ACRS/ACNW Joint Working Group U.S. Nuclear Regulatory Commission, May 11, 1999
UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
MEETING: ACRS/ACNW JOINT WORKING GROUP
U.S. Nuclear Regulatory Commission
Two White Flint
11545 Rockville Pike
Tuesday, May 11, 1999
The Joint Committees met, pursuant to notice, at 8:31 a.m.
B. JOHN GARRICK, ACNW Chairman
GEORGE M. HORNBERGER, ACNW Co-Chairman
RAYMOND G. WYMER, ACNW
CHARLES FAIRHURST, ACNW
THOMAS S. KRESS, ACRS Co-Chairman
GEORGE E. APOSTOLAKIS, ACRS
. STAFF PRESENT:
Andrew C. Campbell, ACNW Staff
Lynn Deering, ACNW Staff
Howard J. Larson, ACNW Staff
Richard K. Major, ACNW Staff
John Sorenson, ACNW and ACRS Fellow
Carl Paperiello, NMSS
Seth M. Coplan, NMSS
. P R O C E E D I N G
DR. GARRICK: Good morning. A little history in the making.
This meeting will come to order. This is a meeting of the
Joint Working Group of the Advisory Committee on Reactor Safeguards and
the Advisory Committee on Nuclear Waste. My name is John Garrick,
Co-Chairman of the Joint Working Group. On my far left is Dr. Thomas
Kress, also Co-Chairman of the Joint Working Group.
The Joint Working Group members in attendance are George
Apostolakis of the Advisory Committee on Reactor Safeguards; George
Hornberger of the Advisory Committee on Nuclear Waste; and other
committee members present are Charles Fairhurst and Ray Wymer of the
The purpose of this meeting is for the Joint Working Group
to discuss the NRC Staff's proposed framework for a risk-informed
regulation in the Office of Nuclear Materials Safety and Safeguards.
The Working Group will gather information, analyze relevant issues and
facts, and formulate proposed positions and actions as appropriate for
deliberation by the full Committees.
Richard Major is the Designated Federal Official for this
The rules for participation in today's meeting have been
announced as part of the notice of this meeting previously published in
the Federal Register on April 21, 1999.
A transcript of the meeting is being kept. As usual, it is
requested that the speakers first identify themselves and speak with
sufficient clarity and volume so that they can be readily heard.
The Committee, the Joint Committee, the Joint Working Group
has received no written comments or requests for time to make oral
statements from members of the public.
As you know, the specific motivation for this particular
Working Group was a recommendation from NMSS to provide technical peer
review of the Staff's future efforts, to develop a framework for
applying risk assessment methods to the regulation of nuclear material
uses and waste disposal.
I think that we will, with that, get right with it and ask
Dr. Carl Paperiello, Director of the Office of Nuclear Materials Safety
and Safeguards, to begin.
DR. PAPERIELLO: Thank you. I am pleased to be here today,
and I would like to talk about risk-informed regulation within NMSS. If
I could have the -- why don't we go on to the third slide. The second
slide just says the -- an outline, to speed things up here.
In my view, what risk-informed regulation is about is two
pieces, risk assessment and risk management, and they are two separate
components of it, and in many ways somewhat different.
Risk assessment is fairly scientific. Risk management has a
great deal of political involvement. The question is, if we know the
risk, what is acceptable, and that has been a target which has varied
over the ages. If we live in an era where people are starving to death
and suffering from infectious diseases, worrying about getting cancer
when one is 70 years old is really not all that important. You may not
live that long.
To talk about the materials program, under the Atomic Energy
Act we have identified more than 40 different kinds of non-reactor
activities, devices and systems that are regulated by approximately
20,000 licensees. Effectively in NMSS we regulate everything the agency
regulates with the exception of reactors, and even reactors we get when
they are in the final stages of decommissioning.
Most of the regulation we do, or most material in the United
States is regulated through the Agreement States Program, not actually
through NMSS. Now we argue about volume of material, I have more curies
than the Agreement States because I have spent fuel. But in terms of
numbers of licensees, most of the licensees are in Agreement States.
The Atomic Energy Act, of course, provides for the Agreement
State Program. We have approximately 30 Agreement States with several
more states, Pennsylvania and Ohio most notably, likely to become
Agreement States in the near future.
Under this program we relinquish -- we don't really
delegate, we relinquish authority to the Agreement States. The states
have to have programs which are adequate and compatible, obviously words
which have a different meaning to different people and always create a
certain amount of controversy. We have a program called IMPEP where we
assess all the programs to ensure that they are adequate and compatible,
and agreements can cover most areas of NMSS regulation with the
exception of high level waste greater than Class C waste. Spent fuel
storage, transportation, and fuel cycle facilities. And states can
decline to regulate in certain areas. They can have partial agreements,
which happens. Some states choose not to regulate mill tailings. Some
states have returned to NMSS the sealed source and devices program.
And also under the Atomic Energy Act, section 274 that
provides for the states to become Agreement States, it also has a
provision which affects my program in that we are to coordinate the
development of radiation standards with the Agreement States. So,
therefore, when I want to change my program -- for example, change the
regulations in the given area -- I have to consult with the Agreement
States, and so therefore this is not a sort of a dictated program,
saying, okay, we're the boss. There is a certain amount of
give-and-take, and that creates, when we are trying to implement a
risk-based, a quantitative -- this is what we are really doing -- a
quantitative risk base, we have always had risk base within NMSS. It
only has been qualitative and has been based on judgment.
For example, we license big licensees out of the
headquarters. The smaller licensees are licensed out of the regions.
We have provisions for general licenses. We have provisions for exempt
quantities. So there is a graded approach in licensing. There is a
graded approach in inspection. We don't inspect everybody at the same
frequency or the same level of effort. But it has not been done on a
quantitative basis in the past.
DR. APOSTOLAKIS: So let me understand this a little better.
If you want -- how do you make sure that the states maintain an adequate
regulatory program? Do you inspect or -- DR. PAPERIELLO: Yes, we
basically have a -- we have teams made up of Agreement State, NMSS and
the Office of State Programs staff. We visit them with a frequency of
once every two years to once every four, depending on previous
performance. We take a look at their regulations, are they up to date.
We take a look at the training and qualification of their licensing and
inspection staff. We review a sampling of their licensing activities to
see that the licensing work is done adequately, review a sample of their
inspection report to see that inspection reports are done adequately.
We accompany Agreement State license inspectors on some inspections
again on a sampling basis, and then we take a look at how the Agreement
States handle allegations, response to events and when they have
programs for sealed source and devices or low level waste or something
like that, we look at -- we do the same thing for those programs.
We then -- the report is written. There is a management
review board that is held for each of these states. By the way, we do
the regions the same way, same program for the regions, too, except the
regions are done at a greater frequency. They do them every two years,
regardless of their performance. From my viewpoint, they work -- what
they do is delegated from me, so I now have an obligation to make sure
that the resource expenditures are reasonable, too. So there is a
little twist with the regions.
The management review board includes people like myself,
Karen Cyr, other senior managers of the agency, as well as an Agreement
State rep. So we review the report, the state people have
representatives there, and then we make a --you know, we put out a
report on our decision, and the findings of whether programs are
adequate and compatible; also whether certain areas -- you could have a
case where the inspections aren't completely up to date. We might turn
around and make a finding of satisfactory with recommendations for
But, anyway, that's the program for the oversight of the
DR. KRESS: Is there any concept in there -- I almost hate
to say the words -- below regulatory concern on some of these things?
DR. PAPERIELLO: With respect to what?
DR. KRESS: Well, I suspect you choose a list of things that
you are going to have to license and regulate. Do you choose --
DR. PAPERIELLO: Well, we do have exempt distributions. I
mean smoke detectors --
DR. KRESS: Okay, that would be one.
DR. PAPERIELLO: Yes, smoke detectors, luminous dial
watches, still the use of source material in certain glasses; there's a
whole list of exempt -- in fact, right now we are reviewing all the
exemptions because most of the exemptions were made up years ago without
a -- with not -- again, not as strong a quantitative view of risk as we
now have today.
DR. KRESS: So risk might be the consideration one would use
for those exemptions?
DR. PAPERIELLO: That's exactly -- well, that would be the
basis, but again, years ago, when many of these exemptions were written,
we weren't as rigorous as we are -- I mean some of the exemptions that
are written in the Part 40 were based on the fact they were uses that
predated the Manhattan District. For example, glazes on dishes, uranium
glazes on dishes. That was something that went back, was in the '30s.
So when they wrote the original Part 40, they made these things exempt.
And the interesting thing, because we are in the process, under
direction from the Commission, of revising Part 40. In the original
Part 40 they were exempt, but you couldn't manufacture them, and the
reason is they wanted to save every gram of uranium for the enrichment
plants, and not waste it. That's my view of the thing because it was
very, very strange; you could watch the evolution of Part 40 and the
expansion of the exemptions and the general license as uranium became
more plentiful. So it was not -- it was a national security issue, not
a matter of health and safety.
Let's talk about them. They vary widely. And again, the
way I put it, I regulate everything from the production of luminous dial
watches and smoke detectors to spent fuel and strategic quantities of
SNM; very drastic differences in the types of things that we regulate.
The hazard accident potential and failure modes vary. Human
error tends to be a very frequent cause of system failures. Think about
university research, where very small quantities are used, people knock
things over, drop things, things boil over, so --
DR. APOSTOLAKIS: So when you say human error, you think of
DR. PAPERIELLO: Yes.
DR. PAPERIELLO: You know, a medicine radiographer. Even
irradiators, read anything, whether they have to be by-product material
or electronic sources, large irradiators, most of the errors are human
errors in judgment, things like that.
We have -- in some systems we have a reasonable amount of
data on errors. In some other systems we have almost no information.
For example, we know from experience that the major cause of
radiography overexposures is failure to survey, but what percentage of
the time do radiographers actually fail to survey? I mean I could draw
a very simple -- and a very simple event tree exists for radiography
overexposures. It's in one of the recent ICRP publications. When I
talked to one of the authors of that document, Don Coole, I said where
did your numbers come from, and they were guesses, because we don't know
the data, we don't have that information. And the communities vary in
terms of technical sophistication. They obviously vary very differently
from reactors, and the people sophistication varies.
You know, fuel facilities will have regulatory affairs
managers who not only are involved with us but are involved with the EPA
and OSHA and other regulatory agencies, down to small gauge users,
engineering companies who happen to use moisture density gauges and the
like will have somebody whose ancillary duty will be to take care of the
I have already -- one showed up early in my career, when I
first became a section chief at a fossil plant that three years before
had installed level gauges, and when we showed up with this license,
what license? And eventually, of about five or six users named on the
license, there was one that was still at the plant. So it was a -- you
know, that --sophistication varies significantly.
In some cases NMSS in the past has given explicit
consideration to risk. The -- we have incorporated an old EPA standard
that's no longer valid into the Part 60 that required that the risk
curve for releases of radioactivity from proposed geological repository
meet certain limiting criteria. Now we know that's not the standard any
more, but at one time we did.
Secondly, we are in fact using risk when we set radiation
dose limits. However, we don't always recognize it explicitly. For
example, when we set a limit of 25 millirem per year for the average
member of the critical population group for high level waste repository,
we in fact have set a risk standard because one can use the linear dose
model and translate that into risk, and we could argue, well, we don't
like the linear model. But the practical matter is on the reactor side,
when you are looking at latent cancers, you are using a linear model.
And the agency in -- and I don't remember the NUREG any more -- back in
around '86 or '87 endorsed not only the linear model but endorsed
certain of what we now call slope factors. And the ones we use today
aren't a whole lot different than what's in that NUREG.
And, in fact, when you use $2000 per person rem, you are in
fact using the linear model. So we -- and I don't want to get in
arguments of whether the linear model is valid at 25 millirem per year
-- my personal believe is it isn't, but the practical matter is that's
what we do -- and so, therefore, you are in fact having -- establishing
a level of acceptable risk.
Since the mid-'70s, we have pursued PRA, probabilistic -- I
shouldn't say PRA, because that's one thing I'm trying to -- if there's
one campaign I have, I'm trying to establish the idea in everybody's
mind that risk assessment and PRA are not the same thing. PRA is one of
many techniques for doing risk assessment, and it may not be applicable
to every area that we regulate in NMSS, but in particularly performance
assessment for waste, whether it's high level waste or low level waste
or decommissioning, or the transportation, the publication we call the
Modal Study essentially is a PRA applied to severe accidents in
DR. APOSTOLAKIS: I think we should clarify the use of the
acronym PRA. From the way you describe it, I think you mean risk
assessment as applied to reactors. Because PRA, in my mind, is really a
conceptual approach that is applied in a certain way to nuclear power
reactors in a different way, but conceptually the same, to high level
waste repositories. So it seems to me that the use of language, since
we are beginning this now, is very important and we should agree on what
we mean by PRA and what we mean by other things.
