right? So that when they apply their I
2l quantitative goal, they are not saying that their best estimate f
~
3 means that the risk is less than 10 They are saying that 4
they are going to apply it so that it is very, very unlikely that e
5 the risk would be greater than 10-6, N
i 6!
MS. WEISS:
That is a proposal I think is worth G
i d
7 considering, in other words, setting the goal so that it is sj 8,
maximally conservative.
We set the goal and we say it will be d
I
)
- [
9l implemented in the way such that considering all of the 1
3
[
10 uncertainties, it is impossible that we are wrong in a non-l E
l 11 conservative way.
l j
12 l MR. LAVE:
.fot impossible.
You never want to say it is j
E i
^
13 I impossible.
That is too strong a test.
Very unlikely.
Okay?
h 14 l MS WEISS:
That we are being non-conservative about 2
15 l it.
That certainly is a far different proposal than what appears s
]
y 16 l here or anything else that I have heard of in the.ast couple of
.i A
2 b.
17 ; days, and I suppose I would be interested in hearing the industry w
=
1 5
18 respond to it. My gut feeling is that it is not something they
}
5 i
E 19 i will be crazy about.
3 n
20 l MR. LAVE:
Well, let me just clarify it a little bit, I
21 and then we will ask for some response. I guess that the 22 clarification is that if you state your goal so that -- the FDA i
23 ' would not want to state a safety goal of 10~
and argue that
-10
-6 4
24 ;
if your best estimate is 10 surely you will be less than 10 25 i Instead, what they are saying is, they are setting the safety goal I.
1 l
1 ALDERSON REPORTING COMPANY. INC.
i i
23 l
o l
-6 1
so that surely you will be better than 10 Those are 2i different.
Those really are different in practice.
Okay?
It was 3
my impression that with respect to many of these calculations, 4l that at-least some of the NRC people and some of the industry l
c 5
people I was talking with meant that they would use a conserva-R
-4 3
6 tive calculation to get to this 10 or whatever it is.
Now, we r
k P
7 can make that much more explicit, but it seems to me that is the t
5 3
8' nature of a proposal that we are talking about for quantitative a
i d
d v'
safety goal.
z C
g 10 l Okay.
Herb, I think you were next in all this.
z
=
i 2
11 l MR. KOUTS:
Yes, I really want to say some words about 5
i d
12 l the Lewis Committee Report.
I not only read it, I helped write z
=
i d
13 '
it, and so I know something about what was in it and what was am 14 ;
meant by it, a
l 2
15,
First, 1 want to say some words about what was meant i
16 by the mistakes in the report.
There were some mistakes in the sx d
17 repo rt.
There were mistakes in the WASH 1400 analysis, and we u
x 5
18 j pointed a number of them out, and I assume that some*of'the
=
H; 19 ;
mistakes that we pointed out are being corrected, and that is 5
i 20 l all to the good.
I would not expect any large report, any study 21 !
of that size to be done without mistakes, and some were 22 certainly made.
I 23 Whether the mistakes were of such a character as to I
24 change the conclusions very much is another matter. We didn't 25 explore that question.
We did make scme comments about
{
l I
'l i
ti ALDERSON REPORTING COMPANY, INC.
i
24 l-l 1 j probabilities and the uncertainties to be attached to probabilities.
These did not mean that the probabilities were 2,
I 3l no good at all.
As a matter of fact, what we meant only was i
4l that there were some error estimates in the report, attached to 1
l g
5 these probabilities, and we did not have confidence in the size N
6!
of the error estimates that were given in the report.
7 Different people on the Lewis Committee had different 8
interpretations of how one might proceed from that point to make n
dc 9
further conclusions.
Hal Lewis's conclusion was that the
?.
i 10 probabilities -- the risks that were calculated by the NASH 1400 3
i 11 '
exercise were too large.
He thought the risks were much smaller
~<
a 4
12 !
than that.
I didn't feel as strongly as he did about this z
5 i
d 13 '
matter, though I was inclined to go in that direction.
There were i
E i
j s
14 some other peop on the committee who thought the error, if w
i M
i E
15 any, was in the other direction.
l
=
j 16 l But that was not really our comment at all.
Our A
j d
17 i comment was simply that the factor of three uncertainty estimate a
=
5 18 i which was given in the report was probably an under-estimate of I
5
]
[
19 '
an error margin to be attached to the calculations.
That was all.
i A
)
20 l MR. COCHRAN:
No, that's not all, Herb, because --
21 MR. KOUTS:
That was all that the report said.
t k
l 22 '
MR. COCHRAN:
No, that's not all that the report said.
1 23 MR. KOUTS:
It said the absolute --
l 24 MR. COCHRAN:
We are unable to determine whether the
{
1 4
i 25 absolute probabilities of accident sequences in the WASH 1400 1
f i
ALDERSON REPORTING COMPANY,INC.
1
-~,~,-, -,,--
,----.-,.----....,,,-.,-.c.,,.,
25 i
l I
i j.
are high or low.
1 MR. KOUTS:
That's right.
2 f
3 MR. COCHRAN:
That's a statement about the best 4) estimate, not the uncertainty, error bars.
5j MR. KOUTS:
It means the same thing.
l 4
E I
n l
l 8
6; MR. COCHRAN:
Well, it doesn't.
I n
i f7 !
MR. KOUTS:
Now, the report also said the methodology f
8I which was used in WASH 1400 is the way to go.
It says that d
9 WASH 1400 at the time was the best work of that kind-which had i
h 10 been accomplished, but it had a lot of improvement to do, to make E
5 11 l on the way to making it useful.
It said that the WASH 1400 --
i
<3 l
g j2 j the risk assessment methods should be used with caution where 1
z
=
j l
j 13 they are applicable in the future, and it was very strong on I
E i
i s
14 i that point.
a b!
15 l When the word "inscrutible" was used, it was not used w
1
=
l J
16 l with respect to the methodology or *.ne techniques at all. It was 4
W
{
g 17 l used with respect to the way the rtoort was written, and I have i
d i
=
i 5
18 !
had a lot of trouble with that kind myself, just trying to trar'.
]
=
=
j t
19 some of the calculations through the report, having to go from x
l 5
1 20 l the maf_n report to one appendix to another appendix, and so on, just trying to find out how the calculation was done, eventually 21ll
{
22 l finding out, but having a great deal of trouble.
23 '
So, I wanted to set some of those things.: straight.
l 24 i I do want to say something about the general subject of 4
25j goals.
I think it is really laudable that we have two categories 1
i.'
d i
3 ALDERSON REPORTING COMPANY. INC.
g 26 1
y 1
1 of goals.
The goal that has been talked about with respect to 2
the working paper here is not the kind of goal that Helen has i
4 3
referred to, of course.
