ML20149J863
| ML20149J863 | |
| Person / Time | |
|---|---|
| Issue date: | 12/22/1983 |
| From: | Kerr G NRC |
| To: | Bernero B NRC |
| Shared Package | |
| ML20149B526 | List: |
| References | |
| FOIA-87-743 NUDOCS 8802230257 | |
| Download: ML20149J863 (41) | |
Text
.- . DEO 2 2105) g;pa,, atey[*,, UNITED STATES
- s. n NUCLEAR REGULATORY COMMISSION i l W A&MO TON, D. C. 20$65
%,*..../
NOTE TO: Bob Bernero
! want to express my appreciation to you for participating in our State
~
Liaison Officers' meeting on December 6.
Your presentation was excellent and greatly added to the fruitful exchange of information between the States and our agency.
Attached for your infonnation is a copy of the transcript of your presentation at the SLO meeting, including the questions you received from the States.
If you want to make any corrections, please notify Sue Weissberg by December 28 (4,b'aA G. Wa e Kerr
Attachment:
As stated 8002230257 800210 s00LLhe77430 PDR W1
r g
.. , 430 y f
, i 1 it f rca th0m?
/** 2 MR. P LAINE : I think they're continuing to hold 3 onto the material.
4 You're a very encouraging group. Thank you very 5 much, it was a pleasure to be with you.
6 (Applause.)
I 7 MP. KERR: Thank you very much, Judge Plaine, we 8 appreciate it.
9 You've heard mentioned various times this morning to about the source term evaluation, which is a subject of 11 considerable interest to several'ef you.
12 Bob Bernero is the expert and Program Director of 13 the Accident Source Term Program Office, established early A_ 14 this year as part of the of fice of Nuclear Regulatory 15 Research. Bob has served in a number of positions at NRC 16 and most recently as Director, Division of Risk Analysis, 17 Office of Research.
l 1 I say he 's an expert, and one of the definitions of to an expert is ona who knows no more than you do about a subject I
so but has it organized and has slides. I can assure you he 21 has slides. So, Bob, if you would proceed.
22 MR. BERNERO: Thank you. We're going to talk today 23 about source terms in general and in particular, about a 24 recent initiative that may have come to your attention u through the press.
9s 151
- i 1 (Slide.)
2 I'd like to get the definition of terms straight 3 at the outset. When we speak of source terms, we're 4
generally saying that for a given accident in a nuclear S
reactor in this instance, that we're describing what is 6
released to the environment, what got out -- that is, wnat 7
f raction of the radionuclides in the core -- what was the 8
timing and duration of the releases did it come one hour af ter 9
the start of the event or three hours af ter the start of the 10 event, was it e puff or was it a plume, was it hot or was it 11 cold. In other words, wr 'd it rise up higher, would it move 12 along the ground.
.- 13 There is not one source term, really. There's a 14 source term for every reactor accident sequence, and 15 consequently, when you look at a plant systematically you have 1
l 16 a compencium of source terms, you add them all up, with some l 17 probability weighting to describe the risk of the reactor.
18 And later on I'll show you how we have done that in the past.
19 (Slide.)
20 I'd like to go back historically, though, and take 21 you back to the very early use of source terms because there's 22 a longstanding tradition to use source terms in nuclear reacto r 23 regulation. In 1957, a famous report, WASH-740, was done.
24 This was back in the days ' shen the commercial reactor program 25 was really just getting underway. There was not a body of ,
"70 152
. t system knowledge, design knowledge, design information, and there was an attempt to look at what were believed to be i
3 commercial size reactors. By our standards today, those were small reactors. There was a great escalation since the late 1950s into the 1000-megawatt reactor size. In those days, t the big plant was believed to be 150 megawatt, electric that is So basically, what thit study did is it went in there and said, I would like to look at the probability of severe accidonts; that is, core damage or core melt accidents, 10 and I would also like to look at the consequences of such II accidents.
12 Well, the technology and the data base did not exist for probability.' That was added subjectively because, 14 well, they didn' t even have the systems delineated in those 18 days and their methodology was still in early stages of 16 development. So really what they did was they defined hazard 17 state, hazard conditions, degrees of damage, and they gathered ,
it source terms for those degrees of damage.
19 The first one was called hazard state 1. And what they said is hnzard state 1 is what's going to prevail if you damage the reac' tor core and release the radioactivity from the fuel and from the cladding, but it's still in the reactor '
23 coolant system. Se your first barrier is breached, but not 24 the reactor coolant system, ad certainly not the containment.
s to Then,they said hazard state two, you will not only
v
.. 153
. i
, 1 broach the fuel in cladding, but you will breach -- by melting the core you will breach the reactor coolant system.
3 So the.t gets you to core on-the floor. And in hkzard state 3 you not only breach the fuel cladding, melt the core and 4
breach the reactor coolant system, but you fail the contain-4 ment as well and you have a substantial release outside.
7 So, three hazard states.
8 And what they developed was that they could 9
estimate the radioactive rather crudely but not bad, 10 considering the lack of data at the time. And when they got to hazard state 3 they said well, it's a volatile activity 12 alone, you know, the gaseous and very volatile things got 13 out, and you'd get this kind of of f-site consequence. Or
! 14 if 50 percent of the activity got out, a bigger source term, i
[
you'd get even bigger consequences. And the famous numbers 16 of thousands of deaths and many dollars of damage, which led 17 to Price Anderson and things like that, came from that last 18 hazard state, hazard state 3 and the 50 percent release.
n It led to our regulations, the regulations for
! n nuclear reactors, having design basis accidents that you
) 21 mustn't fail the containment and a core on the floor source 22 term, TID-14844, which was published in 1962 or 63, and 23 basically, that core on the floor source term was a test of 24 whether containment was adequate. Unfortunately, it looxed at fission product activity; how much is there, but it did
154 - -
(
, i
=I i '
not include thermal hydraulic force. 'So the hea t, the
- f. .