In other words, are PRA and performance assessments
fundamentally different, or they differ in the tools that they use to
implement the same concept?
DR. PAPERIELLO: I think it depends on your definition of
DR. APOSTOLAKIS: I know. That's why we should agree on a
DR. PAPERIELLO: That's why I would just like to say risk
assessment and then recognize that some of my risk assessment could be
purely actuarial. Because in my side of the house, I have a lot of
actuarial data. I have data all the way to the severe consequence. I
know -- I have empirical data on what happens when radiography sources
get loose, both in injuries and death to radiographers and injuries and
death to members of the public, at least on a worldwide basis. So, you
know, there are some differences. Now I --
DR. APOSTOLAKIS: Go ahead.
DR. PAPERIELLO: I'm not a practitioner of any of these
things. I'm sort of a -- you know, at my level I'm sort of just a --
I'm a physicist by background, so I -- it's --but we're looking at it,
at least within the agency. PRA is a very, you know, a bunch of event
trees, you have your fault trees, you have data from mechanical
failures, and I deal with a number, and the differences I see between
performance assessment and probabilistic risk assessment is at the
various Ts, or your various points where things come together, you use
binary values in PRA for each of these, where we use distributions in
performance assessment, at least. And I may be wrong. I'm telling you
as a person who is not a practitioner of any of these methodologies what
it looks like from the outside in.
DR. APOSTOLAKIS: I don't think it's a matter of being right
or wrong. It's a matter of convention, really, of language, and maybe
that should be one of the first issues that this subcommittee takes up,
to agree on the language.
DR. GARRICK: Yes.
DR. PAPERIELLO: If you look at the American Chemical --
American Society of Chemical Engineers books on integrated safety
assessment, you will find PRA in that paradigm as a subset of ISA, okay?
So it's the most quantitative, you know, version in there, but the PRA
then becomes a subset.
DR. APOSTOLAKIS: No, I know that there are different -- and
also they call it QRA, quantitative risk assessment. So I think before
we go into more detail of these things, we should agree on the language,
perhaps not right now, but certainly our letter should say something
DR. GARRICK: Yes, the language is a major part of the
problem, and you are right, there's actually three acronyms that are
used internationally that really mean the same thing, and they are PRA,
QRA, and PSA, with probabilistic safety analysis being the preference of
the international community for PRA, and PRA being pretty much an NRC
acronym, and QRA being the preference of the space industry and the
chemical industry, simply because they don't like PRA. So -- but they
all really are talking pretty much about the same thing.
I would agree with one thing, that risk assessment is
something that is rather fundamental, and I think there is agreement
pretty much on what we mean by a risk assessment. I would not
necessarily agree that when I think of PRA, for example, I necessarily
associate with it fault trees and event trees. I think that's a detail.
It's like saying where do you get your probabilities from, and some
people say MONTE CARLO. You don't get probabilities from MONTE CARLO.
MONTE CARLO does the probability arithmetic. It does not generate the
fundamental probability. So I think a way to get to this issue of
language is maybe to think in terms of very basic definitions, and then
go from there. And you're right, it's a big subject and we need to
evolve to it.
DR. PAPERIELLO: Okay. Well, in any case, when I use
probabilistic risk assessment, I use it within the framework of what we
do in this agency for reactors, and from my view it works extremely well
in mechanical systems; not just mechanical, but complex systems, many
interact as -- again, I'm saying from the outside in, I have a system
that should not fail, yet it fails, as multiple redundancies and all
that kind of stuff. I have a lot of systems that I regulate with almost
no redundancy, or the redundancy are human -- are procedures and human
What we are looking at and the subject of our paper is we
have identified sets of risk assessment techniques that we believe cover
most of the things we do in NMSS. This is something we did not really
have identified a couple years ago. We believe that using performance
assessment is effective for dealing with waste disposal systems and
decommissioning. The set of techniques that are encompassed by
integrated safety analysis, at least as defined by the chemical
industry, is an effective tool for analyzing fuel cycle facilities and
perhaps some transportation. And hazard barrier analysis is being
developed for the multiplicity of systems that we regulate in what we
call our materials program, but basically through Parts 30 through 39
and some of the manufacturing and non-fuel uses of source material.
We want to increase our use of risk assessment. The current
revision of Part 70 will require a fuel fabrication and enrichment
facility to perform integrated safety analysis. Actuarial methods are
being considered for our various material uses, byproduct material uses,
where we have large sets of data that involve many similar licensing
systems. We now have responsibility for the nuclear materials events
data base, and we gather data not only from NRC licensees, but also from
Agreement State licensees. And we are looking for simple predictive
risk analysis methods for our simple systems.
The Commission has directed us to develop a framework for
the use of risk analysis and decision-making, and of course we discussed
this in 99-100. The problem is -- okay, and we now have risk -- let's
suppose that we have a risk assessment technique, or we have -- we can
cover most of our -- the things that we regulate with some kind of
quantitative or semi-quantitative risk assessment, what should be our
I think it's important to define a safety goal for NMSS.
And I call it the problem of the target. The -- who should be the
target? I could pick a goal. I'll make a number up, one in a million.
Everybody assumes that one in a million is negligible, at least for the
society that we have today. I assume if we lived forever, the
probability of death in an accident like automobile accidents would be
considered completely unacceptable. But as a practical matter, the
automobile death rate shortens the average life span in the United
States for about a year or so out of 70 something years. I assume if we
have much longer life spans, it would be a bigger -- natural life spans,
it would be a bigger number and we wouldn't find that acceptable any
But the question is, what's the target? If the entire
population of the United States is the target, you know, if I use 10
miles around every materials facility, I'd probably blanket the country.
That would mean I could accept 270 deaths a year, or a half a million
person rem. And any estimate of dose to the American population from
things that I regulate, putting aside medicine, is probably only on the
order of a couple hundred, perhaps a thousand, person rem. So I mean
drastically higher exposures than now.
If you turn around and make it no one individual is going to
be given a risk of 10 to the minus 6, you are going to be dealing with
limiting the public dose to 2 millirem a year. So the question, I see
as a major function defining my target for a given practice that I
So the reactor safety goal, at least as written, is not
There's the matter of cost/benefit. Now, remember, I put
aside medicine, and it's clear in the regulation of medicine, we allow a
lot higher exposures, and we don't even -- and I'm putting aside the
patients. We are talking about patient release criteria right now of
500 millirem to a member of the public, we assume mostly family, because
the person who would get the treatment was presumed to be suffering from
cancer, and probably be returned to a family environment.
We have petitions for rulemaking which would allow visitors
to patients undergoing implant therapy or radionuclide therapy to get
doses up to 500 millirem, at the same time we are setting dose standards
of 25 millirem a year for high level waste. So I'm just saying -- and
it's estimated by NCRP that the release of patients with -- having
undergone diagnostic procedures is roughly 8000 person rem a year, NCRP
1.24, and that would give you, if you used the linear dose model
collective dose and put aside all the things which I don't believe, to
about four deaths a year, four induced cancers from patients being
released -- not to the patients, but to members of the public.
So clearly cost/benefit comes into the thing. Medicine gets
a wide degree of latitude because it's presumed to benefit -- be of
great benefit to the patient.
So, anyway, this is an issue. So we are interested, we are
very much interested in when the agent being included in any agency
effort to establish a safety goal for the agency and a decision needs to
be made whether or not there is going to be either consistent with the
reactor safety goal, which effectively is going to be a question of
who's defining who to target, or else just having multiple safety goals
for materials. You know, decommissioning safety goal will be different
than an operating safety goal. Just as long as there's a decision.
Because the question is then how do I use risk-informed regulation?
What am I -- at what point do I stop regulating.
DR. APOSTOLAKIS: Are you going to ask the Commission to
DR. PAPERIELLO: I'm sorry?
DR. APOSTOLAKIS: Are you going to ask the -- or suggest to
the Commission that they should consider establishing safety goals for
DR. PAPERIELLO: Yes. I mean this is part of -- you know,
what we've had -- this is our second -- 99-100 is our second paper on
the topic, and we are progressively moving, you know, moving down.
Because I think we have established now a good sweep of risk analysis
I'm in the process of forming a group within IMNS that will
deal with risk assessment. In other words, I don't -- I have
practitioners for some of these techniques; I don't have practitioners
for all of them; and I need a group that can use all the tools we are
relying on for some things on contractors right now.
As I pointed out, we have -- we don't have a lot of
uniformity in materials. We have different degrees of sophistication in
different users. We have incorporated risk in varying degrees. I think
we are moving -- what we want to do is move to greater rigor. The
technique we choose has to be appropriate to the technology we are
regulating. We think what we are doing -- we are trying to accomplish
-- to use things which are analogous to the reactor regime, and I guess
we are looking to this committee for, you know, advice and
recommendations on the proposals that we have, you know, offered up to
DR. GARRICK: Questions for Carl? George? Tom?
DR. KRESS: Well, I agree with you that the primary need
before you can do what we are trying to do is to have targets. I call
them acceptance limits on regulatory objectives. And I also agree with
you that safety goals for reactors are not going to be very useful to
you, but when you get around to setting these, I think you will have to
come up with some sort of common metric for all of these different
things. And it seems to me like the common metric is a cost/benefit.
Somehow you need to decide what your target cost/benefit is, and then
figure out how to set the goals based on those, and I would -- you know,
there should be a concept in there of lives saved are worth more than
statistical lives lost, especially in the medical field, when you do the
cost/benefit. But I do think that's where the biggest problem is going
DR. APOSTOLAKIS: But I think there are some uses, though,
where there are other dimensions of risk that become important, and one
that comes to mind is the distinction between voluntary and involuntary
risk. I mean can we really dictate to somebody what his life or her
life is worth? And when it comes to medical uses, as you pointed out,
DR. PAPERIELLO: Right. And of course, that's the basis
between public dose limit and worker dose limit. To my mind, the
essence of occupational worker is informed consent to the higher risk.
DR. APOSTOLAKIS: Yes. Well, I actually had a question. I
forgot. It's really a comment. In the reactor arena, as you pointed
out, the regulations have been risk-informed from day one, it's just
that risk was not quantified, and what drove us to the risk-informed
approach is that in the last 25 year we made significant progress in
quantifying risk, PRA, you know, since '74, with the reactor safety
study and all the other studies that followed. So there is a driver
Now it seems to me in your case, the driver is still the
reactor side, because you have not -- have you really made the same
advances in quantifying risk in the last 25 years, or are you just being
forced to consider risk-informed regulation because the other side is
doing it and the agency wants to have a uniform approach to things?
DR. PAPERIELLO: I think it's a combination. I think it's a
combination. I think we never thought about it in some areas. In other
areas we were forced to, we had to do it when we started dealing with,
you know, waste.
DR. APOSTOLAKIS: Sure.
DR. PAPERIELLO: We did it for transportation because
transportation, at least the rules up to now, Part 71 has been based on
international standards, they have been based on international consensus
that was deterministic. And the question came down to us, okay, I test
a Type B package by burning -- by having a certain temperature fire for
a period of time, submerging it in water and dropping it and all that.
Well, then you get the -- okay, what happens if this canister is
crossing the Mississippi River and winds up going into the Mississippi
River or falls down the Grand Canyon gorge, you know. We can all make
up accidents. So back in the '80s, we did the modal study to look at
what would be the equivalent of severe accidents for packages, and we
defined a set of risks, and spent fuel is the issue that has driven
I think in the balance of what we have done, we are being
driven to be more quantitative. On the other hand, when you look
around, you turn around and say why not? It's just been ignored. And,
in fact, when you start looking for techniques, you find the techniques.
The chemical industry was driven to use ISA because they couldn't
tolerate severe accidents with chemical plants like occurred in Bhopal
in India, and some other accidents they had. But the question comes
down to, when you can cover a certain amount of area, you are going to
find -- in my view, I find it attractive, is asking why we have a given
regulation, you know, in a given area.
I have been around about almost 25 years in the agency, and
I have seen singular events lead to significant changes in a regulation.
So we have an event and we wind up changing the regulations for one
event. I'm talking about on the material side. We know about Three
Mile Island and what happened after Three Mile Island, but you turn
around, and now -- that's no consideration of probability. It was based
on consequence. I mean risk the probability of an occurrence times the
event. So you have one event, it's serious without considering the
probability of it occurring. We have revised the rules. And so the
question now is, is do we need -- is explicit consideration of the risk.