The goal which is in the working 1
1 4
draft is something which was intended to be offered to the 1
o 5
Commission as a means for them to use to decide, do we do this A"
3 6l or do we do that in a regulatory context, and the question then R
l 7
is, if we do -- for instance, if you apply that aspect of the goal e
4 g
8 which deals with alternate technol.ogies, you are taking into i
dc 9
account the fact that if the Commission says, no, that is too i
l h
10 big a risk, don't build a nuclear power plant, for instance, and z
j h
11 therutility then turns to an alternate technology which has a a
4 j
j 12 greater social; context, greater social impact, that sort of thing
)
5 g
13 I should be taken into account.
l
=
i 14 !
In establishing goals of this kind, you look i
E l
j 15 i realistica11p'at what the alternatives are in the world today,
\\
i j
16 l and that is the. kind of thinking that I believe went into the
^
\\
d 17,
choices of numbers,in this document.
It is quite good to 5
{
18 establish real goals for the future which are much better, which E
i g
19 l have a greater social value than the ones that you are using i
5 20 tempo rarily.
I donft-think you should set them at such a point I
21l that they become absolutely impossible.
I think you should set
)
I j
22 them where you think you might achieve scme success and with a 23 '
reasonable investment of resources, considering that you have a 24 lot of goals in a number of other places which are going to i
i
{
25 compete with these goals.
4 i
e j
ALDERSON REPORTING COMP ANY. INC.
i i
+
i..
m._.,..,..,-
m.m..,_
__-..,-.ry,,
,__,,m.
9 27 I
i 1 i If you are going to try to make nuclear plants safer 2: and safer and safer, there must be some point at which your l
3i dollars are better spent in making something else safer because i
4i something else is threatening you more than the nuclear plants.
I g
5 So you don't want to say absolute zero risk is your absolute j
0 j
6 goal.
I think the water discharge goal of ero discharge is a R
7 very bad goal for that reason.
It doesn't give you any hint at sj 8<
ali where you think you are going to make that transition from 0
9 9,
worrying about' this sort of thing to worrying about the rest of zcy 10l the things in life.
z
_E 11 MR. LAVE:
Neil Wald.
l B
d 12 l MR. WALD:
I think Herb has touched on some of the z
5 d
13 ! points that have been concerning me in the discussion of goals.
5 14 When I come back in this paper to the reason for possibly i
j 15 j enunciating a safety policy statement, the public understanding x
j 16 is the first thought that is mentioned as a justification.
Then A
d 17 there is bhe justification for regulatory purposes, a more w
2 1
18 !
coherent and consistent regulation than the basis for it.
I am
=
r i
E 19 not sure we are not in some kind of semantic trap in calling 3n 20 '
all of these things by t he same word, goals, because I don't i
21 I believe that the kind of words that Helen referred to, which I --
1 I
22 ' I come to this from the standpoint of wanting to see something l
i 23 that improves public understanding so that the biological, i
24 psychological stresses that we had to deal with at TMI and that i
I 25 i we deal with every time a radiation accident occurs are I
i i
l a
i 1
l s
ALDERSON REPORTING COMPANY, INC.
28 I
i i
1 minimized.
That is my objective.
I don't think you can 2l simultaneously have goals which contribute to public understanding 3
and also support coherent and consistent regulation, and I think I
4l you made the point very clear that some of this is directed at 1
g 5j the Commission and some of this is directed at the public, and N
i 3
6I I am not sure that they should be given the same terms.
Maybe G
f 6
7l that would alleviate some of the problem or question in whether-
~u j
8 quantitative goals are appropriate and can be understood by the d
C 9
public, for instance, or probabilistic approach and the like.
z, g
10 ;
s I would speak in support of qualitative goals that do z
E 11 ;
make a commitment to the public.
I think that is one of the a
i j
12 l most important things that can be accomplished by these goals, 5
i l
13 and I would also support quantitation of objectives which meet
=
x 3
14,
the second test, but I would not try to make them also publicly
{
15 understandable.
I j
16 One other point about PRA, the approach that Mr. Ernst A
d 17 described seems to me to be the richt one, that it is a tool in w
5 y
18 '
assessing how close one is coming to a goal which may never be P
l r
19 I a
met bat which at least can be -- progress can be evaluated by n
20 l this tool.
My concern about this tool,and I am not sure, 21 somebody yesterday separated PRA in effect into the failure rate 22 or the risk of failure of the hardware, and then the more general i
l 23 risk assessment, which includes the consequence model, and I think' 24 some people have been using the word PRA aiming much more at that l l
l 25 i failu.e assessment of the equipment.
I am not clear on whether l
ALDERSON REPORTING COMPANY. INC.
29 I
i PRA is under dynamic and continual review and update which would 1 !
2 in effect go past the state of the art when the Lewis Committee I
3! looked at it, and I am a little concerned because in at least i
4l one area which was discussed specifically last nignt by Matt, tne 5
bottom line is the consequences, and at least from my background a
E I-n 3
6 and training it is the health consequences that are the most e
R R
7, important thing, and the basis by mandate for the Commission to
,j 8l make all of its assessments in terms of protecting public health d
i d
9j and safety.
I
,5 10 l I don't get the feeling that there is continuing review z
2 11 i of the health side of this consequence model, and I say this 5
i s
i d
12 l partly bec ause I was responsible for a large part of that, and z
5 i
d 13 at the time anticipated, Number One, that it would be a 2
14 l controversial thing to be doing in the health area, just as it dv!
15 l may be in the engineering area.
Number Two, that;it would be w
i
=
.?
16 ;
subjected to continual update and revision just as the fine s
A t
17 tuning has apparently gone en in the hardware area. Possibly
=
5 18 l because of the nature of the NRC itself and the very low i
E 19 l profile of representation or skill in the health area, it is taken 5
i 20 !
as a given that what was given then is today many years later l
21,
still the right consequence model which provides the bottom line, 22 :i and I am concerned that this updating process does not seem to be e
i 23 going on, and that it is not recognized by those of you who are 1
24 kind enough to consider that the health is a given and we don ' t l
t 25, have to worry about the numbers in that part of it.
j i
i
.i ALDERSON REPORTING COMP ANY, INC.
30 1
End of statement.
l 2
MR. LAVE: Gordon Corey.
3, MR. COREY:
I simply had my card up because I* wanted to i
4l second waat Herb said for the record.
I think that I have e
5 expressed myself as being'in strong disagreement-with Tom M
i a
i 3
6; Cochran and Ellen as to the validity and character of the WASH R
a 7
1400 and the PRA techniques.
s g
8!
MR. LAVE:
Ed O'Donnell.