3 !
temperature, the pressure, the gases generated by core melt ,
were not in that source term; only the fission products.
4 Now, in early centainments you had a big contain-8 !
ment and a small reactor. Later on, we evolved to a rather I small containment relative to the sir
- of the reactor, and i 7 :
that turns out to se a loophole in our design and regulation.
8 (slide.)
' i Later on -- this is now up in the 1970s -- the i 10 i 1 reactor safety study went back, took two reactors, one '
l II pressurized water reactor and one boiling water reactor, and !
5 II
('
looked through and looked at all the accident sequences, did ,
I3 I-the best job it could on the probability of those sequences, ;
I' ;
analyzed core melt phenomena and analyzed containment perform-II f ance in detail.
18 i And sure enough this thing showed up, that from a !
II :
containment point of view, the pressurized water reactor, I8 which had a relatively large passive containment, looked I' i 1
better than the boiling water reactor which had a pressure
- suppression containment, a LOCA condenser. Really, not a II big , passive containment. And that study showed that for both t
" (
reactors, containment is the principal fission product barrier.
It is the ultimate defense and it's the largest and most 4
I' !
effective defense against the release of fission product. !
i Now, many probabilistic risk analyses since that !
. As>
g I .
t I
time h ve used the same models that WASH-1400 did for the !
<* 8 fission product transport, the physical chemistry of it, i 3
There's a report we published -- one of our contractors, ,
4 Sandia Labs did it. We published this in the fall cf 1982, ;
8 ;
about a year ago, a little over a year ago. At that time, '
8 }
we were trying to develop siting criteria. The Congress i &
7
. wanted us, then and thth., after TMI to develop siting i
e f
criteria and regulativas, a i we took the available models 8
really based on WASH '400 of accident releases or source i
10 terms.
And we synthesized the risk characteristics of a :
11 typical light water reactor, just averaging basically what '
18 we knew. And then took that typical set of source terms --
13 it was a much simplified one -- and we analyzed 91 U.S. sites.
14
. So if you go to that report you can find a 16 detailed site risk analysis for a hypothetical or a typical 14 light water reactor, and every site that has a reactor on it --
17 and many sites that don't; for instance, we did Allen's is Creek and that plant was cancelled. So you'll find 91 U.S.
19 sites all analyzed, and what was done was a generic risk 30 comparison, sensitivity studies to find out what the more 21 important aspects are, and that demonstrated that the source 22 term itself, particularly the source term of the largest 23 release, was very important.
24 (slide.)
as We are right now in the nidst of a systematic
156 D I I
consideration of -- oh, it's 140 plants now what with cancel-
- - 2 lations, but of all the types,dif ferent vendors, boilers, PWRs, different architect engineers, different containment i
types. We at the NRC and our contractors are doing an -
5 I enormous analysis of severe accident risk in these plants. i 8 i And in addition, the industry's IDCOR program, industry 7
degraded core program, the industry program is doing a program 8
similar in concept but quite independent in who's doing it g 8
and how they're doing it. Different computer codes, different to models, in many respects.
11 They have nearly completed their work and we're just 12 now at the stage of comparing notes technically in open 13 meetings where we say this is how we modeled it; how did you
- 14 model it, what data did you use, where do we agree, where do 15 we not agree.
(Slide.)
17 In this major prer .or m part of it is what people 18 source term, highly specialized toward fission product 18 transport. If you go back to 1981, the NRC published a document, NUREG-0772; the technical basis for assessing source 21 terms. We're now going to publish a sequel to that, the 22 technical basis for a reassessment of source terms.
23 We have four elements to that work. Much of this 24 material is available, if you wish to pursue it. The first 25 element is, what basis do you have for predicting fission
at
. i 3
, , product behavior during a cora melt. That is certainly some
/~ ' . 2 tough science; very high temperatures, very complex physical 3
chemistry.
4 We have the lead at Oak Ridge National Laboratory 5
in a report that is very nearly complete right now. A draft 6
was available earlier,this year, and we hope to have this 7
better version available in about ene week, in which we 8
summarized. for all regimes of the accident what is the data base and what are the models.
10 Meanwhile, us_ing these models, these computer codes ,
II we have Battelle Columbus Laboratories preparing a multi-12 volume report, 3MI 2104, Battelle Memorial Institute 2104.
13 It's the Source Term Estimates for Selected Plants and
~
14 J
Accident Sequences. So basically, it takes postulated or 15 identified important accident requences and goes through very 18 carefully, using the latest data and the latest models to 17 calculate what gets out, how much of the fist 'on product 18 activity will move from one place to the other, and what are 18 the attendant generation of steam, hydrogen, gas, other non-condensible gases and the like.
21 We have a third element in this work and it's 22 extremely important.
Because of the dif ficult scientific basis, 23 we want a thorough peer review.
24 We divided it into two levels of effort. What we call the technical experts review is we 25 deliberately went out and selected people for their personal
1 158 1
. i expertise, whether they worked for the government, for a 2
university, for a seller of reactors, for an architect 3 i >
engineer. i Never mind conflicts of interest; we want people :
4 t who really understand this stuff. We knew who they are, we 5
know who they work for. If there's a conflict of interest, 6
we'll live with it because we 're going to do it openly.