But once you explicitly consider the risk assessment, you
now have got to ask what's your risk management. Because what you did
in the past is you blended the distinction. My assertion is that some
of that is still going on.
When I asked at the Seville conference about a year and a
half ago, the ICRP what their -- when they made a statement about
acceptable risk, what the scientific basis was, they said they didn't
have any. So whereas a lot of the conference was devoted to risk
assessment, and they talked about their standards, when you took the
risk management piece of it and you started delving into it, they didn't
have a good basis. It was an assertion. And so now -- and so in my
mind, I have been very conscious of the risk management because it's a
-- it's not a completely scientific decision. It's a decision that's
going to be -- society's going to make a distinction what the risk is.
For example, the statement the difference between voluntary
and involuntary risk. Why? Psychologically I understand it, but the
end state is the same. You know. But I'm just saying it's -- but, see,
that's my whole point, it's a -- and we recognize it in other cases. I
would assert that many of our voluntary risks are not voluntary at all.
If I want to work, I have to drive a car, so I mean I would assert a lot
of our voluntary risks are not voluntary.
DR. KRESS: But you recognize, of course, that you always
end up in this business with this input, that society -- that you don't
have a technical basis for those risk acceptance levels.
DR. PAPERIELLO: Oh, I well --
DR. KRESS: Yes, they always have to be fixed some way and
it's mostly societal judgment type thing.
DR. PAPERIELLO: Oh, I understand that, and I recognize it
in a democracy, and it's an incredible societal judgment. And I don't
think that's wrong. I'm just saying you need to -- as a regulator, I
need to recognize -- I need to be very conscious in what space I'm in
when I deal with risk, that management of risk is a very different
animal than assessment of risk.
DR. GARRICK: There are a couple of things that kind of
impact my thinking based on this discussion. One is that it seems that
whatever direction we go with materials, there ought to be some
connections with what we have done in the past including what we've done
in the reactors.
MR. PAPERIELLO: Sure.
DR. GARRICK: In other words, the logic has to be there. If
we are talking about a safety goal for materials, somehow there has to
be a nexus between that safety goal and the reactor safety goal, given
the overall responsibility of the Agency. And maybe that is the idea
that Tom was getting to with the cost-benefit concept. There has to be
MR. PAPERIELLO: I understand.
DR. GARRICK: That bridges --
MR. PAPERIELLO: Let me --
DR. GARRICK: That bridges across these things in order to
assure the public that the NRC --
MR. PAPERIELLO: Right.
DR. GARRICK: Is not ad hoc already doing this, that there
is consistency in their view of how safety should be regulated. That's
MR. PAPERIELLO: I would say I would agree with you
emotionally, but on the logical level I'll point out we don't do it in
the United States.
Let me point out transportation. If I recollect, and don't
hold me to the year, I believe the Department of Transportation's safety
goal for the United States, under GRPA, their outcome indicator, is
43,000 transportation-related deaths in the United States for '99.
Don't hold me to the year; I don't know, '98, '99, or the year 2000,
something like that. Now obviously in a split mostly with automobiles
and very small for aircraft, and rationally we all accept why shouldn't
we have the same safety goal for automobiles as aircraft? Yet we don't,
because we all know that aircraft mile for mile is a lot safer.
Why don't we have the same safety goal in the United States
for my view, my side of the house, technologically enhanced natural
radiation as Atomic Energy Act and we don't? I'm just saying
emotionally I agree with you, we ought to have the same safety goal for
the both sides of the NRC house, and I kind of endorse that. Yet I
recognize that's not true for many comparable risks in the United
DR. GARRICK: But to get to that, we don't want to set the
example of not being able to --
MR. PAPERIELLO: I understand that. I'm just --
DR. APOSTOLAKIS: But it seems to me, though, again it's a
matter of language, when you say comparable risk, you mean comparable
numerical values, but risk is multidimensional. The numerical value is
one aspect of it. We talked about voluntary/involuntary controllability
of risk, for example. That's perhaps a better argument, that when you
drive your car, you feel that you're controlling your fate.
MR. PAPERIELLO: Let me --
DR. APOSTOLAKIS: So the numerical values do not necessarily
have to be consistent with each other, because there are other
dimensions that make us use different numerical values. But I would
agree with Dr. Garrick that we have to agree on the basic philosophy.
The philosophy can be the same; the approach, the conceptual approach,
can be the same among the various offices of the Agency, even though the
particular implementation may differ in terms of numbers or how we do
things. So we have to go to a higher plane, I think.
DR. KRESS: And to expand on that just a bit, I think George
would say a risk analysis by any other name is still a risk analysis.
DR. APOSTOLAKIS: It's like a rose.
DR. KRESS: Yes.
MR. PAPERIELLO: As I said, there's risk analysis, and then
there's risk management. They're two very, very different animals.
DR. APOSTOLAKIS: Yes.
DR. KRESS: Oh, I agree with that certainly.
MR. PAPERIELLO: Let me make an observation about reactor
risk. When you look at risk over a ten-mile radius, if we truly had a
metric system and we used kilometers, would we have chosen the risk over
a 16-kilometer radius or would we have rounded it to 15 or rounded it to
DR. KRESS: It doesn't matter, and the reason is is you
fixed your acceptance criteria that you have for that risk in
conjunction with that number. So, you know, you can make it smaller and
make your acceptance value smaller; make it bigger, make the acceptance
values bigger. You can balance those things off.
DR. GARRICK: In fact, from a risk assessment, there have
been a number of studies done that indicate that if you want to do it on
a PRA basis, it's probably not 10 miles, it's probably a couple miles,
because 95 percent of the fatalities are within -- on a large early
release are within a mile or mile-and-a-half. So that right, you pick
the number, and then -- the other point I wanted to make was I think
that in the materials area we do have an advantage, and that is to take
note of all the mistakes that have been made in the past, and to try our
best to avoid them.
One of the things in that regard that I think we want to
focus on is the matter of simplicity. Is there a way to keep this
process from going exponential in terms of process?
So the lessons learned from the safety-goal experience on
the reactor side is critically important here. There's a lot of people
in the industry that would argue that there are no safety goals in the
reactor side either, because they're not implemented rigorously. The
core damage frequency has kind of come into use more in a use basis than
an actual requirement basis. And there are those who would even argue
that the safety-goal idea is not particularly logical because it is a
single-attribute concept. It is a speed-limit mentality, and we all
know that everybody that stays within the speed limit is not necessarily
safe, and everybody that exceeds the speed limit is not necessarily
unsafe, that maybe what we should be adopting here is more of a decision
framework, that is to say, the choice between alternatives.
And so I think all I'm saying by this with respect to my
second point is that we ought to be darn careful here to make sure that
indeed we do want a safety goal, because the experience with the safety
goal in some people's minds has not been all that reassuring, primarily
because it hasn't really been implemented.
MR. PAPERIELLO: That may be the case. I wouldn't want
to -- I don't have enough experience on that side of thinking about it.
When I was on the reactor side I dealt with the implementation of the
program in the field and not the more philosophical or regulatory
interpinnings here in headquarters. I do know in the case of what I
regulate I am looking for some way of bounding what we do. You just
can't keep, every time something goes awry, adding another requirement.
DR. GARRICK: That's right.
MR. PAPERIELLO: For one thing, somebody can't keep track of
all the rules in the rule book. That's part of my problem. I mean, one
of the things we're doing when we're revising all of our licensing
guidance is to keep things simple. Because I deal with a lot of
entities, you know, gauges, what five things do I need to know about my
gauge? You know, don't put your hands on the source, you know, keep it
locked up when you're not using it, if it gets damaged, don't abandon
it. I forget what the other two things are, but you can do it very,
very -- keep it very, very simple.
I think in some cases for a lot of my licensees if I could
keep the rules on one of these cheat sheets. You know, I used to play
bridge years ago, and you get a plastic bidding convention that was on a
sheet of plastic and read both sides. If you keep the rules on
something like that, for many of my licensees that's about as much as
they can remember or refer to. And it's better to have a handful of
reliable rules rather than a whole bunch of extra little things that
people will never follow because they're going to forget it.
Again, it has to be gauged to the sophistication of the
DR. GARRICK: George.
DR. HORNBERGER: Is it important to have the Agreement
States on board as you do this, and if so, what are you doing to --
MR. PAPERIELLO: Well, it would be nice to have the
Agreement States on board, and of course what we do when we write -- for
the last several years, at least since I've been in charge, every time
we have a rule revision, when we revise all our licensing guidance, and
in some cases when we revise our inspection procedures, we have working
groups, and we include Agreement State people on the working group. But
the Agreement States are clearly not where we are in terms of
quantitative risk assessment in, you know, in regulation.
DR. APOSTOLAKIS: I assume we'll come back to the issue of
risk assessment when we talk about the framework.
DR. GARRICK: Yes. Yes. I guess we have another
presentation to do. Yes. Yes.
Are there any other questions for Carl? We really do
appreciate your spending the time with us, Carl.
MR. PAPERIELLO: Glad to be here.
DR. GARRICK: Yes. Okay? All right. Thank you.
Seth, I'll let you introduce yourself and your topic.
MR. COPLAN: Good morning. I am Seth Coplan. I am in the
Division of Waste Management in the Office of Nuclear Materials, Safety
and Safeguards, and I am the contact on the paper that is kind of the
main part of the discussion here this morning.
What I would like to talk about is, first of all, some
background, how the paper came to be written, what its purpose is, then
I would like to spend some time taking about risk assessment, starting
with risk assessment methods that we have used in the Office of Nuclear
Materials, Safety and Safeguards and what associations we believe can be
made overall within the office between such methods and the things we
regulate. Then I would like to turn to the topic of risk management and
with all of that in place, we can talk about the framework and the need
for stakeholder involvement as we implement it and finally how this
links with the PRA Implementation Plan.
Obviously there is a lot of material there and I am going to
try to move along but will be happy to answer any questions as they come
Basically the paper had its origins in the strategic
planning effort that the Commission went through a couple of years ago.
One of the issues that the Staff identified, that it looked to the
Commission for policy guidance on, was something called risk-informed
and performance-based regulation.
In 1997 on April 15th the Commission issued a Staff
Requirements Memorandum, an SRM, that gave direction on this broad
topic. Most of it was directed agency-wide. A very substantial portion
of what was not agency-wide was directed towards reactors, but there was
one paragraph in particular that was focused in the materials area, and
there were basically three things that it brought up.
First of all, it asked us to review the risk-informed and
performance-based approaches that we had in the high level waste area
and in the industrial and nuclear medical areas to assure that the needs
of those licensees were receiving adequate consideration.
Secondly, it asked that we review the bases for our
materials regulations to identify areas that could be made amenable to
either risk-informed or performance-based regulation with minimal
Finally, we were asked to develop a framework for using risk
technology in materials regulation that would be similar to the
framework that was developed for reactors and provided to the Commission
in I think it was SECY 95-280.
We provided a preliminary response to the SRM in SECY
98-138. The purpose of this paper was to address several commitments
that we made in SECY 98-138, which I will get into in a minute, and to
request that the Commission approve, first of all, a proposed framework,
secondly, our approach for addressing risk management issues, and in
particular the approach that we were going to take or were planning to
take to develop risk metrics and goals, and finally Commission approval
of the formation of this subcommittee, because we thought we needed to
have some means of getting good technical peer review on this.
In SECY 98-138, we committed to develop or to scope the
development of a framework, and estimate the resources that it would
take to do it, make recommendations to the Commission on how we were
going to proceed from there, and I probably ought to explain one thing
there. In SECY 98-138, the proposal that we had for developing a
framework was really going to be a pretty drawn-out, cumbersome process
that would have brought the Agreement States in and lots of stakeholder
interaction just to develop a framework.
When we got together as a task group, which was part of what
we said we were going to do, it came out pretty quickly we were
over-cooking things and that the reactor framework was really a fairly
high level, logical structure that was just intended to make sure that
all of the right bases got touched and that even with the more
complicated situations that we have in terms of the variety of
activities it would be possible to come up with a fairly simple
framework analogous to the reactor framework, and so we ended up doing
We also committed to making an association of risk
assessment methods with the regulated uses of materials and disposal,
and finally, to make some preliminary identification of how we might use
risk-informed approaches in materials regulation.
We also committed to re-examining what activities we had
listed in the materials area in the PRA Implementation Plan from the
standpoint of are these consistent with the framework with the approach
that we are planning.