O d
9 MR. O ' DONNELL:
I would like to, I guess, respond first i
h 10 !
to some of Tom's earlier comments, and then address the question 3
5 11 Ellen raisedi d
12 I would like to lay to rest, if I can -- I have tried z
5 i
d 13 to do this a number of times, as far as the industry's conten-E 14 tions in using PRA in safety goals, is quite definitely not to I
E 15 j undermine the existing defense in depth concept, and I think the I
16 l industry would be willing to stipulate that, make that part of
\\
d i
d 17 the qualitative goals, that regardless of what PRA and quantita-I E
18 i tive safety goals say, that the concept of defense in depth must E
y 19,
be retained.
n i
20 l The problem as the industry sees '.t is not one that i
21 the existing regulations are so onerous that they can't be met.
22 They are in fact being met.
The problem is one that they are 23 '
under constant change and instability, and this is the area that 24 l we see the most valuable application of PRA and quantitative li 25 ] safety goals in making those -- that change, decision-making ALDERSON REPORTING COMPANY, INC.
~
31 1
j !
process more rational, and using PRA techniques, and using the i
2 quantitative safety goals as benchmarks to assist in that i
decision-making process.
That is, do the existing regulations 3
i 4l need to be changed?
Do they need to be increaseci 4
l 5l And aside from that, even if we did not support the e
3 i
i 3
6 idea of using PRA, which we do, NRC staff is requiring applicants e
f f7 to go out and do PRA and to submit the results.
4 y
8!
In the Zion,, Indian Point case, they are going to be I
d 9l the subject ofi discussion at the hearings, and itz.is beyond me g
-2 I' jo to see how the NRC can go out and say, well, you have met the e
M j
j) regulations but we still want you to do the quantitative PRA, i
t B
l d
12 which we are somehow going to use as part of cur decision J
z=
j
'=
13 !
process, without having some quantitative goals to guide those a=
E 14 decisions.
w i
i b
I 15 l So, I think that is how the industry would view this, d
16,
as something to control change, to make things more rational, 3
A g
17 '
in terms of decision making and change, retaining the 5
E 18 l underlying concepts of the existing regulations.
a E
Now, with respect to the issue that Ellen posed 19 x
4 n
l j
20 l regarding the conservatism required in these things, I think the i
21 '
use of, us'you described, Lester, by the FDA in adopting a very 22 j conservative approach, is probably under the legislative 23 mandate of that agency, probably appropriate. That is, they have 24,
a mandate to regulate carcinogens to a zero risk level, and in i
25 fact designing or showing that you have met some number higher
)
f l
ALDERSON REPORTING COMPANY. INC.
j
- ~.. _
32 i
l
[
j !
than that is really non-conservative with respect to that very i
l 2i absolute standard.
l l
3l The problem there I:think is that the standard in the i
l I
first place doesn't make sense.
When you turn to nuclear power, 4
3 the NRC does not have the mandate to reduce risk to zero but to 1
i N
1 j
6l establish that the risks are not undue to the public.
The 4
k 7l problem of getting very conservative gets into the issue of 8:
resource allocation.
If youaare really concerned with cost a
a d
i g
9l benefit balancing, doing things very conservatively does not i
i h
10 serve that objective of making sure that you have wise decisions 3
i s
11 with respect co where you are putting your money in getting the J
B
)
d 12 l best benefit out of it.
3 So, biasing things unduly towards very conservative 13 z=
i s
14 l approaches does in f act bias the results and may be putting your j
s i
j 4
m j
2 15 investments in things that aren't really protecting the public a
J
=
i 16 :
in the best manner.
4 a
w i
j g
17 l MR. LAVE:
Just let me respond to that rery quickly.
ax i
5 18 You are certainly right that -- I think the Delaney clause that i
d j
t 19 l the FDA works under in the case I was talking about doesn't make x
4 5
20 l any sense, but I think that in any kind of regulation, 1
21 l quantitative or qualitative, where you have uncertainty, then i
l l
22 l in the safety -- health and safety area, then regulators ~have
{
23 tended within the United States to resolve that uncertainty in 24 favor of, or being more conservative because of the uncertainty, t
i l
25 and whether you make that explicit or not, I don't think that 1
4 i
ALDERSON REPORTING COMPANY,INC.
l
33 4
1 Certainly if you take a look at past regulatory practice within l
4
'the NRC or in other areas, if there uncertainty, then you 2
j 3
don't take a best estimate, you take semethiag that is down 4
towards the lower bound of what ycu can do, and I was simply e
5 trying to make that explicit.
}
N i
3 6i MR. O'DONNELL:
I would agree that the uncertainty --
o 4
R R
7 somehow you've got to account for the uncertainty, and somehow A
j 8
take a more conservative approach where the uncertainty is d
9l greater.
Now, whether you do that explicitly in the PRA or you d
1 i
h 10 do that through having additional defense in depth concepts and i
z i
11 additional deterministic requirements, is a matter of debate.
t s
MR. LAVE:
No, I think that is fine, but the other d
12 I
z i
l 13 l action that you get there, that again came out of the I
=
i g
14 !
carcinogen testing, was, if you have a penalty that is in direct J.
F 2
15 proportion to the unresolved uncertair +
at the moment, then you U
i 7
16 l are going to push very hard towards both the government and the i,
m
]i g
17 i interested industrial parties resolving the uncertainty, because w
i j
t 5
18 there is a huge payoff for getting certain kinds of answers.
That E
19 will push forward the science, and at least if one believes in a A
4 j
20,
-- if not a perfectable world, at least a world where you can i
21 i make some progress, that will help.
1 22 t Okay, I have sort of the idea of the order in which 1
1 i
23 people have put up their signs, but let me ask you to start 24 bringing another subject into all this, namely, the second one l
I 25 that I posed, of implementation, so that now we are -- I don't I
3 3
ALDERSON REPORTING COMPANY,INC.
i
_ _ - - -, _ - ~, _.., - -
- _,--.~ _ - - --, ~ --------
34 l
I 1
want to get away from quantitative goals, and PRA, but I want to l
2; have us st'.rt to talk abou*. implementation kind of issues.
We have l
l 3j been sliding into this a little bit already, bu
! would like us 1
I to slide in a little bit more explicitly.
4 i
5l I guess Steve Derby was next.
R 3
6, MR. DERBY:
I think it is pretty clear that PRA is R
7:
not a verifiable, accurate prediction of the body cocnt.
I l
s[
8 don't think there's any disagreement.
Being true, or being the d
C 9
opinion of this group, I think it is inappropriate to try to set io g
10 body count goals for the nuclear industry.
On the other hand, i
I ac j
11 I don't see PRA being thrown out as a useful engineering standard.
t j
12 j It is an excellent way to be proceu ral, although it is 3
i j
13 artificial, for coordinating, organizing very complex, very mj 14l dense information, and summarizing it.