7 We selected technical experts on that basis, and 8
we have been meeting for a year now periodically as elements 9
of this work necame available. We have these technical expert a 10 looking at it and openly criticizing it. They would read it, 11 they'd study it, we'd meet publicly and they would say, ah, 12 you're misinterpreting that test, or, you're being too 13 C- conservative over here, or, too optimistic over there.
14 A very good feedback mechanism.
15 In addition, we have made arrangements for an 16 independent scient.ific organization, the American Physical 17 Society, to do a wholly uncoupled scientific review, a broad-15 based review of the basic science here; is that good science 19 or not.
And they are presently underway with their investiga-20 tion.
21 We also have a lot of homework to do in the staff n
for the appraisal of the risk and regulatory significance,
' 23 because when I lay out these accident sequences, it fairly 24 described the reactor when I take a balanced or pro rated set, 25 when I look at the . accident characteristics of the reactor and i
159
'77
. s I
say the predominant sequence is this one; 30 percent of *.ha
.- g time it's this accident, 40 percent of the time it's that i 3
accident, and so forth. Then I get a matched set that describt 4
the reactor.
5 l And I have to do this here (indicating ) . This is i
also where I look~at IDCOR'o work and compare notes, where I l look at uncertainties and sensitivities.
8 So those are the four elements of the source term 8
work. The plants we're doing -- some of you may have these 10 in your states.
II (Slide.)
12 This is the Battelle report, and you have copies 13
,,. of these notes, I bell' eve. We're doing the Surry plant in C I4 Virginia, the Peach Bottom plant in Pennsylvania, Grand 15 Gulf in Mississippi, Sequoyah. Surry -- again, the reason 16 for Surry being done twice is a very important computer code, I7 the MARCH code. When we did the first experts ' review, they 18 dumped on that version of the code so. heavily and said you've
-I' got to fix those deficiencies in the code that we've known about, and we have fixed it to do the MARCH-2 code from here 21 on. But we had to recycle and go back and do that one over 22 again. And we're doing the Zion plant.
3 Another peer review comment was that we had to pick 24 up that plant, also, to have a representative set. So it's a
, 25 set of five plants here. There 's one more plant that we 'll
)
78 160 I
add to this, it's the Limerick plant in Pennsylvania. It's
- /"
2 a boiling water reactor with a MARK 2 containment. It is not 3
represented by any one of these reactors, and it does represen t 4
a set of about a dozen reactors.
5 And we have an extensive effort in the licens'ing 6
st.af f to review that, so we' re going to amalgamate that and 7
have a set of six.
8 Now, all this shorthand is just nomenclature.
8 They are shorthand abbreviations for what accident sequence 10 is being analy' zed.
11 (Slide.)
12 Basically, let it suffice that I say that those are 13 the important accident sequences.
14 Now, I included a picture, in case someone wants 15
, to refer to it, of a boiling water reactor, a MARK 1 contain-16 ment, and should questions come up we can refer to that.
17 But basically, let me just talk now on more programmatic I8 things.
19 The American Physical Society study is focusing on 20 those elements 1 and 2. What is the data base that you're 21 working with, how can you be sure these codes describe things, 22 how can you be sure of your source term estimates. They need 23 to understand the context of how we would use these, but their 24 basic focus is, is this sound science or is it mumbo-junbo.
25 In fact, it's popular nowadays in the comments on the use of
79 .
161 1
probabilistic risk analysis -- many people in the public 2
interest groups have adopted the term "voo-doo engineering" 3 and they like to say that. l You know, their taking a hard i 4
look at a severe accident is called voo-doc.
5 (Slide.)
6 There are some schedule factors that make it very 7
difficult to give you a hard time line, so I hope you can 8 understand. There's a lot of pressure in early or mid-1924; 9
in other words, the beginning of next year to mid next year, 10 to stop wringing our hands after TMI.
11 Remember, the Three Mile Island accident said 12 these accidents really can happen, and maybe we ought to 13 fix something. Now, w6 fixed hydrogen. The hydrogen control
- k. . .
14 systems on plants were fixed, but when are we going to drop 15 the other shoe? Many people in the industry get a lot of 16 feedback. Their bankers keep jacking up the interest rate, 17 there's too much regulatory uncertainty, there's a strong 18 pressure to close the book on TMI.
19 There's also a strong pressure -- we get letters 20 from people constantly: re-evaluate the source term. Many l
21 people think that Three Mile Island brought us news that the 22 source term is dramatically lowered; if you really look hard j 23 you'll see that it is so low, you can eliminate half the i
24 regulations. And we get letters, people clamoring, please, 25 sort it out, stop this nonsense , stop the hysteria.
1
i 163
'I Those are pressures to come to a decision. The 2
work that we 're doing and the peer review -- and it turns out 3
the American Physical Society is a sort of critical path --
4 their work is they're doing their study during 1983 and 84; 5
they won't sit down to write until the summer of 1984 at the 6
earliest.
And the earliest publication or public availability 7
of their comments is November l'984. They have a routine 8
public meeting at that time.
9 So that leaves us with the prospect that until we 10 hear from them, it's most unlikely for the NRC to say here's 11 the final source term. You know, this vital peer review just 12 won ' t be available.