I would like to turn now to risk assessment in NMSS. I
think that in the talk that Carl gave it started to come out that we
have been involved in risk assessment in the materials area for a pretty
long time actually. We started developing performance assessment back
in the middle to late 1970s in part because a risk-based or risk type of
approach seemed to be the only reasonable way to account for the kinds
of upset conditions that might affect a repository over the long period
of interest that was involved, and so we got involved in a number of
activities within the United States and international activities that
led to the development of the performance assessment methods that are in
Carl described earlier how we got into applying risk methods
to transportation. Some time later we got into a concern about
accidents at some of the fuel cycle facilities that were kind of
unanticipated and again look to system analytic approaches that might
help to ferret out things like that before they happen and we became
interested in integrated safety analysis.
My point is that over a period of time, we picked up
techniques or developed them as they seemed to be applicable and needed
and when we started to look more carefully at what we do in the office
in connection with this paper, it struck us that really we deal in one
sense with four types of activities.
We have got activities that involve a long-term presence of
nuclear material at a site at some planned level. This is high level
waste disposal, low level waste disposal, and decommissioned facilities
where some contamination is left at the site so that people would have
exposures that are within the Part 20 limit and no more than that.
The second broad area are activities that involve casks to
keep material from exposing people, and the casks then need to be
protected. We use them for shipping, use them for storage of spent
The third type of activity is physical and chemical
processing of materials. It includes things like enrichment, fuel
fabrication. We think of the preclosure activities at a high level
waste repository as falling into that category.
Finally, there is just an almost remarkable variety of
activities that involve the use of byproduct material in industry and
medicine, and what we found is, not surprisingly, performance assessment
is a good tool for the first group. Should be -- it was developed for
one of the principal areas in that group.
Secondly, we found that PRA and ISA both seemed to work
pretty well for transportation and spent fuel storage analyses.
Thirdly, that where we're dealing with chemical and physical
processing of material we're dealing with processes that are really in
many ways similar to what goes on in the chemical industry and it made
sense to pick up techniques that were being developed in that area and
that is essentially what we did and what we are working with.
The area that was problematic is the one that involves this
wide variety of things that involved sealed sources, loose material,
just a whole variety of things, and there we had a contractor study that
has been working and developing something that we have come to call
hazard/barrier analysis and we think that that might provide a pretty
good tool for doing risk assessment in this industrial and medical area.
DR. APOSTOLAKIS: Can we get a copy of that? i don't have a
copy of that.
MR. COPLAN: You don't have a copy of it? Okay --
DR. APOSTOLAKIS: And also are there any reports or papers
where we can go and learn more about ISA?
MR. COPLAN: Yes. We will get that for you too.
DR. GARRICK: I assume we are going to come back to this
because this is --
DR. APOSTOLAKIS: Not today, I don't think.
DR. KRESS: That request is not just for George. It's for
MR. COPLAN: Understood.
DR. GARRICK: Seth, just to pick up on something that Carl
said, from a curie standpoint, aren't the first three the real issue as
far as risk is concerned and curies are, is that so?
MR. COPLAN: I think generally yes. The only reason that I
hesitated at all on that is that some of the sealed sources have a fair
amount of activity in them and they get lost from time to time and there
have been in some parts of the world some pretty nasty accidents that
involved material from sealed sources getting dispersed, but generally
speaking it is the first three that have the real concentrations of
DR. PAPERIELLO: I would beg to disagree and I think if you
ask Commissioner Dicus, because I know she made the remark, she believes
the last group is the more -- or has the greater risk. You are talking
to distinguish between probability and consequence.
DR. GARRICK: I was talking about curies.
DR. PAPERIELLO: Oh -- I'll admit. No, the source terms in
the top ones are much higher --
DR. GARRICK: Right.
DR. PAPERIELLO: But if one actually takes a look at people
being really injured or killed, it comes from radiographic sources, it
comes from -- and people have been injured in misuse of medical sources,
so, you know, depending on how you define the numbers, if you turn
around and take a look at person-rem, I suspect your last group
dominates, except perhaps maybe occupational workers is dominated by the
last group because there is no barriers physically much between the
public and the source.
DR. GARRICK: I think that one of the reasons for mentioning
it is that to the extent that we can have pointed out to the committee
the differences, the fundamental differences between the materials and
the reactors, it is helpful.
Another area that strikes me where the differences are
rather profound is that when I think of materials I think of the real
risk of being operational. When I think of reactors, I think of the
risk of being upset conditions because the operational is much tighter
controlled and much -- in many respects more manageable than the
materials risk, so there is an interesting difference right there is
that we are very focused in the high level waste on the long-term
disposal, but on the other hand, there's a considerable amount of
evidence that says the real risk in materials is not in disposal but
rather in operational, and it is one of the reasons why there has been
some challenges by the ACNW as to why there isn't more attention given
to the operational phase of, say, Yucca Mountain -- but that is another
difference that is very, very distinctive with respect to materials over
MR. COPLAN: Yes. It is fundamental. I think one of the
slides that Carl used focused on a number of the differences and the
degree of hazard is one of them -- just the curies associated -- and
another important one is how material can get dispersed. With the
reactor you have an energy source built in to do the dispersal, and in
most instances with materials you have to depend on some outside agent
of some kind doing whatever dispersal you are going to get if it is
going to affect the public.
DR. GARRICK: Yes.
MR. COPLAN: And I am sure there are exceptions to that.
There always are when one makes generalizations about materials
DR. GARRICK: Of course.
MR. COPLAN: The PRA policy statement provides general
guidance on what regulatory use should be made of risk assessment.
Basically it says that it should be applied to reduce regulatory burden
where possible and to find gaps where there may be such gaps in
regulation and there are two principal considerations that I think one
goes through in trying to decide how are you actually going to use risk
Basically they are what is the Staff going to do, what will
it do with risk insights and risk assessment in developing regulations
and guidance, licensing, inspection, assessment and enforcement.
An example might help. In Part 63 we decided that we would
put a risk criterion front and center as the basis for licensing a high
level waste repository and so you have a performance objective that
anticipates that the Department of Energy is going to do a performance
assessment, that the Staff is in some fashion going to review that
performance assessment, and it is all embedded in the regulation.
On the other hand, the approach could be something that
might be more applicable in some of the other materials areas where we
wouldn't expect a licensee to have the capability of doing risk
assessments and instead the Staff would do some assessments for the
purpose of developing guidance and developing what may be fairly
prescriptive criteria that the licensee would be expected to follow but
that would be more risk-informed than just some of the judgmental bases
that are in place now.
So in any event, the point is that those are choices that
have to be made when one starts thinking about risk-informing
The complement of that is what do you expect the licensee to
be doing? Do you expect them to be actually doing calculations as part
of their operation or do you expect that they are going to implement
criteria that are risk-informed because of what the Staff did?
There are a number of factors that are important in making
those decisions. They include the hazard and complexity of the
activity, the degree of human involvement in the activity, the technical
sophistication of the licensee community, NRC staffing and training
issues, Agreement State issues and others.
The consideration of these factors in the -- across the full
spectrum of activities that we regulate is something that is going to
have to have a lot of stakeholder involvement so that we are really sure
that we are understanding the right issues and addressing them and we
would expect that we are going to end up with a pretty broad spectrum of
regulatory approaches coming into the regulations at the end.
DR. GARRICK: Seth, aren't you already contradicting your
approach? It seemed to me I heard you say at the outset that the task
group in trying to formulate recommendations and what have you was
forced in the interest of getting the job done and simplicity to -- what
I read -- minimize stakeholder participation?
MR. COPLAN: No. No, let me clarify. It is the difference
between development of the framework and implementation. In a sense
what we did is we moved what I think is really essential stakeholder
involvement from a development phase of the framework to an
implementation phase, and there I think it is unavoidable.
DR. GARRICK: But isn't that one of the concerns and
arguments that a lot of the stakeholders have is that they are brought
into the process later rather than early?
MR. COPLAN: Again I have to agree. I don't think that we
are really running afoul with that because the framework itself I don't
think is where the stakeholders would find that if there is an ox that's
going to be gored that that is where it would happen.
In fact, we did attempt to have an Agreement State
representative on the task group and when we discussed the project with
the organization of Agreement States they in effect said we are covered
indirectly through another project that we are involved with and when
you get to the implementation, come back -- then we will have a pretty
The paper has a tabulation of what are really very
preliminary thoughts that the Staff has had about where all of this
might lead in regulatory terms. By the time we get through working with
stakeholders and so on, I would be surprised if it ended up that way.
It will probably look quite different but it is kind of a starting
Risk management, I guess the way that I think about a lot of
this is that risk assessment gives you tools for being able to ascertain
what the risk may be that's associated with some activity. It doesn't
tell you anything about how you should react to that risk as a
regulatory agency. And that side of things is risk management. It
involves deciding what risks are you going to look at, what metrics are
you going to use, what individuals or populations are you interested in.
Are you concerned about property or aren't you concerned about property,
and so on. And then once you've looked at what risks you want to
measure, then comes the issue of what targets or goals do you want to
In the nuclear area I think that it makes sense to look at
things in terms of -- I guess I have sort of a mental matrix that
involves workers and public and normal operations and accidents, normal
exposures and accidents, so that you really, when you start thinking
about managing risks, you have to start thinking about managing worker
risk during normal operations, worker risk during accidents, public risk
during normal operations, and public risk during accidents. And for
both the reactor and the materials area, there are metrics and goals
that apply to normal operations. They stem from international groups
like ICRP, there is Federal radiation protection guidance, and so on.
And for the Commission's purpose, this is all embodied in Part 20.
For upset and accident conditions at power reactors, there
are the qualitative safety goals and the quantitative health objectives,
and, as Carl pointed out, they don't really quite fit the materials
area, for a number of reasons, unfortunately. It would be nice to
really be able to pick them up. And we're figuring though that we are
going to have to, you know, somehow come to grips with targets, and in a
way that takes into account the very substantial differences between all
the different kinds of activities that we regulate. And it's something
that I think we're going to have to work very carefully and completely
with stakeholders on.
Some of the things that are pertinent to developing accident
goals for materials. First of all, with reactors, as Carl pointed out,
deciding on what the target groups were is fairly straightforward. The
reactors stay in one place, and you have a population around the
reactors, and depending on what happens at the reactor, the population
may or may not be affected to some degree.
For materials it gets more complicated. You have some fixed
facilities, a number of them, some of which do have enough energy
involved in the processes that go on in the facility to disperse
material, some not. You've got a lot of things, though, that take a
fair number of curies and move them around from one place to another.
Radiography is an example of that, a radiography source has a lot of
curies in it, and it goes from one place to another, and then we have
the problem of gauges that get lost and end up in places that they, you
know, you don't expect that they should be, but the point is they're
moving around through a population. And deciding how, you know, how
to -- who the target is is a problem that we're going to have to
confront that is a little bit different than the reactor situation.
The kind of -- the way the risk works is different, too.
With reactors, you're kind of in a situation where the real concern is
low-probability, high-consequence events. Our situation is that most of
the risk involves relatively low-consequence events relative to the
reactor situation but with much higher probability. They happen.
We, you know, we have a materials event data base, as Carl
pointed out, that involves fatalities and that sort of thing. And so
the fact that we actually see kind of a continuous stream of events
because of the large number of licensees that we have and the fact that
you have these relatively high-probability, relatively low-consequence
things going on, you see things happening, which is unlike the reactor
situation and is again something that I think would have to be addressed
in developing a safety goal.
And then in addition to those factors there are some purely
institutional considerations, namely that there are an awful lot of
people who have an interest in standard-setting where materials are
concerned. EPA is heavily involved, ICRP, NCRP, and they have different
levels of authority, but all are involved, and all of them have to be
dealt with in however we would proceed in that area.
And then, at the risk of belaboring a point, there's just
this tremendous difference in the capability of licensees to address
risk in their activities, in the -- and the value that society gets from
what they do. And somehow that has to be accounted for.
DR. GARRICK: But that fact cannot remove the
MR. COPLAN: That's right. That's right. But, you know, I
think it's something that you have to address at least at a
philosophical level. In a sense the reactor safety goal recognizes that
you don't want to have I guess it's a societal risk associated with the
generation of nuclear power that's in excess of the risks associated
with generating electricity by other viable means, so that there's a
recognition that the risk doesn't come in a vacuum, it's because you're
trying to get something, and it's part of what you pay for what you're
The framework that we came up with is fairly similar to the
reactor framework. It has four parts, basically.