2 15 l Engineering standards could probably use that kind of w
=
j 16 assistance.
The uncertainties in whether or not you can:_be j
17 '
procedural enough and organized enough to bring some bit of M
18 l j
clarity I think are small and manageable tha.
crying to tackle E
I b
I 19,
the idea of whether we have good body count statistics.
If
-5" i
20 people want to take the position that you cannot set goals or --
21 because you cannot verify the body count which you set as a goal, 22 '
I think it is sort of ridiculous.
23 So the use of this, say in implementation, it's an 24 +
engineering standard.
It is artificial, but it is good, better l
i lot of things that are around n s,
and it doesn't mean l
fthana 25 i
i ALDERSON REPORTING COMPANY, INC.
l
35 I
i 1
that that is the body count.
And if you are going to implement 2!
this, you implement it on the level of engineering standard in 3i an area that requires that -- the appropriateness of clarity and 4 f trying to sift through a lot of information, and I think that is l
5:
manageable.
e E
I N
3 6
MR. COCHRAN:
What do you mean, implement it as an e
M engineering standard?
R 7'
4 j
8' MR. DERBY:
I mean, it is going to be written as a d
I d
9 !
reg. guide, or you've got, I don't know, several hundred 2
c b
10 '
thousand pieces of numbers that -- you've got a very hard to 3
j 11 i understand system that-- all you've got is line diagrams, and a
j 12 you look at that thing and say, I don't know, I didn't even know
=
i i
13 -
the pipes were su :ased to be hooked up. It 's very dense.
And
~
e j
14 j you use chis type of analysis to try to cut through that,and u
2 15 i find which line diagrams it is most important to understand.
It d
j 16 is a reflection of your state of knowledge.
It is a way of s
g 17 expressing or summarizing your ability to know what that system F
8 5
18 !
looks like and how it responds in certain situations.
That is E
i j
19 certainly a very lower level endeavor than trying to get the n
20 !
body count from application of the technology.
21 !
MR. COCHRAN:
Do you know what a regulatory guide is?
I 22 '
MR. DERBY:
Yes, I have used them.
6 23 MR. COCHRAN:
I mean, this is cracy. You are suggesting l i
24 ] that they write a regulatory gcide that would in effect say that 25 l exercise this conputer code somehow, and if you get the right i
ALDERSON REPORTING COMPANY. INC.
36 l
1 answer you have passed on the stcff will agree that that 2l represents a sufficient basis for having passed Regulation X.
Is that what I understand you to say?
3, MR. DERBY:
At a lower level, yes.
I am not saying 4
I that if you run this code, that the plant is safe for all time i
e 5
2 3
6 and you can walk away and be happy about the whole thing.
I 7
don't know what regulation X is applied to.
If it is applied to f
8l some auxiliary feedwater pump, well, yes.
O d
9l MR. LAVE:
The lesson of much of the discussion in the i
i 10 l last hour has been,to paraphrase the name of that popular jj TV show of different goals for different purposes, and I think thac 5
I i
'i 12 j Steve made that point nicely in Palo Alto and thereafter, that 6
^
I S
13 we have to think about the consistency of statements at various E
S 14 l analytical levels.
The one you -- what you communicate to the a
b!
15 :
public cannot be opposed to what you are telling the design s
16 engineers at the lowest level, but it should certainly be s
'n g
37 different, different in content, different in specificity, and d
E 18 ;
so on.
But it seems to me that a lot of the wisdom here is to
=
I 19 separate out very carefully the statement intended for the 5
m 20 ;
public from the statement intended for the~ industry from tha statement intended for the design engineers.
Okay?
21 22 Vojin, I think you are next.
I 23 MR. JOKSIMOVIC:
Several points.
With regard to these i
24 allegations that the industry is coming in with PRA and j
i I
25 i quantitative safety goals as some kind of ploy to undermine the l
l ALDERSON REPORTING COMPANY, INC.
i
37 j
i defense in depth concept, I would like to endorse what Ed has l
2; said, and in addition to that I would like to say that the PPA f
3 better focuses that, and I have written many papers on that 4
subject, and I would be glad to provide copies.
J 1
g 5
Now, the second subject of accounting for uncertainties, i
9 i
j 6l it is one of the key activities in any PRA study.
We spend a lot R
d 7
of time quantifying uncertainties.
The results come out in i
j 8; terms of probability distribdtions, and not best estimates.
d I
i c
9l However, people for the purpose of communications that talk in
)
i I
Og 10 terms of best estimates or mean measures, whatever, like we z
1
=
l 5
11 i talked about probability of core melt, and let's say the study
^<
j d
12 i comes out, and somebody says, what is the probability of core z
i I
5 i
s 13 l
-5 melt, and somebody says 3 x 10 that is a mean estimate, but E
i
]
14 l it doesn't mean that the same study, being the reactor safety 9
2 2
j 2
15 'l study or the -- or the Zion-Indian Point study or any other 5
l 16f study, we don't have the probability distributions which are in j
j i
d 17 there, and they quantify what are the u mertainties associated 1
5 l
E 18,
with the meaning of the number like 3 x 10 or whatever.
~
=
l 1
19 'l When I cost a PRA study, which these days, as you know,
}
}
I I
l 20 ! is in the $1 million class, we assign something like 15 to 20 a
21 percent of the effort in quantifying the uncertainties, and in
)
22 l doing that we are also trying to quantify the uncertainties not 1
J
{
23 only for the probabilities of occurrence but also for the l
24,
consequences.
25 However, now that I mention the consequences, I do agree l
ALDERSON REPORTING COMPANY. INC.
1 l
38 3'
with Neil that in the last several. years, after the issuance of I
2; the Lewis Report, there has-been significantly more emphasis l
3l placed on the systems analysis, or if you wish on the P side of a
}
4, the equation of the risk rather than the C side of the equation.
e 5
There-have been some efforts on the C side of the equation, too, I
g 1
n 1
3 6l but I don't think they have been proportional, and I do like to o
i I
l
{
7 take this opportunity to suggest that that kind of balance between P and C should be redressed.
I think that is a very good point, 8.
l N
and I would like to endorse that.
9 l
Y E
10 Now, the last point, on the subject of verification, E
}
l gj PRA results are verifiable in the same vein like calculations t
a d
12 '
for the dams or bridges or anything else.
It is an engineering z
=
E 13 exercise and it can be verified.
The verification process is not
=
i
=
I s
j4 simple.
It is tough.
And one has to do it from the standpoint wH!