I3 1 I
Many other people are looking at it and saying i 14 well, do you really have to go to a revised source term? l l
15 Maybe it is lower, somewhat lower, but what difference will 18 it make? Do you really need to go that far? And there are I
many possibilities for doing these now for an orderly I8 progress without chasing the Holy Grail, without looking for 18 that new source term.
(Slide.)
21 If you go in -- this is a near-term subject. When 22 we recognized that the American Physical Society's schedule 23 was, if anything, stretching our whole re-evaluation, what we 24 said to the Commission was hey, we're thinking of looking at 25 emergency preparedness rulemaking now; take advantage of what
81 163 g
<- we've learned f rom experience, make sure that we don ' t do 2
something foolish that would be overturned tomorrow by a.
3 revised source term, but maybe we can do something now.
4 Let me go back to the documents that are the basis 5
of energency preparedness. Some of you may be familiar with these. There are two documents; NUREG-0396 from 1978, and 7
NUREG-0654 from 1980. They are joint documents. The EPA and 8
NRC published this one, which basically says 'aere is how risky a nuclear reactor is. It used the WASH-1400 analysis. It to basically uses the WASH-1400 source terms and describes the 11 lethal reach or health-threatening reach of a light water 12 reactor of that size and type. It uses the EPA protective l *'
action guidelines and 'says you ought to take steps so that people do not get radiological doses in excess of the 15 guidelines. And certainly take great pains to plan so that 16 the people's lives won't be threatened, which of course is 17 at much higher doses.
18 Generally, people will identify 200 rem whole body 19
. as immediate life-threatening doses. That's a very grave l radiation dose. It was on that basis that this report said 21 for the immersion path you ought to plan ahead of time a 22 radius of about 10 miles around the reactor. And for the food 23 j
chains, you know, the milk and the iodine and so forth, plan 4 l
I 24 on about a 50-mile radius. And in a joint report we prepared 25 with FEMA we took that and based on that and other factors,
164 ,!.;
l
- ' I \
said here are good ways and planning standards for preparing
/ 2 emergency plans, criteria, planning standards, for emergency 3
preparedness.
4 (Slide.)
5 We' re looking with FEMA at that. And in the mean-6 while, if we look at this curve -- this is a curve drawn or 7
copied directly from NUREG-0396, the one about how risky is 8
a reactor.
' And notice here it says if you have a core melt; 8
that is, given a core melt accident, the conditional probabili q 10 of people suffering a dose looks like this. And the curve 11 for 200 rem, life-threatening dose, is very interesting.
- 12 What it says is th t if you have a core melt and 13 you recognize that a core melt without containment -- that is, 14 where the containment failed very early -- will release a lot 15 of the radioactivity, there's enough of it there and the odds 16 of the weather dispersing it are such that a life-threatening 17 dose is of almost constant probability for quite a distance 18 out from the reactor. And then the amount of radioactivity 18 involved and the dispersing effect of the weather is such that 20 the likelihood of having a life-threatening dose drops 21 dramatically af ter you get out to about 10 miles.
22 Now, if you look at that curve, it says it's the 23 unmitigated core melt that is a very significant factor in 24 selecting the 10-mile planning radius. And if I say all those ,
25 y
scientists who have been working on source term, what do they
'03 165 j
)
I see so far, they see a lot of cases where an accident I 2
sequence had an early containment failure, but you studied it 3
more carefully, the containment failed later. It's a lcwer 4
l category accident.
5 Well, if you re-analyze the accidents and move them 0
l from the earlier failure containment category to the late !
i 7 I containment failure category, what you have done is you've j t
8 taken that whole curve and you slide it down, and you change i j
8 !
its probability.
i 10 But'if you have an accident that releases that 11 large inventory, it will still have this shape, it will still 12 reach 10 miles. You just have fewer of them.
13 So if you'rs going to change the source term that 14 l way, by just refined containment analysis or better containment 15 performance, you have to move this thing down, down, down, 16 until it's literally off-scale. So that you say it's so low 17 now in probability I'm not going to worry about it.
18 That's like'the big jumbo jet falling in the 19 football stadium. You know, the odds are so low as to be 20 negligible. It's going to have to come down, I would assert, 21 a f actor of 100 or something like that. I don't see that.
22 I see it coming down maybe a factor of 10, but I don't see a 23 ~
factor of 100.
24 Then if I turn to the physical chemists and s 25 I say look, fellows, NASH-1400 said a certain amount of
soc 84 I radioactivity will come out of the reactor coolant system 2
whether or not you have a containment there -- and these are 3
accidents where the containment is not going to do you any 4
good; only the attenuation or holding power of the reactor 5
coolant system is working for you -- are you coming up with 6 different numbers ?
7 As I said, we have copies of BMI-2104, the physical 8
chemists' analysis, and they are changing one nuclide by a 9
l factor of 2, another by a factor of 3, another by a factor 10 og 4, but they're Tefinements , they're subtle changes, they 11 minor changes. And what they amount to is maybe they reduce 12 what gets out by a factor cf 2.
13 Well, if I draw this curve, a factor of 2 would 14 be right parallel to it and it would come down something 15 like that (indicating) .
16 You ask me to stand across the room and what's 17 an approximate planning radius? Ten miles. I'm certainly 18 not going to go out and send you word, dear folks in the l'
, states, change from 10-mile planning to 9-mile planning. That 20 would be absurd. Unless I see a dramatic change in the 21 physical chemistry and a really, really dramatic change in 22 the probability, 10 miles is a good zone.
23 What other problems do I have?
24 (Slide.)