The first part is simply identifying all of the areas in
NMSS where risk-informed regulation is a possibility. A little bit
different than the first part of the reactor framework, in that the
reactor framework deals with only one activity, reactors. We've got to
deal with fuel-cycle facilities, with geologic repositories, and so on,
and each one of these has the same collection of activities that -- or
regulatory areas that you might have with a power reactor. And so the
idea is to be able to look at these and say is this something where
risk-informed regulation can bring a benefit or not?
There are so many of them that we have no hope of kind of
working through in a kind of a perfectly logical fashion from start to
finish we're going to be somewhat resource-limited in what we do, and so
we're going to work through these in a prioritized sort of way, and one
of the things that we'll have to do early on is establish how we're
going to do the prioritizing, one thing has to be external drivers. For
example, we've been driven to a risk-informed approach in high-level
waste regulation by certain external factors and by some factors that
are just kind of intrinsic to the repository problem.
The second part is having identified areas where the staff
thinks that there's promise, that we would work with stakeholders to
identify metrics and to -- I'm sorry, I jumped ahead here a bit.
The second part of the framework, I started to think about
the implementation. The second part of the framework is to make sure
that we understand what underlies the current approach. What are the
deterministic considerations that are there, what sort of institutional
concerns, and so on, make it what it is now.
Thirdly, what are the elements that risk considerations
could improve, and what could come out of that.
And finally, how can the things that are there now that we
like be integrated with the risk aspects?
And to implement, we came up with a five-step process, the
first of which I described in talking about the first part of the
The second step is one of deciding how to modify the current
regulations, with this understanding of what it is you want to keep
about it, what is it that you can gain from risk. We would start
thinking about metrics and goals and the issue of what is it that the
Staff would do, what is it that the Licensee would do, and bring in
stakeholders pretty heavily at that stage.
The third step would be to make the appropriate changes to
the regs and guidance.
The fourth step would be to make sure that the Staff is
knowledgeable in the new approaches to do whatever training is needed to
And then the fifth step is developing and adapting the tools
needed to make these things go.
The -- I think that I have really covered most of the points
on this slide. Basically let me just say that we see this as just an
extremely important part of making this work. It would --
DR. APOSTOLAKIS: You know, the issue of stakeholder
involvement in risk management is something that is of great interest
these days, and the National Academy of Sciences and Engineering have
published -- issued reports through the National Research Council, the
other NRC, where they outlined approaches to using or to bringing
stakeholders into the process and so on. I wonder whether something,
especially in the analytic deliberative process that is now something
that people are trying to recommend, I wonder whether these things would
work here. One of the principles is that you should bring the
stakeholders into the process very early, and have them involved also in
the risk assessment process. You are not really assessing any risks
here, you are planning to revise the regulatory system?
MR. COPLAN: That's right.
DR. APOSTOLAKIS: But, even so, it seems to me that bringing
them in early in the process would help, and I assume you are familiar
more or less with that literature.
MR. COPLAN: Yes.
DR. APOSTOLAKIS: By the way, is this committee a
MR. COPLAN: Yes.
DR. APOSTOLAKIS: So you would involve us --
MR. COPLAN: After a fashion.
DR. APOSTOLAKIS: -- you would involve us early in the
MR. COPLAN: You couldn't get much earlier than this.
DR. APOSTOLAKIS: In the future I would really appreciate it
not being presented with a SECY when it's final. When you plan to do
something, it would be a good idea to come before us and get some ideas,
because this is a new area, I think, for most of us, and reviewing a
finished product can only lead to trouble. And, you know, the committee
will make sure that it reacts to your proposals appropriately, that they
are not final and, you know, trying to develop something, if my
colleagues agree, because reviewing the final product sometimes is a
risky proposition. This is the fifth area of risk, you know, in your --
MR. COPLAN: I am hoping that this will be unique.
DR. APOSTOLAKIS: Okay.
MR. COPLAN: In this. I think that the model that we have
in mind for the interaction with the subcommittee is one that would be
similar to the interaction that the Staff had with the ACRS over the
development of the risk-informed guidance.
DR. APOSTOLAKIS: You are very well informed.
MR. COPLAN: That's what we're -- that's our goal here.
DR. APOSTOLAKIS: So we are in agreement.
MR. COPLAN: There is not a lot to say about the next chart
here, just that we had committed in SECY 98-138 to look at the PRA
implementation plan, and to look at the activities we have listed in it,
decide whether we thought they remained appropriate. Given the
framework, we did that, we think they are. The other point is that we
are planning to come up with a -- well, as we lay out plans for
implementing the framework, they will be incorporated into the PRA
implementation plan, and that will become the way of tracking our
Just in conclusion, we have used risk assessment in the
materials area for a long time, but the degree of development of
applicable methods and experience in using them has varied from one area
to another, and I think consistent with really the importance of
bringing risk assessment into play in those areas, we have proposed a
framework to the Commission for expanding the use of risk-informed
approaches in the materials area. We would see implementation as
involving a pretty significant effort. We think that stakeholder
involvement is extremely important, and we will value input from the
subcommittee for technical peer review.
DR. GARRICK: Thank you. We've already passed our break
time, but I think that while Seth is up front that maybe we should take
a little more time and ask the questions now, and then take a break. So
going in a little different order, George?
DR. HORNBERGER: Seth, one of the concerns, of course, about
risk-informed performance-based regulation is that licensees could see
it as regulation on top of existing regulations and, of course, that's a
concern that we all are sensitive to and want to avoid. It's pretty
clear in some of the areas that you mentioned -- for example, high level
waste disposal performance assessment as well, it's well known to be the
method of choice. If we look at something like sealed sources, do you
have confidence that the framework that you have outlined is not going
to add a burden in any way to the licensees?
MR. COPLAN: If we do our job right, it shouldn't. And I
guess the reason that I would say that is that starting from the general
guidance in the PRA policy statement, what we'd be looking to do is to
reduce regulatory burden where possible by using risk methods, and where
using them reveals, you know, some sort of a gap in the regulations, we
might make a change there. So, I mean, to begin with, the basic idea is
one of using it to reduce regulatory burden.
Secondly, we recognize that the -- that that licensee
community is one that doesn't -- it doesn't make sense to burden them
with doing risk assessments; that for them, the kind of regulatory
framework that exists now, that's fairly prescriptive, that has fairly
prescriptive guidance, is probably the right one, but it may need to be
reexamined from the standpoint of are we focused on the right areas.
DR. GARRICK: George?
DR. APOSTOLAKIS: Well, where do we go from here? I mean
DR. GARRICK: Well, if you have any questions --
DR. APOSTOLAKIS: No, I don't have any questions. The Staff
is not asking for a letter or anything at this time. Are you?
MR. COPLAN: No.
DR. GARRICK: Tom?
DR. KRESS: No.
DR. GARRICK: One of the things that as a practitioner in
the risk field, and therefore a little bit short-sighted on some of
these global issues, that strikes me is that when you talk about the
implementation of the framework, I am a great believer in it, in the
top-down thought process, and you say that the first step is to identify
the specific regulatory applications that are amenable to expanded use
of risk, et cetera, et cetera. As a practitioner, of course, I think
the first step ought to be based on our vast knowledge and experience,
what does NMSS think the real risk issues are in materials, one through
10? And, of course, this has a process implementation, and I'm asking
the question, well, you've got a lot of experience, and you have been,
as you've said, doing risk work for a long time and safety work for a
long time. I like the categorization into the four, I'd make it five,
and I'll tell you why in a minute.
But one thing that would really, I think, help this
committee would be for us to have the full benefit of your knowledge of
what the risk issues are. Where is the problem here? We have
identified some fundamental differences, the upset versus the
operational, and the populations that are affected, and the variety of
other things. But I would guess that there are some opportunities here
to narrow this problem down, whether it's done on a curie basis and, as
Carl has pointed out, that's different from a risk basis, isn't so much
important as us knowing what those things are. Because as I look at
your categories, I look at them -- I looked at one category that would
be basically the disposal activity. That's where you get rid of it for
all time. And the other category that I'd put as maybe the next one
would be the process and manufacturing facilities, because that's a
category where we can take maximum advantage of the reactor experience
And then the third category that I would think of, and it
lines up with yours except you have joined it with transportation, I
would have storage and then I would have transportation as a separate
because transportation is a bogeyman in the risk field.
There is this tendency to think that you can't move
hazardous materials. I am reminded of the chemical weapons program
where a very preliminary transportation risk analysis resulted in about
a $15 billion decision to build processing plants at each of the storage
sites rather than either a central processing plant or a regional site
because of the transportation risk.
I think transportation is another area that is very amenable
to the kinds of analyses and methods that have been developed and tuned
and that you have used, and so I would consider it separate.
The other thing I wanted to comment on the categorization is
that -- and I will be struck by lightning on this one -- because I have
always said that the biggest lesson that we have learned in reactor risk
is that risk is plant-specific and therefore you have never known of a
generic plant to fail or have an accident, but on the other hand when
you talk about something like sealed sources, radioactive sources, this
strikes me as something that really lends itself to a very good generic
analysis of asking the questions of what can go wrong with the sources
and what are the consequences and what have you, so that strikes me as
very much an exception to the approaches of getting consumed in the fine
structure of individual seals and individual sources. Now there may be
categories of sources you would want to ferret out.
But back to the question of the five step process. Is there
a way that you could get in front of this committee your collective and
expert judgment as to what you consider to be the top 10 risk issues
associated with materials handling?
MR. COPLAN: Oh, I suspect we could, yes.
DR. GARRICK: Wouldn't that be helpful, really helpful to
us, to know from the real experts what the judgment is right now from a
MR. COPLAN: Yes, I suspect we could do that.
DR. KRESS: Wasn't there a Scientech report on the risks
associated with it?
DR. GARRICK: There was something, yes.
DR. KRESS: That would help, to get that. I have read it in
here somewhere but I couldn't find it in here. It exists. It is in the
MR. COPLAN: If I am not mistaken, which I could easily be,
but that sounds like it may be the report that deals with this
barrier/hazard analysis or --
DR. APOSTOLAKIS: Yes.
MR. COPLAN: Or with diamond trees and all that stuff.
DR. APOSTOLAKIS: It is the same one?
DR. KRESS: No. No, I don't think that was the one they had
MR. COPLAN: Okay.
DR. APOSTOLAKIS: Well, maybe it is not Scientech. They
just say a contractor.
Who was your contractor? Why is he a secret? In 98-138 you
refer to a contractor who developed the report?
MR. COPLAN: It was Scientech.
DR. APOSTOLAKIS: Scientech. It is not the same report
MR. COPLAN: I think it is, yes. I am getting nods from the
back of the room that it is.
DR. APOSTOLAKIS: But that was for reactors.
DR. KRESS: Yes.
MR. COPLAN: Oh, oh, oh, oh -- there were two that were
done. There was one done for reactors and one done for us, very similar
in terms of the methodology.
DR. APOSTOLAKIS: So it was a diamond tree again?
MR. COPLAN: Yes, yes.
DR. APOSTOLAKIS: Wasn't there an ACRS letter on that? Have
you read that letter?
MR. COPLAN: I have not.
DR. APOSTOLAKIS: Yes. It is a recent letter.
MR. COPLAN: Okay. We will have to get it.
DR. GARRICK: Any other comments, questions for Seth from
either the subcommittee, the other committee members or the Staff?
DR. GARRICK: Okay. Well, thank you very much.
MR. COPLAN: Thank you.
DR. GARRICK: And I think with the permission and the
concurrence of the co-chairman, we will take a break for 15 minutes.
DR. GARRICK: I think there are a couple of things we want
to cover in the remaining time, which is about an hour and fifteen
Number one is based on what we have heard today, what does
the subcommittee think is the direction we ought to be taking? We
through the questions that we asked implied some directions. It sounds
like that on the surface at least there is additional information that
the subcommittee wants to see, that maybe a letter is an appropriate.
We'll get some more guidance on that issue from the Staff I am told, but
for the time being at least it seems that there's these other documents
that we would like to see and maybe even hear about, although I think it
would be appropriate based on what we have heard for us to discuss the
The second thing we want to see if we can resolve is, as you
know, the subcommittee process is a work in progress in terms of how it
is going to work and the supporting documentation for it, and we need to
talk about that a little bit and see if we can come to grips with any
remaining issues for how the subcommittee will function.
So I think with that I would like to hear from the
subcommittee or the members, their views on how they think on the basis
of what we heard we should proceed. I am open to suggestions as to who
would like to comment first -- maybe Tom, you would like to comment on
DR. KRESS: Yes, I will. First off, I think the issue is
one of risk. We're controlling risk and I think we need to recognize
the different methods that have been suggested like PA and ISA, PRA are
all risk assessments by different names and that we are interested in
what I guess you would call the risk triad or the triplet.