15 of verifying the calculations that have been done, and one has 1
d
~
to then corroborate that with the plant experience, like for 16 j i
B e
g 17 ; instance when we talk about, let's say, a plant specific PRA y
i 5
18 : profile, let ts say the plant is just about to start operating, l
t C
I 19,
and we assess what the profile is.
We recognize that that is an 4
5 i
a e
j
)
20l assessment for the plant which hasn ' t started operating. When the i
j 21 l plant starts operating, the numerous patterns are exhibited which 1
4 22, may or may not have been contemplated.
23 If they have not been contemplated, then one has to 24 update the plant specific PRA study in such a manner that at any l
l 25 particular point in time we have a living PRA profile as opposed I
L e:
I ALDERSON REPORTING COMPANY. INC.
,l i
1
.__-..,____._-_,.__....m.
39 j
1 i
I to some kind of static profile which was once assessed and not l
l 2'
updated, and that will then enable us to know at any point in 3
time whether that plant meets the goals or it doesn't.
If the i
4 pattern is such that the plant is outside the goals, then we have e
5 to do something to bring the plant into the acceptable envelope, a
4 0
3 6I and I see in the document that is before us the timing of e
k E
j R
7 corrective actions has been proposed.which you haven't discussed
~
s j
j 8
at all, and I suggest that we spend a few minutes discussing that d
d 9l subsequently.
i h
10 MR. LAVE:
Okay.
You know, again, if I can persuade z
i E
11 some people to put down their signs, I would love to do that,
<3 d
12 but those people who do have their signs up, I want to smear over z
r y
13 I into the implementation stuff, including the timing issue that i
=
l l
E 14 l Vojin just raised.
Dale?
t w
15 l MR. BRIDENBAUGH:
Well, I put up my sign because I see 2
16!
that we have 20 minutes left, and I think there is a part of
~
.-s l
^
l 17 )
implementation we haven't really talked about specifically.
1 w
=
l M
18 MR. LAVE:
Good.
c j
{
19,
MR. BRIDENBAUGH:
And I don't have a solution for it, n
20 ;
but the issue is public participation and public acceptance of i
21 whatever comes out of this exercise.
I want to just raise a i
22 caution or a flag that I am concerned from my standpoint as what I
4 l
23 '
I will call a token dissident in this exercise that this may be 24 l I may be represented as having represented the public, and i
t 25 the public will not have had a chance to really review I
l l
ALDERSON REPORTING COMPANY. INC.
l l
i I
-. -. ~ -,. -..,, -,, -
40 l
1 or consider what is involved in these goals.
I am specifically 2'
concerned that I guess the public meetings that were scheduled 3
have been cancelled.
I am concerned that Brookhaven is looking 4j at one month from now to put together the results of this 1
j g
5 meeting, and perhaps one month.from that the Commissioners are
]
O i
j 6j going to act on this safety goal issue, and I really don't think R
8 7
that there has been a concerted effort to get outside involvement i
- j 8j in looking at the direction that is being taken.
U l
2 o
9!
MR. LAVE:
Okay, I tnink -- let me just react to that N
l 10 I very briefly.
First of all, I assume that each of us around the z
i c
l i
l g
11 table represent ourselves, whatever claims we may make when we 3
j 12,
make a speech.
Actually, I represent 225 million Americans, l
5 g
13 l and all the rest of you people are superfluous.
i j
j 14 l (General laughter.)
i 15 l 2
l MR. LAVE:
So whenever I hear anybody saying about, s
a
=
y 16 l you know, the industry thinks, or intervenors think, we all 1
s j
g 17 represent ourselves, with all due respect, and so these work-s i
M 18 !
shops could not possibly do more than collect a sampling of
=
l c
19 e people and hear what you have to say.
.i 5
20 l The second point of all that is, I think that i
21 i inevitably there will be adequate time, adequate time will be l
l 22 i made for the Commission to hear public opinion.
I would have
+
f 23 thought it, as Dale did, more valuable to get some public i
24 -
opinion in early, because y2u have a tendency to get locked into I
25 nomething, and then it gets much more painful to change it.
ALDERSON REPORTING COMPANY. INC.
I
1 But I think that point is well taken, and I think that 2
the other point being represented generally is that any day when 3
experts alone thought they could represent things and the public l
4 could eat cake is just wrong.
That is not the way we play our i
I g
5 game now.
l j
6 MR. RATHBUN:
Give me just a second now.
As you know, t
g 7
as you pointed out, we did have four public meetings scheduled 4
s j
8l and proposed, beginning, I think, May 19th, and through.the d
i i
9l middle of June.
We weren't able to proceed with that on the Y
I h
10 basis of Congressional action.
I would suppose that we will z=
1 j
11 I prepare a -- for the Commission's consideration a craft policy 4
12 l statement on this subject, but that they would submit it for g
5 I
y 13 i public comment before any final action.
=
m I
g 14 j MR. SEGE: If I may add just one observation to what i
b j
=
1 15 Dennis said, even though the public meetings were cancelled as a g
l
=
I j
g 16 result of Congressional action, we did solicit public comment in s
i d
17 '
writing on the basis of the preliminary policy paper, and when i
5 1
C t
3 18 l the public meetings were cancelled, we extended the comment i
E I
19 l period to provide people who would have come to the meetings an j
g i
20l opportunity to comment in writing.
We have so far received a
)
j 21 I little over 2,000 comments, some fairly extensive, some more
]
22 i limited.
j 23 '
So it is..notta matter of doing entirely without public
~
24 views.
i
)
25 MR. LAVE:
Okay.
Chronologically, I think we have got l
I
\\
l ALDERSON REPORTING COMPANY. INC.
42 l
1 Toby Page next.
1 l
3 2i MR. PAGE:
I want to make a couple of comments about i
3 implementation.
i I
)
4 MR. LAVE:
Wonderful.
=
5l MR. PAGE:
In. passing, I sort of want to tie it together 2
l n
2 6 !,
with some of the things that have been said about benchmarks and e
k7 other regulations.
I think that -- I have to disagree with you i
M I
8 8
about the Delaney amendment.
Basically, it is not a zero risk
]
a i
d 9l concept, I think.
It is perfectly clear, I think, when you read o
l 5
10 '
the language, it says, if after appropriate tests, FDA finds z=
1 2
11 that the chamical is a carcinogen in animals, then you take it I.
j d
12 i out of food.
The appropriate test they are allowed have as much 4
z J
5 1
j 13 i as 40 percent false negative rates.
So if you are allowi*ng such m
l 14 i a high false negative rate, you don't have a zero tisk concept.
]
j 2
15,
So what is the wisdom of this observation in d
i
~
l f
j 16 i the question of implementation here.
That is the question I want j
i p
17 to address.