25 If I look at experience, -- and my favorite item
.....g,...-.--
85 1
of experience is the TMI accident itself -- let me show you 2 f rom NUREG-06 5 4, this -- I'm sure you've seen this figure 3
before -- this is the keyhole picture where you have the 4
reactor site and the illustration to show the various steps 5
of emergency preparedness priority.
6 Obviously, if you're going to cope with a severe 7
accident, the highest pricrity goes to the people closest to 8
the site in the emerging pathway. You go for a short radius 9
and then the downwind sector, and that's why you get that 10 keyhole-shaped thing , and you work your way out.
11 But what has happened in the years since TMI, in 12 the years since this strong attention to emergency preparednes s 13 came in? How many pechle do we all know who think 10 miles
(- 14 is not a planning zone; it's an action zone. Actions are taken 15 in increments of 10-mile radius.
16 But the emergency preparedness response is everyone 17 i in 10 ndles, run, all at once. teolly inappropriate, wholly 18 wrong. You go to the existing guidance and it says over and 18 over again things like, it would be' a good idea to start with 20
[ 2 miles and work out. Initial evacuation of 2 miles to 3 21 l miles, or 2 to 5. But over and over again, it is a suggestion ,
i 22 It is an exhortation. The police power, the authority to 23 direct evacuation or any other protective measure, is yours; 24 You know that, it is not ours.
25
- s. So what we said was maybe we could simplify the ,
1 l
l
co 168 1 process.
The guidance virtually enshrines 2 miles, so why
- ~'
2 don't we make 2 miles as mandatory as we can make it, and 3
link it directly to the definition of a general emergency.
4 So what we did was we sat down one day, recognizing the 5
schedule problem, and we said let's clean up the regulations 6
and emergency preparedness in the following respects.
7 (Slide.)
8 Let's see if we can't sharpen the emergency zone 9
thinking with a little more teeth for that 2-mile action so 10 that people won' t sit around there wasting valuable time, 11 trying to decide whether or not they should follow the 12 guidance.
Let's be sure that we look at the sensitivity of 13 the source term changes so we aren't doing something dumb 14 that's going to be overturned in 12 months.
15 We've got problems, as you know. Is FEMA finding 16 acceptability of the planning standards , or is FEMA finding 17 acceptable off-site safety? We 've had grave problems about 18 exactly what is the legal chara ter of the FEMA finding as 19 against the NRC finding. Wh'o is responsible for the health 20 and safety of the public and the finding thereof ? We've had 21 real problems with that. We said, you know, we've got to 22 clean up the FEMA memorandum of understanding and corres-23 pondingly, clean up the regulations to match it, to clarify 24 the finding.
25 We've had problems with the four-month clock.
'87 169 1
Remember, it went in there in the first place where the 2
licensee might become involved in a situation where the change 3
needed was in your hands, not his. However, any emergency 4
preparedness thing is argued sometimes that you ought to have 5
a four-month grace period. But if there's something that's 6
clearly in the licensee's control and we find that, you know, 7
our normal enforcement action would not just automatically 8
give you four months we say look, what's the problem? Fix 8
it. We need to sharpen that and clarify that.
10 And lastly, this is one that causes us some concern.
11 Appendix E for emergency planning is an appendix to our 12 regulations. The Federal Register, which is a body, not
~
13 merely a booklet, the Federal Register authority has criticized 14 the NRC many times and said it's bad regulatory practice to 15 put regulations in appendices. It leads to legal problems.
16 And we have been urged for a long time to move the contents 17 of the appendices up into the body of the regulation for 18 clarity, for legal robustness.
19 Now, that doesn't change the requirement. It 20 doesn't just change the standards; it moves them, it 21 '
editorially shifts them. We have prepared to do so, but l l 22 we're concerned. We don't want to confuse people and we don't!
23 l want te create a smokescreen. We don't want to have an l
24 apparently substantive change confuse people and make them 25 think there's something there that isn't there. So a lot of l
l 1
88
. . 170
. o 1
l word movement in our current draf t is merely lif ting a whole l 2
section from the appendix lock, stock and barrel, and putting 3
it in the particular paragraph of the regulations. But we're .
t 4
very apprehensive that we don't mislead by that.
5 (Slide.) I 6
The one substantive part of that regulation that's 7
getting a lot of attention -- we said on emergency zones would 8
we change planning? Answer: r.o . We clarify thet the 9
emergency planning zone , about 10 miles, is a planning zone 10 and remains 10' zone. Would we change notification? The 11 answer is no. We would retain the capability for prompt i 12 notification throughout the 10-mile emergency planning zone.
13
(.. ,
How about rdsponse? And this is where we say 14 let's put a knee-jerk response in here within the legal 15 authority of the MRC. Let us link action to the declaration 16 of a general emergency. Remember, if you look at the guidance 17 now it says in a general emergency you either wrecked the core 18 already or you're about to. You have damaged the core. It's 19
, not an everyday thing. It's not an alert or a minor thing; 20 it's a very grave situation.
21 And this says here is an action; move. The 22 recommendation is automatic. And you must show -- and we've 23 included a time line -- show that you can evacuate promptly.
24 From our studies, it appears to us that a 2-mile zone around 25 the reactor which can be evacuated in about two hours under
89
.s 171 O
- 1 typical conditions is consistent with this kind of immediate 2 action. When you look at the different accident sequences 3
and you look at how long it actually physically takes for 4
a core to melt, you will have a compatible match. You'll l
5 outrun most of them.