DR. GARRICK: Triplet.
DR. KRESS: Triplet, okay, and that somehow in all the
activities we have to address that triplet and regardless of how the
process is whether it is formal one, whether it is qualitative or
quantitative and whether it adds up all the sequences or just the major
ones or not, we -- I think somehow we need to get that principle in
The other one is I think we do need to develop some high
level principles as a committee to guide our view of how to review this
stuff and some of those high level principles have to deal with
acceptance criteria and risk acceptance criteria as well as the
uncertainties associated with those.
My own feeling is that we have to have a common metric on
risk acceptance criteria and I don't know what it is right now, but I
suspect somehow it has to do with the risk benefit is a common metric
that ties them all together, and the uncertainties have to be dealt with
some way by defense-in-depth, so I think we have a problem on how to
wrestle with the defense-in-depth in this arena because it may mean
something besides barriers when we talk about the wide range of things
that we are dealing with.
I think among ourselves we have to figure out what our
principles are on addressing this and those are some of them that I
think we have to deal with.
George, did you have anything to add to that?
DR. APOSTOLAKIS: Well, who is going to do this, Tom?
DR. KRESS: That's the --
DR. APOSTOLAKIS: Not the committee?
DR. KRESS: Well, I think the committee ought to -- you
know, I don't think we can sit by and wait for somebody to come and
bring a set of principles to us. We need to think about high level
principles. We don't have to quantify them but what are the principles
associated with this?
DR. APOSTOLAKIS: Or we can write a letter recommending that
the Staff do this, because I think it is going to take some considerable
effort to think about this and come up with proposals, and we can
certainly work with the Staff but to go through another exercise where
we have to do it, I mean we are just finishing the defense-in-depth
letter and you know how much trouble that was.
DR. GARRICK: Just to help me understand, Tom and George,
when you talk about principles are you talking about something along the
lines of what was generated through the ACRS interaction with the NRR
Staff on the set of principles?
DR. APOSTOLAKIS: 1.174, yes.
DR. GARRICK: Yes, guiding the implementation of --
DR. APOSTOLAKIS: Yes -- high level principles, which is
related really to your top-down comment earlier.
DR. GARRICK: Right.
DR. APOSTOLAKIS: I mean at the very top what is it that we
want to do and that will give us also an opportunity to say a few words
about language and establish, you know, the common framework.
DR. KRESS: And I think we should have some sort of a graded
approach to regulations in this area particularly. I think there are
some beyond regulatory concern issues here.
DR. GARRICK: Yes.
DR. KRESS: And, you know, it appalls me that we are
regulating at the very low millirem level in this area. I think we need
to think about where are or what are the real issues in terms of health
effects and risk associated with that but that is treading on land
DR. HORNBERGER: Quicksand.
DR. KRESS: Yes, quicksand. That is the word I was trying
to get in.
DR. GARRICK: George, do you -- based on what we heard, what
is your thought on where we should go?
DR. HORNBERGER: Well, I mean I think that the only other
thing in addition to what has already been talked about, and perhaps
that is part of what we have talked about, is these issues of how the
licensee uses risks and exactly how the NRC uses -- makes the connection
between the risk assessment and a regulation. Perhaps that is really
just in the risk acceptance criteria. Maybe that is just another way of
DR. KRESS: Yes. I meant to mention that also. I think
it's probably not practical to ask every licensee to make his own risk
assessment in this area.
DR. HORNBERGER: Right.
DR. GARRICK: Right.
DR. KRESS: Somehow the NRC has to do a lot of that and
reflect that in the acceptance criteria that they come up with, which
very well may be very prescriptive.
DR. GARRICK: Yes. The materials field doesn't have the
great focuser that the reactor field has in terms of everything centers
around avoiding core damage. There is not a counterpart here on the
materials side that aids the whole process of organizing the issue in
the same way, and so it seems as though they kind of got a start on what
may be a direction that we will have to go of maybe four or five
different categories of issues that have to be addressed.
And then there is one, the only one that sits apart from all
the others perhaps is the Yucca Mountain issue. That is clearly a
singularity in the whole arena of materials risk. But the rest of it
needs some real organization to be able to provide focus to what the
risk are and understanding the problem. I think that that kind of
information has to be before the staff before anybody can be thinking
about the concept of safety goals and what-have-you.
I think one of the things that this committee probably feels
a little different about than maybe what we heard is that -- what the
world of risk assessment is all about. I think of us see the four
categories of activities that they have as really being a variation of
scope of the application of risk assessment more than different methods
being applied. You still ask the same questions. You only deal with
the answers to the extent that is necessary within that category and no
more. So I think there is an opportunity there maybe for some advice
down the road. But I do believe that we need to have the benefit of
more of their understanding of what the risk issues are before we are in
a position to do that.
DR. APOSTOLAKIS: If we did nothing, I mean we wrote a
letter this time and so on, what are the staff's plans? What do you
plan to do and when will be the next time you will come before us, and
with what? It is not clear to me what -- I mean do you have an SRM now
that you have to respond to and at some point you want to come before
MR. COPLAN: We don't have an SRM yet.
DR. APOSTOLAKIS: But you know there is one coming?
MR. COPLAN: Eventually we will get an SRM. We don't know
what it is going to say.
DR. APOSTOLAKIS: So if we write a letter now, we may
influence the SRM.
MR. COPLAN: That is true. That's true.
DR. APOSTOLAKIS: Okay. So what are we going to do? I mean
when are we meeting again, and what will you present?
MR. COPLAN: I think we were figuring that we wanted to see
an SRM before we got real deep into, you know, into planning exactly
where we were going to go next. I mean that is part of the reason that
we didn't get to this subcommittee until the paper was written actually.
I think that one of the first things that we would
anticipate doing is identifying what we would put into the PRA
implementation plan. And it just seems to me that the thinking that
goes into that might be something to discuss with the subcommittee.
DR. APOSTOLAKIS: So, well, it seems to me then that if we
want to influence things, we ought to write a letter stating what we
DR. HORNBERGER: Should be done.
DR. APOSTOLAKIS: Should be done. And recommending to the
Commission a certain course of action and leave it up to them whether
they want to issue SRMs or not. Because the process is til in the very
early stages and we don't know which way we are going to go, right. So
maybe a letter is appropriate.
DR. GARRICK: Yes, I was thrown off a little bit on the
letter issue by the question that was raised earlier by Tom when you
were making your presentation. But it seems that if we do want to
impact, and there is value added in our impacting the SRM, that a letter
with some preliminary thoughts and ideas on how to proceed might be
Now, if we were to write such a letter, Tom, is it your
thought that we should attempt a stab at some basic principles?
DR. KRESS: Yes, that would be my thought, but George seems
to think that is more of a task than we are up to at the moment.
DR. APOSTOLAKIS: Well, Tom, if you can produce something, I
would be happy to read it. But I know how hard it is for part-time
advisors to put together something that is archival.
DR. KRESS: That's tue.
DR. APOSTOLAKIS: And, you know, the defense-in-depth
letter, we are writing a letter on defense-in-depth on the ACRS and Tom
had the lead on that.
DR. KRESS: It is very painful.
DR. APOSTOLAKIS: It is extremely painful. So my
inclination would be state our concerns and what we think ought to be
done at a high level, and then recommend -- and maybe give some
preliminary thoughts, I mean we can do that, especially if we give
alternatives where you really don't have to vote then that you like this
one versus the other one, and then recommend to the Commission that they
ask the staff to actually spend some serious time on those. And that
will help everybody, I think, you know.
Like, for example, one could be the principles. Tie it to
the top-down approach. Another one would be the language, clear up the
language and make it consistent with the white paper of the Commission.
Another one, I would like to see some words about other aspects of risk
like controllability and so on, that influence the decisions. The
necessity for goals. Maybe the Commission ought to think seriously
about it, you know, that they should ask the staff to think about
establishing goals before they do anything else.
DR. HORNBERGER: Goals?
DR. APOSTOLAKIS: Goals, safety goal.
DR. GARRICK: Well, actually, it seems to me that one of the
things we could do that would be rather specific, because it embraces
the prior work to some extent, is for the committee to comment on
SECY-99-100. That has in it most of the issues that we have been
talking about, including the categorization, including the activities
that are involved. So maybe a first cut comment on SECY-99-100 would be
a constructive initial product.
DR. HORNBERGER: But that would be basically the platform to
promote the ideas.
DR. GARRICK: Yes. Right. Including the idea of the need
for some basic principles.
DR. HORNBERGER: Yes.
DR. GARRICK: Seth, would you think that that as a strategy
would be constructive?
MR. COPLAN: Yes, I think it would. I would just mention
one thing and that is that we haven't really set any resources at all
aside to do this effort for FY '99, we are not really planning to get
started until FY 2000 and then we are figuring on a level of effort that
would be about 8 FTEs per year office-wide, including some things that
are ongoing, so there are going to be some other demands on resources,
DR. LARKINS: Yes. To add to that, I think plan for about
three meetings of the subcommittee per year in the request.
DR. GARRICK: Okay. All right. Well, I think if the
subcommittee is in agreement with this, then we will attempt to draft a
letter on our view on what we heard and particularly on 99-100.
DR. APOSTOLAKIS: Yes, and I would appreciate receiving
those documents we requested as soon as possible, because I would like
to have a chance to look at them before we finish the letter.
DR. GARRICK: Yes. Yes, I think so. Okay.
MR. COPLAN: Do you have any idea what the timing of this
letter might be?
DR. GARRICK: Well, that depends on our next topic.
DR. HORNBERGER: That's right.
DR. GARRICK: If we can resolve a procedure here that
doesn't get totally bogged down in practicing what we preach, namely, of
trying to keep things simple, it shouldn't be too long.
DR. FAIRHURST: Can I just ask a question? I am a little
hesitant here, but being the fool, I will rush in where all the experts
apparently fear to tread. But it is interesting to listen to what Carl
and Seth were saying, that there appears to be an implicit assumption
that you can define the risk. And yet we just came from a workshop last
time where this linear no-threshold hypothesis is highly -- well, a lot
of people think it is extremely uncertain.
And I am wondering if you are going to build a large
superstructure of regulations and codes which are built on possibly an
unsure foundation unless they get too rigorously built in, which it will
take a lot of effort to get them in.
And let's say five, eight years down the road, somebody
comes up with some biological research in the human genome studies and
says, look, you are all wrong. There is a threshold that is 5 rems or
whatever there is. Would it be perceived in the public yes, if one
wants to shift to that as a -- whatever. You would arouse an enormous
amount of misgivings and upsets and so on.
So at the outset, one would need to predicate all of the
results drawn from the study on a strong discussion of the uncertainty
of the foundation. I don't know how much that -- because when you are
talking about, for example, as Carl was, risk benefit, that would change
dramatically depending on the definition of the risk.
DR. GARRICK: Right. If we take the measure of risk as a
dose, for example, we get around part of that problem because we are not
having to work through the dose --
DR. FAIRHURST: That is what I am saying, the experts must
know a way around of comparing A to B, rather than comparing absolutes,
et cetera, et cetera.
DR. KRESS: Yes, I think that might be one of our
principles. Let's deal with dose.
DR. GARRICK: Yes. Right.
DR. KRESS: Get around those problems.
DR. GARRICK: Get around the dose response curve.
DR. KRESS: As long as your acceptance criteria properly
reflect things of risk, why, I think dose is a perfectly good measure.
DR. GARRICK: Right, but --
DR. FAIRHURST: But in the waste isolation, that is going to
DR. GARRICK: Yes. Well, Yucca Mountain is moving in the
direction of a dose standard.
DR. FAIRHURST: I know, the definition of what dose
constitutes a risk.
DR. GARRICK: Yes. But I don't think they are going to
answer that, even --
DR. FAIRHURST: No, I know they are not.
DR. GARRICK: Yes. Right. Any comments from the staff?
DR. LARKINS: Yes. It seems to me one of the things that
might be beneficial, if you could help the staff in terms of
prioritization, in terms of approaches. Now, there's a number of
parameters to consider, risk, the impact on stakeholder, the likelihood
of success in doing something along this line and various methods to do
that. So I think if you are going to early on provide some guidance,
maybe give some thought to how they might want to proritize, or develop
an approach for prioritizing which regulations and things they take on.
DR. GARRICK: Well, you know, my problem with that is I need
to know what the problem is in order to do that.
DR. LARKINS: Yes. Well, I understand. And that goes along
with your request, I think, for those areas where there seems to be the
greatest risk significance in the regulations.
DR. GARRICK: Right.
DR. LARKINS: And that is certainly one parameter, but there
are probably another five or six that might be considered.