And it seems to me that what is really going on is w
18 !
that in a 11 of these risk benefit kinds of decisions where 5
i
]
}
19 ! you have a technology that is risky, and you take some steps n
l 20 to lower the risk of possible harm, there i going to be a
)
l j
21 ! balancing process, and the question is, where does the balancing i
f 22 l process take place, what factors are used in the balancing I
i 1
23 '
process, and what sort of tools are used to implement it.
24 So that is the third thing I am trying to get to, and i
25 j in the -- well, okay.
I think in the discussion of the 1
i f
l 1
I l
j i
ALDERSON REPCRTING COMPANY. INC.
~.
43 i
j Delaney amendment, the balancing process took place in Congress, i
thb false negative is high, the false positive is low, we are 2
3 going to have a simple process that econcmizes on the number of 4j false negatives at the price of a lot of false positives, and that j
e 5
was a value decision that was made in Congress, and the tools that 2H r
l
{
6!
are used to implement it are passed on to the agency to do R
1 2
7 something with.
So in principle you want a decision mechanism A
i j
8 that is fast, allows lots of chemicals to go through it, a d
i o
9 chemical that doesn't go through it is a potential false 10 negative.
You know, and so on.
Okay.
I don't want to belabor I
3 I
5 11 that point.
<a i
'd 12 But I wanted to say that if we try to make a comparison z
E j
13 of that with the nuclear safety problem, you might be tempted q
=
i 4
)
j 14 l' to sav, well, in the old says the approach of the AEC was to 0
l 2-15 ;
claim zero risk.
They said accidents were so unlikely we just az i
16 l don't worry about them, no problem.
1
- k
^
g' 17 Now, that doesn't appear to be the case now, and what
=
18 l we are seeing now are pretty high probability numbers for
=
I i
H l
j
{
19 ;
core melt and containment break, so that they are so high in j
M j
20 !
fact that we are beginning to worry that we may nou be able to i
1 i
l!
21 !
finesse the equity issue with the numbers being low enough to 22 !
finesse the equity issue.
So it is just coming back.
t 23 So where does this. lead to?
It seems to me that it 24 leads to the kind of proposal that Tom was making, as I; understand 25 it, which is to, if you look at the Goal Number 3 in the discussion 1
I ALDERSON REPORTING COMPANY. INC.
s
-4, r
---_.t
--.m 1
44 1
i 1!
paper, and say, well, as Kouts said yesterday, this number seems l
-4 2i awfully high, perhaps it should be lower, instead of 10 or
-5
~
3 10 plus another 10 for breach of containment probability 4
so it's some"here ar'ound 10-for a consequence of a large i
5 accident, you know, then you begin to getiin the range where 8
a j
6l maybe the equity issue again becomes less important, but for a G
l i
7 number that low, then the implementation issue becomes much A
j
{
8, tougher, and what I have heard Tom say is that he fears that if d
c; 9
you come up with a low number for this goal, which is not i
z h
10 l sufficiently low that it means the industry can't live with it, 3
l I
h 11 then it may be used to undermine the defense in depth concept.
4 W
- j 12 It seems to me that the only practical way of 4
y 13 I implementing a goal of a low probability of a large scale
=
i 14 accident is to have some assurance that there will be a i
l 2
15 defense in depth strategy that goes along with the goal.
So s
e:,' 16 i it seems to me that by choosing the gcal the way that Tom is x
i d
17 suggesting that it be chosen, this almost necessarily-requires a
4
=
l
]
{
18l us to choose certain instruments for making this more than just I
{F 19 noble words, and I think it means that the goal itself has to be j
i n
i 20 i tied along with the defense in depth strategy that goes not as a 21 substitute but as an implementation policy.
22 So the question I would like to see is, we have heard 23 '
Ed O'Donnell say that in no way does the safety goal undermine a 24 i defense in depth concept.
We have heard Tom say that it could i
25 very well lead to that.
The question I have is, what sorts of I
1 l
ALDERSON REPORTING COMPANY, INC.
3
,.,,,,__...----.r-
+ -,. - -.., -. ~-
t 45 l
L 1
, -- what sort of connection or tie should there be between a l
2 low number for the probability of a major accident with the 3
defense in depth strategy, so that people would feel that this 4
goal is somehow meaningful and also leads to a defense in depth e
5 strategy as apposed to substituting for it.
b So I guess I am starting out with that as a question l
j 6
3 7
for Tom, but it seems to me it is important to all of us.
y A]
8 MR. LAVE:
Before Tom has a chance to answer it, we 0
0 9
will ask Helen Ingram to talk, and then Tom can.
I I
j 10 MS. INGRAM:
Well, I feel a little uncomfortable because j
11 I think Toby's question is very worthwhile to follow up on.
At a
j j
12 the same time I have been sitting over here waiting to say 5
I 13 something that I want to say, so, cecause I don't sort of allow m
i l
14 !
it, please let's don't let this go.
E 2
15 I have been thinking about the poor people who I
g 16 revised this documer.t, and trying to think about how we might be A
l a
l d
17 i useful to them, and Lester, it seems to me that you said some-E i
5 18 l thing a while ago that is really helpful in changing something i
=
I I
19 on a page hern, and that is that goals serve different purposes.
M 20 l There are different goals for different folks. All.right. So i
21l on Page 3 where there are purposes of a policy statement, there I
22 l is the beginning of some statements about the purposes goals 23 '
should serve.
1 24 ;
I think those purposes are not extended long enough.
25 We have talked, about lots of other purposes.
You like the one on l
l i
ALDERSON REPORTING COMPANY. INC.
l l
l
46 l-I i
1 understanding.
I notice that there isn't there a purpose of l
2i to reflect and articulate public concern.
Had that been a l
I 3l purpose stated here, then there is an inalementation strategy I
4 which follows that might well contai-a recommendation for public I
g 5
meetings and other ways to elicit public reaction.
There would S
i j
6l be a process recommended there, and I note in reviewing some of R
l 5
7i the documents I read before I got here, there was a lot;of j
8, discussion about process goals. That is not here any more, and I
d n;
9l it is not here because that purpose isn't here, the whole thing E.
i y
10 I has been dropped out.
I z-i j
11 l Perhaps we want'_to'do that, perhaps we don't, but I B
j 12 l think it ought to be noted.
All right.
Again, had it been 5
l 13 l stated here that there are different sorts of goals, one is an x
14 aspiratien, somewhere there in the future that we don't E
15 :
operationally know how to reach yet.
It is our fundmmental x
i y
16 l concern, our kind of fundamental value pgsition.
- The,
^
j 17 >
measures.not are we there yet but are we making progress, and one
=
5 18 l could devise an implementation strategy to test progress.