6 I can still give you what they call the "web 7
scenario," you know, where you're in some far-fetched situatiot 8
that's going to fail the containment in 120 seconds, but their-9 probability is very, very low. I don't know exactly what it to is but it is very low. So this makes a logical set, that 11 you have a h,ead start, it is done in a time-consistent fashion; ,
12 There are not a whole lot of people there in those first 2 13 miles, and in effect, what it does is it gets you started so 14 that now your dose projection or condition projection thinking --
15 at least you have these people out of the way. And the fact 16 of the matter is, 9 times out of 10 or 99 times out of 100 17 that's probably all you'll have to do. It's precautionary 18 action and you'll be in a much more favorable position to 19 say, take a little more time to think it through before 20 anything else is done.
21 You've got the concentration, the priority of 22 emergency action focused where the risk is highest; right at 23 the fence, The people who so close they can see the cooling 24 tower.
25 Now, the remainder of the EPZ remains as it was
90 l 172 I before; you link the action to conditions. You've got the
/ e 2 ,
capability to shelter or to evacuate but it's a much more 1 3
deliberate thing, you ' ve got more time. The accident can 4
develop a little more, you can see where the scenario is I 5
going -- really bad or not really bad. '
l l ,
6 We think this is the best way to deal with that.
7 Unfortunately, we prepared this draft for internal review, 8
I signed a memo on September 16th,1983, to Division Directors 9
within the agency to get a broad review. On September 28th, 10 just 12 days later, Congressman Markey wrote a request that 11 said, I want the memo signed by Robert M. Bernero on 12 September 16th that said so-and-so and so-and-so. He may even 13 have had it by then for all I know.
A.
14 It went out and it appeared in the New York Times.
15 A description of it appeared in the New York Times that was 16 erroneous. It said, 10 miles is reduced to 2 miles. Could 17 you interpret that as 10 miles is reduced 2 miles (indicating) ?
18 But that's what went out and many people think that's the 19 case. I know Lieutenant Governor Delbello of New York is 20 still very upset about it; we just got another letter from 21 him. It's an unfortunate thing when people take things out U
of context and move them out prematurely. Lieutenant Governor 23 Delbello is very upset. He d oe sn ' t like reading about in 24
. the New York Times, and neither do I.
25 You understand -- and I took great pains to cover
173 4
1 this here -- this is what it is. I'd be happy to answer 2
questions on it or the source term in general, but we have 3
not only no intention to engage in subterfuge with emergency 4
preparedness regulations now. As I've indicated to you, I 5
don't think a whole lot is going to change in the near future.
6 MR. ALLARD: Allard from Virginia. As I see it, 7
I don't see a great deal that has changed as it exists in 8
0654 right now, and as we conduct -- in other words,the 9
scenarios are written that way with NRC and FEMA blessing 10 that we emphasize the 2 miles. Yes, we emphasize the 11 evacuation of 2 miles, and then we go as far as we want, 12 wherever we want --
13 I,
MR. BERNERCH Yes. The only problem we see is --
' 14 it's not a perfect example because it's TMI -- at 7:00 o' clock 15 in the morning at TMI they had a little procedure where if you 16 get a high range gamma reading in the building as they 17 did, you fill in the blank with the reading on the gamma 18 monitor and you feel in the blank with the current reading 18 on the meteorology meter, you know which is a chi over Q, and 20 you run out the calculation like it says, and it gives you 21 the dose rate in Goldsboro, Pennsylvania. And they did that, 22 and it was a dose rate in excess of lor per hour; however, 23 the equation presumes that leakage is at the design basis 24 leakage, which is a tenth of a volume percent per day.
25 A monitoring team showed that there wasn't such
174 il
. . l I
leakage; you know, that nothing was getting out. So that situation went into an evaluation phase.
In our current procedurer., if you look at the 4 ,
circulars that Ed Jordan has put out for plant operators to 1 1
5 try to figure out, or if you look at Appendix 1 of NUREG-0654, 6
there is here and there a prediction of whether the containment will fail, a prognosis. It puts too much prognosis burden 8
on the plant operators or on the state and local authorities 9
to say do we really think it is going to come out.
10 You remember at TMI it was a threat to come out, 11 it was down to the last barrier, and two days later there was 12 a lame decision to start moving people. What should have been
/ done was the very first day, a precautionary evacuation, an
~~
orderly thing.
The public would certainly be better served 15 by that.
That is a responsible attitude, and then they would 16 feel confident that their interests are being served, instead 17 of milling around in confusion for two days and then selecting 18 pregnant women and children.
19 Now, if you look at the current regulations, this 20 is clearly there.
The current guidance, I should say, not 21 regulations.
22 It's clearly there. This philosophy is clearly there.
The only thing absent is this last push to get it 23 started, and that's all we felt was needed.
24 Now, you may not agree with that. You may feel 25 that the authority is there, the advice is there, the
. 93 -
175 -
I recommendation is there , that's enough. You may not share our 2 feeling, or at least my feeling, that that push is valuable.
3 That mandatory recommendation -- that's really what it is.
4 The NRC feels don't sit there and analyze it, don't wait for 5 some guy in the plant to predict whether or not his contain-6 ment is going to fail.' You know, no matter how hard he's 7 trying, he's no seer. He's got a lot of pressure and a lot of 8 confusion.
9 The NRC feels -- or at least I want the regulation 10 to say the NRC, feels that you ought to move right away as a 11 precaution. If you've got the symptoms to declare a general 12 emergency, you've got the symptoms that ought to move the peopl-13 out at 2 miles, and the rest is up to you.