DR. APOSTOLAKIS: Now, just a comment here on the
principles. The staff came before the ACRS about a month ago and they
proposed to develop overarching principles of regulation that -- and
goals, right -- that would apply to reactors and NMSS activities and so
on, and the ACRS -- and I assume NMSS participated in that proposal.
MR. COPLAN: Yes.
DR. APOSTOLAKIS: The ACRS was very cool to the idea. They
felt that we can do much more that would be useful in the reactor area
in the next several months and trying to bring everything under the same
But, on the other hand, we did suggest that they spend some
time thinking about these overarching principles and goals. So you may
want -- I mean that certainly should appear in our letter someplace so
that the staff has already thought about it. And I don't know how --
the staff has not responded yet, have they? I have heard through the
grapevine that you are actually implementing the recommendation, but I
haven't seen it in writing. This is just a comment, that the staff has
already thought about some of what we are talking about.
DR. GARRICK: Yes. Right. In other words, there is the
conflicts here, there is -- on the one hand, you would like to be able
to assure the public that there is consistency in the thought process
behind the formulation of regulations be they for materials or reactors.
But on the other hand you've got to get the job done.
DR. APOSTOLAKIS: That's right.
DR. GARRICK: And if you can see a way to solve it for one
group of things and not yet see a way for another one, you shouldn't
DR. APOSTOLAKIS: Exactly. That was the idea, really.
DR. GARRICK: Yes. Right. Right. Yes, okay.
DR. LARKINS: Along that line, there are overarching issues
which come out of some of the things that have been done on the PRA,
which have been done on Part 63, that this group should probably think
about. George and Tom were mentioning the problem that we're having
with defense in depth on the Part 50 stuff. Certainly there's an issue
on Part 63 and some of the other regulations, and if you're going to
have consistency or coherency in the regulations, you might want to
The other thing is John Garrick made a point of not making
the mistake of some of the things that came out of implementation or use
of PRA. I think also you want to take a look at lessons learned from
the use of ISA, hazards analysis, and PA, and fold that into your
So you might want to ask the staff the next time they come
in to talk about lessons learned from the use of these various
methodologies, since they are proposing to use that as various tools or
methods for different regulations. And there's been a lot done by DOE
and Defense Department and others in some of these areas.
DR. GARRICK: Yes, Ray.
DR. WYMER: I'm in the same position that Charles is in of
really not knowing much about this field, but I'll make a comment
The SRM of April 15, 1997, says to develop a framework for
applying PRA to nuclear materials uses similar to the reactor framework.
There's a big difference between that and developing overarching
principles that cover both the reactors and the nuclear materials. You
get down in to substantially more detail addressing what the SRM says
than if you try to do these overarching principles which have to be very
general in nature. What are you really trying to do here, and how
closely are you tied to having to do what the SRM says?
DR. GARRICK: Well, I think what we're trying to do is find
something we can anchor ourselves to.
DR. WYMER: Something you can do.
DR. GARRICK: That we can evolve from, that does deal with
the question of consistency and the kind of things that were discussed
in the strategic initiative document, but we have to be very careful
about that, that's right. That's one of the reasons for the joint
committees is to begin to, since we're both offering advice on these
very broad issues like how to implement risk-informed, performance-based
regulatory practices, we need to begin to think about the overarching,
but not to the extent that it keeps us from getting other things done
where the path is a little clearer. But, yes, there are some tough
problems here, and we obviously don't have the answers.
DR. HORNBERGER: Andy has a comment.
DR. GARRICK: Andy?
DR. CAMPBELL: One of the issues that came out in Carl's
presentation and also Seth's presentation are the incidents that occur
with sealed sources, and what I don't think the Committee has in front
of it is, you know, some sort of statistical analysis of what the
frequency of those incidents are and the consequences. Are they mostly
burns? Are there -- you know, there are fatalities that are reported
But that would be helpful, and I think also one of the
considerations in applying a risk-informed basis would be can a
risk-informed approach reduce that frequency and, you know, what are the
costs and benefits of that.
DR. GARRICK: Yes, I think the main --
DR. CAMPBELL: What is the connection to having a
risk-informed approach as applied to sealed sources in terms of actual
safety which is reducing the incidents that occur?
DR. GARRICK: Right. Right. The main value received from a
risk-informed approach in my opinion is to help us gain perspective on
issues that affect risk, and to help us therefore be in a position to
allocate resources accordingly as to what is most important. And I
think that also we need to be very conscious of these marked differences
between the two problems.
DR. APOSTOLAKIS: But I have a question about that. I mean,
the whole idea of a risk-informed approach, as you just said, is that we
want to focus Agency and licensee resources on what's important. Now in
reactors, there is a clear need for that, because we are regulating
many, many things that are really not important. Is it the same thing
here? Do you feel that you impose unnecessary burden on the licensees
in some areas that, you know, we may change that by following a
DR. GARRICK: We shouldn't, but we may. You know --
DR. APOSTOLAKIS: Have the licensees complained?
Like in reactors, I mean, they have complained.
Oh, they do complain?
DR. HORNBERGER: No, the reactor --
DR. APOSTOLAKIS: Oh, the reactor for sure. Yes. So I'm
trying to understand, I mean, if we switch to a risk-informed approach
here, what's going to change besides the paper? I don't know.
Seth, do you want to comment on that?
MR. COPLAN: I wasn't sure whether you were asking.
I think, first of all, we're not sure, but there may be some
areas where either the risk is worse than we think and the regulations
may need to be tighter. More likely I think we will find that we may be
focusing in some wrong places.
And the other thing is that in terms of, you know, just
looking at the regulations and how they look, chances are they would not
change a great deal as a result of this process. In other words, you
know, turning again to the IMNS licensees, for the most part these are
not licensees that you can expect to do calculations or that you would
want doing calculations.
So you'd have criteria that for all intents and purposes
looked the same, but the emphasis may be different, there might be
requirements that aren't there now, there might be requirements that are
there now that would disappear and that sort of thing, but you'd still
have Parts 34 through 39 looking pretty prescriptive.
DR. APOSTOLAKIS: It's not clear to me that a risk-informed
approach or a performance-based approach would change much. So we seem
to be doing this just to be consistent, I guess, with what we're doing
in other areas.
DR. KRESS: Well, I don't know how many resources NRC puts
into policing these things. They certainly could focus their attention
DR. APOSTOLAKIS: Well, I hope --
DR. KRESS: You know, even if the regulations don't change
at all --
DR. APOSTOLAKIS: Yes, that could be.
DR. GARRICK: Well, I think that our position right now is
that even though we're just beginning to get into this matter, that
there may be an advantage in us on the basis of what we've heard here to
write a letter and comment principally on SECY-99-100, and we will do
that unless new information suggests otherwise.
I think maybe it might be appropriate so that we can do that
for us to switch to the issue of how the subcommittee is going to work
and see if we can get some resolution of that, and I'm going to look to
Andy to help, and maybe Jack Sorensen to help us with this discussion.
I have two pieces of paper here that are kind of critical in
this regard. One is a memo that was written by Andy on April 27 to the
ACRS and the ACNW that's basically the minutes, I believe, of our
teleconference where we were trying to address procedural issues for a
joint ACRS/ACNW subcommittee meeting, and the other thing that I have is
an excerpt from something.
DR. CAMPBELL: Okay. Let me tell you what that is. The
draft -- the memo which is stamped "Draft" is in your notebook right
behind the agenda, in case you don't have a copy of it. It's before Tab
1, just immediately behind the agenda. But the other piece of paper
John referred to is a recommendation from the P&P Subcommittee of ACRS.
Behind that is an attempt on my part to take the old draft
protocol, which is three years old, and try and update it. I didn't
completely have success in that, in that I still have Tom as Chairman of
the ACRS, but other than that, it was an attempt to update the protocol
in terms of this memo, this conference call. So the issues before you
are a, b, c, and d recommendations from the P&P Subcommittee, and
whether or not the joint subcommittee is going to proceed by those.
It's somewhat different than the discussion during the conference call.
DR. GARRICK: Okay. Maybe we can address these one by one.
In the conference call we sort of compromised from the very
simple approach of the subcommittee writing letters itself and the
cochairman, whoever may be so designated at that time, signing those
letters, and allowing any Member of either committee to attend, and as
long as they participated in the meeting and the discussions, they were
to be a party of the letter.
Now I guess that's not enough based on what I'm hearing.
DR. APOSTOLAKIS: It's not.
There were some strong views expressed at the P&P, Planning
Procedures Subcommittee that this is a subcommittee like any other, and
it's supposed to do what you announced earlier today, namely to collect
information and so on and propose positions to the committees. Some
sort of compromise was reached in the sense that -- well, first of all,
the easy part is when we decide that one topic belongs to ACNW, for
example. Then this subcommittee doesn't get involved, unless you want
some comments from us.
But when we feel that it's a joint letter, then as I recall
we would write a letter, the four of us, we will agree, you know, I
don't know whether we'll do that through a formal vote or among
ourselves. But then that letter will be forwarded to the ACNW and the
ACRS and those two committees will have to approve, and the review,
though, will not be line by line. The review will be on the main points
of the letter, and if there is any disagreement or they want to
emphasize more one of the points, because if it goes line by line, we'll
never really finish. And then they will have to be endorsed.
In fact, as I recall, the official letter will be signed by
the Chairmen of the two committees, and not the cochairmen of the
subcommittee. So in this case it will be Garrick and Powers.
Now if the ACRS, for example -- see, that way you give the
opportunity to the full Committee to express its support or
disagreement. Now if there is a serious disagreement I suspect the
letter will come back to the four of us --
DR. GARRICK: Yes.
DR. APOSTOLAKIS: For some compromise. But also it's very
important, individual Members will have the opportunity for added
DR. GARRICK: Sure.
DR. APOSTOLAKIS: Because if it's only from us, we're not
giving them that opportunity.
DR. GARRICK: It seems that the success of this approach
would be strongly dependent upon the subcommittee Members bringing the
arguments, the discussion, the basis for the letter in a very
DR. APOSTOLAKIS: Yes.
DR. GARRICK: Fashion before the full Committee.
DR. KRESS: Yes, because they will not have had a chance --
DR. GARRICK: Right.
DR. KRESS: To hear what the --
DR. APOSTOLAKIS: Speaking of a chance, Powers was adamant
that there should be a briefing of the full ACRS, so the poor staff
would have to do this three times if you guys want the same thing. And
that's really a heavy burden, unnecessary burden, but he really felt
very strongly about --
DR. GARRICK: Well, I think if you push that to the limit,
you're back to two committees.
DR. APOSTOLAKIS: You're back to two committees.
DR. GARRICK: And there's not much advantage.
DR. APOSTOLAKIS: Well, maybe we can again find a
compromise, and Tom and I go back to the ACRS and get a briefing.
DR. KRESS: Yes --
DR. APOSTOLAKIS: And say, you know, this is what we heard,
these are the points, and this is the way the subcommittee is thinking
of going, and at least it comments from the ACRS itself to bring back to
There may be rare occurrences where the ACRS would feel they
wold need a briefing by Staff but I don't see that happening too often.
We'll try to minimize that -- but in principle each
committee could demand a briefing. We'll try to minimize that.
DR. GARRICK: I think operationally it depends upon what is
the reality here. If every time we do this one of the committees
demands a briefing, you know, we are wasting a lot of time.
DR. APOSTOLAKIS: That's right.
DR. GARRICK: Of a lot of busy people. The other thing that
I guess we have to face up to is is this a consensus position or is this
an individual position, and how do you deal with that? I think there
comes a point at which the consensus has to prevail some way.
DR. APOSTOLAKIS: Another way we are handling it now at the
ACRS, and that was based also on the feedback we got from the Commission
and the Staff, we tried to list only consensus points in the letters,
which meant that a lot of things were eliminated.
DR. GARRICK: Yes.
DR. APOSTOLAKIS: There was a lot of discussion that if you
go to the letter you do not find at all.
DR. GARRICK: It is in the transcript but not the letter.
DR. APOSTOLAKIS: It is in the transcript but not in the
letter, and, you know, sometimes you see added comments, but as a rule
that discussion is lost, so what we are doing now, we are structuring
the letter where we have conclusions and recommendations upfront and
these are the clear consensus of the committee, but then they are
followed by discussion. This is the National Academy format.
DR. GARRICK: Right.
DR. APOSTOLAKIS: In the discussion now we try to present
different points of view without really commenting on whether this is
the consensus of the committee and so on, and at least the Staff finds
this very informative, so we may want to do the same thing here and say
some members of the committee also are uncomfortable with this because
of such and such a reason and leave it at that, but I don't think the
letter should be consensus I mean on everything.