I 19 l If one needs an operational goal, to instruct the 5
\\
20 !
present ongoing regulatory process, then one needs a number you 21 i can verify, and then the implementation of that is getting a l
i 22 verification kind of procedure.
23 '
I would only remark that the implementation section as i
24 it presently is written is written only for implementing PRA.
l l
25 j That is all there is.
There is not any implementation of i
i Ii I
ALDERSON REPORTING COMPANY. INC.
I
47 i
j i
i l
l 1'
anything else here.
So when we talked about it being adequate, l
1 i
2 and Mr. Corey said that several times, insofar as it is adequate, 2
3 it is adequate just to that subject, not to the numbers of other l
]!
4 things that these goals may have been.
(
j i
]
a 5
All right.
That is all I wanted to say.
I
)
6!
MR. LAVE: Okay. Let me just make the point rather I
R l
j M
7 dramatically.
There were complaints in both Panels B and C 3
j j
8 that the current document did not reflect the richness of tne b
d
]
}
9 discussion in Palo Alto.
Well, unless people do what Helen just z
i c
g 10 did, the next document f on't going to reflect the richness of i
i z
{
r j
11 the discussion in Harpers Ferry, and so we don't have time for f
12,
everybody to go through and sort of list the substantive changes
=3 3
13 they want to make in the report, but let me tell you that unless
=
t l
l 14 you submit in writing or by telephone -- I hope in writing --
l E
i 2
15 presumably to Georts, the nature of, rather specifically, the 3
w l
l!
i j
changes that ought co be made in here, then your comments, your 16 d
i d
17 great wisdom are going to be lost.
That is, we have a transcript, u
a
=
i M
18 just as we had a transcript of Palo Alto.
It doesn't turn out 6
l j
E 19 l to be very helpful in practice when you are sitting down to try 5
i 20 :
and decide what to do.
I l
1 i
21 i So let me make that point rather strongly, particularly l
22 f for the people in my panel and the p_cople in Panel C who were l
I 23 dramatically unhappy with the lack of richness in the document.
24 Tom.
l 25 MR. COCHRAN:
Several speakers back there was attention I
i ALDERSON REPORTING COMPANY. INC.
1L...
.. ~
48 1
i j
1 l broight to Page 35 and the timing of corrective action, and I 1
j 2
just wanted to make a~ couple of comments about that.
One, I think 1
)
3 it is somewhat amusing that you have to < jet to Page 35 to find f
l 4li that all the quantitative goals should be raised by a factor out l
]
e 5
of three because you really would not have to meet the goal, you j
6, only have to meet the goal within a factor of three before you 7
would be required to do any corrective action.
N j'
d I
j 8,
Secondly --
j q
9i MR. JOKSIMOVIC:
You are talking about the operating i
\\
i g
10 l requirement there.
1
?
l 5
11 !
MR. COCHRAN:
And the second point is that the 1
4
's f
j 12 I statements about corrective action imply that these are not goals 4
I i
13 I but are standards.
I mean, we have had discussion in the previous 3=
i j
j 14 l meeting and also in this one on several occasions that there is j
2 15 h no clear distinction between -- made by the staff between whether 5
j 16 ;
this is really a goal or a requirement, and if you are requiring 1
A i
j p
17 corrective action if you don't meet the goal and it is a i
a s
e l
E 18 requirement, it is not anything less.
5
)
]
19 f And thirdly, I will just make the observation that n
20 l the.se numbers sort of appear to be codifying sort of business l
j 21 I as usual at the Commission where they find the regulations are not l
l 1
22 !
met, they don't shut.anything down with rare exceptions.
They 1
1 23 '
call for study and so forth, and give the vendors or utility 24,
lots of time to fix things.
'l 25 !
So, I just make that observation.
1 I
1 j
i ALDERSON REPORTING COMPANY. INC.
I L
49 i
l l
i MR. LAVE:
Okay, Tom, we had Toby Page's question.
j I
MR. COCHRAN: I am embarrassed to say I didn't --
2 MR. LAVE:
Toby, would you briefly --
3, i
4l MR. COCHRAN:
-- wasn't focusing on your question when 5l y u were stating it, Toby.
If you would restate it.
e R
l n
8 6;
MR. PAGE:
I will just restate it quickly. It seems to O
k7 i
me that if we have a-- okay, your first ccacern was to say that 8
people don't want large scale accidents, and that if we are I
d i
g 9
going to have a goal number of three, it seems to me that what i
S that means is that it should be stated more in terms of a large 10 i c
E 5
11 i scale accident and more in terms of a greater assurance that
<M d
12 l we are not going to have it, which means a lower probability.
z=
i 5
13 ;
If we do that, then we face the problem of implementa-cion more squarely, because this is a number, how do we know that 14 I i
w i
b i
2 15 '
we are living near it or far away from it or whatever.
The thing 16 l that I heard you say before is that the way of getting some sort s
A I
g-
- 7l of confidence that we do have low probabilities of large scale d
i ig accidents is to have redundant systems, so it seems to r.e that
=
5 the logical kind of implem'ntation for a goal that focuses 39 25 i
20 particularly on minimizing the risk of large scale accidents 21 j means it has to be tied directly to defense in depth concept.
l 22 i But we also heard you say that you fear that if quantitative 23 goals are adopted they are going to be used to undermine the l
24 ] defense in depth concept.
4 l
25j S
the question is, what would have to be done, what l
i I
1
- i ALDERSON REPORTING COMPANY,INC.
. = _ -.
1 i
n 50
~
l l
j j ;
kinds of mechanisms or assurances, rules of the game'would there L
2 have to be in order to tie.an active use of defense in depth i
3 strategy to a quantitative goal like this number three but j
4 sharpened up to focus on a low probability of large scale 1
g 5
accidents?
I R
4 j
6 MR. COCIIRAN:
Well, I think you need an overhaul of J
4
- g 7
the whole system, the whole regulatory syctem.
You know, you are
,f8 asking sort of can you take the mad dog to obedience school and j
e j
c 9
make him sit on command.
I mean, it just may be that you are i
f h
10 not going to get what you are asking for here, and I think that z
4
]
l ij is the situation we are in.
1 a
)
d 12 I think the public looks at -- like myself, they're t
E I
)
pretty simple minded people',. and they say, they have entrusted 13 j
i j
14 j this into the hands of the Commission, and they say, we don't W
l 5
l i
j E
15 l want any accidents.
We don ' t want any TMI's, and we don't want
}
16 li any catastrophic accidents, and we don't want these things to a
1 x
l l
g j7 melt down.
We want them to be safe.
And that ought to be what
~
a w
x M
18 the goal
-- I mean, that ought to be the standard.
And that E
1 19 ought to be what the Commission is articulating.
i 9
i a
j 20 !.