14 SPEAKER: As I say, we read exactly what you say 15 in most of our exercises are carried out, and I don't see 16 anything new. You have added another sector maybe out to 17 5 miles. You know, if it looks like the meteorology is 18 going to hold, that's a good thing to do.
19 MR. BERNERO: Yes. In site-specific features, 20 somebody might say well, I've got a screwy geometry here 21 because of some town that goes out to 3 miles. Sure, that is 22 additional action and we certainly wouldn't say only 2 miles.
n SPEAKER: This also helps to allay the FEMA people 24 and emergency management people who want to ,iump to evacuation u right quick. If you're doing some evacuation --
6
94 176 1
MR. BERNERO:
You've hit the nail on the head. '
,.- 2 SPEAKER:
You've given them busy work and you're l 3
putting them to work, and you're doing something --
4 (Laughter. )
5 MR. BERNERO: You know, the thing that gets me --
6 I tried to make this point. I was talking to congressional F
7 staff last week. If I look at a normal site, it is natural i 8 I for the population density close to the fence to be lower 9
than the population density out here (indicating) . So the 10 population density increases with distance out to 10 miles, 11 and on top of that, since it is a circle, you've got more 12 people in the last mile of the ring, the one-mile ring at 13 the end, than you have .in the first-mile ring.
1,. 14 If you ignore the risk versus distance curve, 15 you're going to be driven subconsciously to give your 16 planning emphasis to the larger population, and you're going
~
17 to start sweating how well and how effectively can I plan to 18 move the people out of this condominium, rather than five 19 people in a farmhouse right at the fence. And they are -- if 20 you look at a risk versus distance curve, the risk here is 21 10 to 100 tires higher than the risk here (indicating).
22 It's just a perversion of priorities. ,
23 MR. GERUSKY: Tom Gerusky, Pennsylvania. Won't 24 this tend to get you into the decisionmaking on Q1en to declare
, 25 a general emergency rather than when to declare an evacuation?
l 95 c:
\ 177 I
I'm not opposed to it, but.I can see us arguing with the 2
utility -- hey, you'd better declare a general emergency and 8
they say ch, no, no, we don' t think so yet. And you know, -- 4 4
if the general emergency doesn't evacuate automatically, 5
you've got a problem.
6 MR. BERNERO: You can't do this without the 7
corollary in NUREG-0654, Appendix 1 on that very point, Tom.
8 The definition of a general -- gee, I meant to bring my 8
NUREG-0654. But if you go there, the' declaration of a general 10 emergency isn't as sharp as I would like to see it be. I'd II like to see it say hey, you have high radiation in the 12 building, like TMI. No doubt that that's a general emergency.
13 And the other thing is' (b) a general emergency can be declared 14 on an approximate condition, and let me give you an example.
15 We're trying to sharpen those words so that we can 16 make that sharp definition that won't be confused. A station 17 blackout. You know, total loss of AC power. All plants have 18 a capability that is independent of AC power to temporarily I'
cool themselves until you can restore the power. The 20 capability, the duration of that capability -- six hours, 21 eight hours -- you know, will vary from plant to plant.
22 Now, I don't want to wait until the core starta 23 to melt to call a general emergency in that plant, but I don't 4
i want to declare a general emergency the minute it loses AC l
25 power. I want to have a definition that says AC power, a L
96 178
. .. ~
I 1 total function, has been lost for a substantial period of time,
(. '
2 and tie it to a condition that, let's say it's a plant that 3
has six hours' capability for AC-independent cooling, and at 4
three hours -- if you haven't cooled it by then, I'm going 5
to, as a precaution, declare a general emergency.
6 But we 've got to get out of that. We don't want to 7
have a Philadelphia lawyer on the plant staff interpreting.
8 That's the very thing we want to get away from, is 8
interpretation. He cannot -- we 've had this complaint many 10 times. The plant staff cannot interpret the likelihood of 11 containment failure. Theycan'treakanytealeavesor 12 crystal balls, and certainly in the upset of a real emergency 13 they would be even less able to, than they could in the 14 common environment.
15 so we must have a definition of general emergency 16 that is pure plant symptom. The meter says it or the meter 17 doesn't say it. The condition is met or the condition is not 18 met. But a very clear condition. And we are working with 18 FEMA on a re-draft of NUREG-0654, including a redraft of 20 Appendix 1, that would be circulated. We have a task force 21 working on it right now and I have pieces of'the drafts coming 22 to me. They will be circulated out through the state review 23 channel very soon. I couldn't tell you exactly. We have a 24 mechanism to get it out to you, and we'll be looking at that 25 very point, the definition of general emergency, among other .
r - _ _w _ - -
179 1 things.
r- 2 MR. EISENBERG: In your chart of the extremely low 3
probability of the high dose -- the life-threatening dose, 4
200r, it didn't seem to make much difference as far as - .
5 There wasn't that much variation as far as distance was 6 concerned.
7 MR. BERNERO: Well, this thing here, -- yes, there's 8
a certain inventory in WASH-1400 which is the basis of this 9
70 percent of the iodine and what have you. And it's going to to get scattered to such a degree that you just don' t have 11 enough in a puf f to get a dose in excess of 200 rem at this 12 distance.
13 MR. EISENBERG: All right. What I'm saying it it
\
14 didn't make that-differentiation in the O to 10 miles.