When the full ACRS reviews those letters, they really
scrutinize the recommendations and the conclusions line by line, commas,
periods, semicolons -- but the discussion is paragraph by paragraph, so
we don't really go into the detailed words, and that speeds up the
process of course.
DR. GARRICK: Yes. Well, I was going to ask you, when you
talk about agreeing on the basic points or agreeing on the
recommendations, how does that work? Is that a paragraph by paragraph
DR. APOSTOLAKIS: No. The recommendations are line by line,
but they are usually one, two, three -- very short.
DR. GARRICK: Yes.
DR. APOSTOLAKIS: Certainly they are line by line and I
believe that when each committee, ACNW or ACRS, review a letter that we
produce they should look at the recommendations and the conclusions line
by line, but the discussion they should not, because as you say,
otherwise we are back to two committees.
DR. GARRICK: Right, right. Yes.
DR. CAMPBELL: Let me just add something here. Dana
received this draft memo but in that there were a couple of concepts
that were discussed at the conference call.
One was that any member of either committee could provide
comments on any -- for consideration by the subcommittee of any topic it
is reviewing or letter it is drafting.
DR. GARRICK: It's Item 7.
DR. CAMPBELL: That was Item 7. The other was that in terms
of voting as it was conceived of during the conference call, any member
of either committee could participate in the subcommittee, so for
example Charles and Ray today are participating in the subcommittee and
under the original idea of a conference call they would have a voting
You could retain something of that in terms of continuing to
invite members of both committees to participate in the subcommittee
activities to avoid the kind of situation where you end up with three
briefings on the same subject matter of three committees.
DR. GARRICK: Well, I think there's some general rules that
we ought to be following.
One is that we ought to be tapping from each of the
respective committees the best expertise we can on the subject at hand
and, two, if somebody has a particular desire to get involved in an
issue, whether they are an expert or not they should get involved.
DR. APOSTOLAKIS: But in terms of voting though, it seems to
me that the first letter should come from the four members of the
DR. GARRICK: Yes, so the four members of the committee
generates the draft letter.
DR. APOSTOLAKIS: Yes, and clearly I mean Ray is not a
voting member, but if he says something, the subcommittee will listen,
so it is not a matter of saying no, you don't vote.
DR. GARRICK: Oh, of course -- and besides which, this
process that we are trying to get underway here, there is going to be a
lot of rotation of members and co-chairmen and what have you, and so we
are trying to maintain a very collegial approach to the whole process.
Well, I haven't studied these thoroughly, Andy. If we were
to embrace these can you interpret it for us as to what it means?
DR. CAMPBELL: The P and P?
DR. GARRICK: Yes, the ACRS subcommittee recommendations.
What is different about this than what we have on letter one?
DR. CAMPBELL: Well, the main difference is that voting will
be by either ACRS or ACNW or in some cases both ACRS and ACNW, full
The original concept in the phone conference was that the
subcommittee would act as its own subcommittee with a charter and would
vote, the four of you or whoever was participating.
DR. APOSTOLAKIS: That's different.
DR. CAMPBELL: That is a fundamental difference.
DR. GARRICK: How do you feel about that, George and Tom?
DR. KRESS: Well, you know, it was my original proposal that
the subcommittee act like an individual view.
DR. GARRICK: Yes.
DR. KRESS: I guess some of the discussions on the ACRS were
persuasive from the point of view that, number one, they didn't think
the Commission desired to set up another independent committee, and
number two, that they may not appreciate three sets of views, one from a
subcommittee and one from the ACNW and one from ACRS, although that may
not be a likely outcome.
Those arguments were persuasive to me, that we probably
shouldn't be putting together another independent committee.
DR. GARRICK: Yes, but the whole idea here again was to get
the two committees to work together.
DR. KRESS: It was a pragmatic, practical idea.
DR. GARRICK: Right, and that there's certain issues -- that
without the subcommittee concept there are certain issues that we both
address and in theory at least we could be offering different advice and
we are trying to at least --
DR. KRESS: Yes, trying to avoid that.
DR. GARRICK: Yes, so the real trick here is going to be
able to implement this additional complication and get something done,
be able to get something done.
DR. KRESS: I think, without -- George and I being very
persuasive in discouraging this business of ACRS -- they need to hear
everything that the subcommittee heard from the Staff over again, I
think that would be the part that would just not be at all workable.
DR. GARRICK: Yes.
DR. KRESS: We have to get around that part some way.
DR. GARRICK: Yes. I wonder if we shouldn't even say
something like that.
DR. KRESS: In the write-up?
DR. GARRICK: In the write-up.
DR. KRESS: Yes. Probably it ought to be better to have it
DR. APOSTOLAKIS: Say something like what?
DR. GARRICK: Say something like every attempt will be made
to avoid the need for presentations to the two committees independently.
DR. KRESS: Through the use of summaries --
DR. GARRICK: Or something of that nature.
DR. CAMPBELL: In terms of procedures that could be
accomplished by having the lead member of the subcommittee for a
particular letter or a particular issue address the ACRS or the ACNW, as
the case may be, or both committees and then have the Staff available to
answer questions but not require that they have to prepare a
presentation because, you know, part of the FTE impact on the Staff is
going through the whole process of preparing presentations.
DR. KRESS: That's a good suggestion.
DR. GARRICK: Can you represent that in words?
DR. CAMPBELL: Yes. I will put something together and run
it by -- well -- I'll get it out to you on e-mail would probably be the
best thing, if that is okay with John Larkins.
DR. GARRICK: Okay. Now as far as the issue of the
signatures of the two chairman, does anybody have a problem with that?
DR. HORNBERGER: No, but it does strike me that the Item C
is really the one that is important. It strikes me that if we decided
that it were either an ACNW or an ACRS matter, it would be more like
acknowledging that we wanted consultation from someone with expertise on
the sister committee, and Item C is really where the -- I don't think
that we would welcome going to more meetings if it weren't truly
something of interest to both committees, and therefore that is -- it
all depends on how it plays out.
If both committees insist on going through it line by line,
it is going to die. It's just going to die.
DR. GARRICK: So are you suggesting maybe some modification
to Item C?
DR. HORNBERGER: No, not at all, just an acknowledgement
that it is just going to depend upon the individuals being willing to
recognize that, okay, this is the way we are going to do it, but there
are pragmatic considerations that, as George said, go through the
recommendations and conclusions line by line because you are signing on
to that. They are the important part but to not -- because if we have
to go back and forth and back and forth four times amongst the
committees, first of all we will never get a letter out. Second of all,
we will get worn out. We won't want to participate in those.
DR. GARRICK: Right, okay -- so my summary of where we are
is that Andy is going to make an attempt to consolidate the two sets
here of guidance that we have, with the additional narrative that we
referred to earlier.
DR. CAMPBELL: Okay. I think --
DR. APOSTOLAKIS: I want to make one comment.
DR. GARRICK: And get that out to us for consideration.
DR. APOSTOLAKIS: To avoid problems in the future it is
always best to bring the stakeholders into the process early, so what I
think we should do is Tom and I should inform the ACRS in June that this
subcommittee is about to prepare a letter with these five points we
mentioned, and maybe we should half an hour for a committee meeting or
45 minutes to discuss this and see whether the members disagree, agree,
they don't think it is the right list --
DR. GARRICK: You mean the procedure?
DR. APOSTOLAKIS: No. No, the letter we plan to write --
about principles and language and so forth.
DR. GARRICK: All right, all right.
DR. APOSTOLAKIS: So we prepare them that, hey, this is
coming -- you are going to see it in July or August.
DR. GARRICK: Yes.
DR. APOSTOLAKIS: And maybe we will get some useful input.
I think that will help.
DR. KRESS: Maybe if we had a draft all ready by then.
DR. APOSTOLAKIS: Well, I would rather do it before there is
a draft, for psychological reasons.
DR. KRESS: Okay.
DR. APOSTOLAKIS: I mean if you have a draft that's great,
but it is a "rough" draft.
DR. KRESS: Very rough.
DR. APOSTOLAKIS: But I think that would be very helpful, to
give them, you know, we are thinking of putting down these
recommendations, what do you think? Because we just went through a
similar debate when we were talking about the overarching principles,
and maybe there will be some good ideas. You never know.
DR. KRESS: Oh, absolutely.
DR. APOSTOLAKIS: The important thing is not to surprise
people with a letter.
DR. GARRICK: Do we want to talk about what we want to put
in this letter for a few moments --
DR. APOSTOLAKIS: Sure.
DR. GARRICK: -- at this point in the way of
We've already identified something in the way of a first
principle, namely that we ought to be thinking as far as risk measures
are concerned in terms of dose as the parameter, and -- a risk
DR. APOSTOLAKIS: Is that really something we want to bring
up in this letter?
DR. GARRICK: Well, I don't know.
DR. KRESS: I think that's the right way to go. It's a good
time to bring it up.
DR. APOSTOLAKIS: Dose.
DR. KRESS: As a surrogate for risk.
DR. APOSTOLAKIS: I thought we were talking about higher
DR. KRESS: Frequency and dose.
DR. GARRICK: Yes, it's not just dose.
DR. APOSTOLAKIS: Oh, so what you're saying is stay away
from estimating individual risk and --
DR. GARRICK: Right. Right.
DR. KRESS: And cancers --
DR. APOSTOLAKIS: Oh, oh, oh.
DR. GARRICK: Yes. All we're saying is that --
DR. KRESS: That is a principle.
DR. GARRICK: We'll do either a PDF of dose or a CCDF. Yes.
DR. APOSTOLAKIS: The surrogate goals, I mean, we can
perhaps recommend that the staff think about it, but to actually say
what they are at this point.
DR. HORNBERGER: Well, I mean, the decision we have to make
is whether we're going to recommend principles or whether we're going to
recommend that the staff think about what the principles should be.
DR. APOSTOLAKIS: That's right. Maybe that's the first
DR. HORNBERGER: And if we're going to do this in short
order, then I think that we ourselves are not going to put forward the
principles, we are going to perhaps give some for examples.
DR. APOSTOLAKIS: Exactly.
DR. GARRICK: All right. Then it seems to me what we're
going to have in this letter is, number 1, a recommendation that the
staff develop those principles and maybe give an example so as to convey
to them when we talk about principles what we mean, and, number 2, maybe
some brief comments on this document itself, including the
categorization of activities that they have done and some of our
thoughts about that. Number 3, some discussion of methods and the need
to perhaps illustrate the continuity of the methods from one category to
another, that it's not so much a discrete individual method for one
category versus a discrete individual method for another category, it's
more a matter of degree and scope than it is kind and fundamental
thought process. So those are the kind of things --
DR. HORNBERGER: Okay. Now, John, on that last one, is that
different from the language, or is that part and parcel?
DR. GARRICK: I think it's part of it, but the language
should be part of this too, because we've commented on it, we've
discussed it --
DR. HORNBERGER: Well, as a white paper.
DR. GARRICK: Yes. Yes. Yes. So -- and usually what we
try to do after we discuss what we want to put in a letter, we put the
finger on somebody to say okay, on the basis of what the Committee has
discussed, would you be so kind as to try to generate a first draft? Is
there any reason why that procedure wouldn't work here, and is there a
DR. HORNBERGER: Is there an alternative to that procedure?
DR. APOSTOLAKIS: Well, do you want to also include the need
for quantitative objectives?
DR. GARRICK: For what?
DR. APOSTOLAKIS: Quantitative objectives, safety goals.
DR. GARRICK: Oh, yes. Right, right. The issue of safety
DR. HORNBERGER: The safety goals. Make sure we get the
DR. APOSTOLAKIS: That's right.
DR. KRESS: That and the metric one.
DR. APOSTOLAKIS: Right.
DR. KRESS: Well, it sure wouldn't upset me very much if you
made a stab at this draft, John.
DR. APOSTOLAKIS: I'll second that.
DR. HORNBERGER: We have a volunteer.
DR. APOSTOLAKIS: I move that John Garrick volunteer to do
DR. KRESS: He has a way with words anyway.
DR. GARRICK: I think that's unfair, but -- okay, we'll try
to put something together and get it out to the Members.
Very good. Any other issues, topics, relative to the
Subcommittee process from either the Committee, the staff, or the NRC
staff that we want to bring forward at this point?
MR. COPLAN: Not from the NRC staff.
DR. GARRICK: Okay. Then I think unless we have such
comments, we will adjourn. Okay?
[Whereupon, at 11:43 a.m., the meeting was concluded.]
Page Last Reviewed/Updated Friday, September 29, 2017