And then they've got to have a regulatory structure i
21 and some decisions coming out of the NRC that gives people i
22 confidence that that is occurring, and you are not going to get
-24 l
23 that with this mumbo-jumbo of 10 mean probability estimate l
24,
trade off with some cost benefit stuff.
I mean, you are going l
25 to get it with a lot of other things, I mean, with some, we are l
1 i
ALDERSON REPORTING COMPANY, INC.
l i
51 l
~
I 1 l not going to have any more hydrogen explosions because we are l
2i going to switch'to stainless steel rods, or something, so we I
3l don't get the same kind of reactions, and we are not going to l build any more containments that won't contain a full core 4'
5l meltdown.
I think you've got to have some very sweeping changes.
e 9
j 6i The staff in this document and the Commission are a 5
7 zillion miles away, I think, from what people are looking for.
i "j
8, They have got a goal here that will allow 21,000 cancers from the d
l d
9; reactors they are going to build.
I mean, that is not what i
10! people are looking for.
z 1
=
i g
11 MR. LAVE:
Okay.
We are now already eating into our 3
d 12 '
lunch hour.
We have three people who have asked to talk, and E
=
1 s
13 i I would like to say a brief word _ after the three people, but E
l s
14 : why don't we do the following?
Let's change the procedure a little d
2 15 ;
bit.
Whenever you think that it is time to go to lunch, turn up J
16 j your card, and when I see a majority of cards turned up, I will E
y 17 go to lunch.
a
=
5 18 '
(General laughter.)
E l
I 19,
=n 1
20 j 21 '
22 23 '
24 l
25 li
{
i i
I i
I 1
ALDERSON REPORTING COMPANY. INC.
i
I l
52
!1 i
alRC I
gj7. jg y l
MR. PERPOW :
I think that I'm a little surprised lKIL~dLEA/ 2
]StlTTON at your statement that if we want to make changes to this, we j
3 ITAP-3 have to submit it in writing. If my recollection of the last j
4i j
meeting is -- we spent a lot of time discussing all these goals; 5
e 5
i and at past meetings, we spent a lot of time writing up a thing l
I N
6!
l l
about it and sending it around.
I know that I gave statements g
I E
7
~
l
-- comments -- on this thing in writing.
I know the Court gave I
8 8
]
long comments in writing; I know that Paul revised in on that 1
]
d 9
y basis.
And then all of that in writing went in, and this docu-3 E
10 I j
j l
ment that I got in the mail has less to do with what we were E
11 concerned about than the ACRS document -- which was much more 1
i d
12 ;
j open to these concerns than this one.
j d
13 1
}
E So this was stuf f in writing, with some effort on I
E y 14li the part -- a lot of effort -- on the part of all of us; some l
2 15 i s
l ef fort -- a little bit -- by the rest of us.
And the effect was
)
3.
16 '
i minus.
So to say now look, if you want your concerns considered j
g 17
)
E in this thing, you've go t to write out a new version of this
)
E 18 l
E report and send it in.
I don't -- just don't believe that we l
{
I 19 -
A representing -- if we call ourselves experts, I don't think we're 20 representing expert opinion let alone public opinion to this way 21 l i
of thinking.
We're going backwards.
)
22 l
MR. SEGE:
Let me give you two reactions to that:
One
]
23
)
is, you don't need to write it all up again, but simply a note 3
I
' 24 -
l that says you didn' t pay enouch attention to what was in the i
25 l j
j panel's C Report is fine.
The second one is -- not now, but ALDERSON REPORTING COMPANY, INC.
1 t
j ri.r
i 53 I
~
sk 2 af ter lunch --- I think that you ought to ask the question that 2i why is it that what was in that was not as helpful as it might I
I 3'
have been, in revising it since that might tell you how it is l
4 I that it might be reformulated to be of more help next time.
e 5
g i
Mel Ernst.
n 3
6-MR. ERNST:
I think Tom raised some excellent quus-L'n 7
tions about the table, and I might like to spend one minut.'
n 3
8I l
trying to get some perspective which isn' t in here.
o c
9!
j MR. PERROW:
Okay; I'll give you exactly 60 seconds.
~
c 1
b 10 i y
MR. E RNS'2 :
Okay.
! 11 l!
j In the first place, the table sounds very precise in d
12 l j
factors of three and ten and so forth; the intent is actually 5
13 '
quite the opposite I think, and we ' re trying to come up with some d
14 d
decision guidelines.
There is clearly no clear go or no-go i
=
9 15 j
break point on acceptable risk as posed. in PRA because of uncer-T 16 j
tainty and the fact that there is not that clear an idea of what F
17 j
would be acceptable even it it's completely certain.
5 18 i
There has to be room for some cost effective deci-r C
19 3
sional process as to whether one recuires a chance or not; so 20 '
l that is some of the reason fc-the so-called sloppiness or what-21 ever in the factor of three.
Also, if you have a plant that 22 l exceeds the risk by a certain amount, there will probably be a 23 dominant sequence contributing to that.
So even though it looks i
24 i
Ilike we' re trying to call the shot very fine, in practicality, 25 :
you will find the doninant secuence, more than likely, which can l 4
ALDERSON REPORTING COMPANY, INC.
l 1
l 54 l
.k 3 I
be fixed and will make a substitute change to the risk --
rather l
i i
l 2:I than just driving for a.99 of the acceptable limit as opposed l
l 3
to 1. 2 er :omething like that.
i l
4I So that was the intent of the table.
J g
5 MR. SEGE:
Okay.
S j
3 0 l, I was going to spend a minute giving you a surmary 1
m*
7 of this session; but given the hour, I will use some time after u
g 8
lunch to do that.
So, why don' t we co to lunch?
J e
i 9'
.{
Oihereupon, the morning session was adjourned, with the hearing z
i i
=
"h IO l to be resumed the same day. )
i "A
11 I I
3 l
- 5 12 l z
l Ei 13 !'
g I
14 '
id M
l 2
15 e,
u
?
16 ;
a
=
l d
17 E
i 5
18 7
19 5
t n
1 2a!
21l ni 23 24 1
25.
I i
l t
- )
NUCLEAR REGULATORY CO.WISSICN This is to certify tha: :he attached proceeding: before the
.)
~
4 d
in the satter CfL SAFETY GOAL WORKSHOP - SESSION 5 - IMPLEMENTATION
- Dane of ?roceecing: July 24, 1981 Decket !!u=b er:
? lace of ?roceeding:
Harpers Ferry, N.
VA.
were held as herein appears, and tha: this is the Original transcrip
- hereo f for the fi' e o f the C~ mission.,
Rossie Sutton l
Official Reporter (Typec)
\\:3 b
2Al JJ) i Cfficial Reporter (Signature)
F
!'.3
.