15 MR.'BERNERO: Oh, no.
16 MR. EISENBERG: What I'm getting at, then, is: is 17 your 0 to 2 precautionary evacuation tied to the emergency 18 declaration for this kind of a situation? In other words, if 19 there's no rationale for the O to 2 --
20 MR. BERNERO: Oh, yes, I see what you mean. No, 21 the O to 2 -- in the first draf t we called it a prompt action 22 zone, and one of our people criticized that and said you know, 23 that has an implication that that's the only prompt action.
24 And we're thinking, you know, it's more like a minimum action 25 zone. It may be the only action zone, but for one of these
- 98. 180
\. . .
1 it's a minimum, it's the headstart and you're going to be 2
doing other things immediately right after it. Because if 3
you have a grave core melt accident with containment failure, 4
depending on the weather, this is average. The weather could a
reach even farther. That 's another point. You may be moving 6
even beyond 10 miles.
7 And you know, we've had feedback where people say .
8 I'm safe, I live outside the 10-mile zone. You know, in a 8
! really bad accident with bad weather it could reach beyond 10.
l 10 So the action k- this would be a minimum. Tvo ndles would be 11 a minimum and other action would be logically dependent on it.
12 MR. EISENBERG: I guess the point I was trying to 13 make, in that low probebility situation where you do get the I 14 core melt with the proposed release of 200r, there's no
- 15
, rationale for choosing 0 to 2 in that kind of a situation.
I 16 i
MR. BEDNEDO: For only 0 to 2.-
17 MR. EISENBERG : That's correct.
18 r MR. BERNERO: In other words, we give you the l'
, automatic 0 to 2, the plant conditions have to give you every-20 thing beyond it. There is a burden, there 's a shif t, a decisio:
21 burden, to the plant staff and state and local authorities 22 I to read the leaves there and say boy, this is even worse than l 23 the minimum, it's developing badly. We 've got to do -- but ,
24 remember, we've also automatically started the process by I
' 25 taking the people at the highest risk and moving them first, f l
s:
181 1
The cloud gets to them first, they're at the highest risk.
/ -
2 MR. WILLIAMS: During the NRC's directive and the '
3 nuclear utilities' installation of the emergency notification 4
system and the health physics network for communications , will 5
the NRC headquarters team abide by this particular method in 6
making recommendations to the utilities and to the states?
7 or will they continue to use the emergency action level 8
predictions and make recommendations for protective actions 9
to the state governors when we 're at site area emergency anc 10 when we're in conditions beyond containment releases? <
11 MR. BERNERO: Well, in the revision of the rule 12 that we're considering and NUREG-0654, we're going to reduce I l
,._ 13 yes, the unusual event. gets eliminated. We keep the alert, k' 14 j
we keep the site emergency and we have the general emergency. I 15 Recommended off-site actions are only for general emergency.
16 For site emergency it's alert and stand-by but you don't 17 actually make any recommundation off-site.
18 And we feel that this mandatory recommendation is 19 already in the system, and we're just reinforcing it, making 20 it stronger. And that what we need is a set of coherent, 21 symptom-oriented emergency action levels that define these 22 conditions in the plant,which define a general emergency so l
l 23 that they're not subject to interpretation and confusion.
24 And once you get to those defined symptoms or plant 25 conditions, the definition of emergency action is as l
100 182
.s .
1 appropriate; you know, if there are conditions of a general
~f*.
2 emergency, you've got a general emergency. And then having i 3
this minimum 2-mile action as an already pre-decided .
4 recommendation, the next burden is judging and recommending 5
whether anything else is in order. And that's basically --
6 the burden is still there. The only way you can get out of 7
that burden is to pre-decide everything, and that means 8
you're going to have a knee-jerk response of within 2 miles, 9
and in two hours the next 3 riles, and within one hour the 10 next 5 miles, or some sort of great, never-ending action, 11 you know, you take all of it.
12 We feel you can have a knee-jerk response like r- 13 this for 2 dies and it' puts you in a great position to
(. ,
14 deliberately take any further action. It just buys you time.
15 MR. WILLIAMS: Would this then also include your 16 redefinition of those emergency action codes?
18 MR. WILLIAMS: Would that include the potential 19 barrier reductions between fuel cladding to containment?
20 MR. BERNERO: Yes, yes, indeed. And what we're 21 struggling for is the exact right, simple words, so any 22 operator at anytime can say it's yes or no. You know, no 23 interpretation.
24 MR. JAGER: I think what I'm hearing you say is 25 there's going to be a regulation to put all of this into place.
1
101 183 I i MR. BERNERO: Yes. We have a draft regulation.
2 i
MR. JAGER: And within this 2-mile zone there's l
3 going to be an automatic reaction to start evacuation --
4 i
MR. BERNERO: Well, we can 't do that. You could do that. It would be an automatic recommendation.
6 MR. JAGER: We have a reactor under construction 7
where within or closer than 2 miles there's a major ehemical manufacturer --
a MR. BERNERO: On the Tittabawassee River,yes.
MR. JAGER: And an evacuation of that can cause 11 greater public health and ecological damage than a reactor 12 going haywire if it were abandoned. Now, are you going to deal with that in terms of revisiting the basic question of
\' 14 site suitability or -- ?
15 MR. BERNERO: No. What we would do is this. We 16 recognize -- that's a very good case. The Midland ' plant is directly across the Tittabawassee River from the Dow Chemical 18 Company, and to run across there and say evacuate the chemical
, plant might -- as a precaution -- might be more of a hazard 20 for the public than sit tight, as a precautionary -- at least, l
leaving the plant staff tight.
A local interpretation such as that, or I'll 23 postulate another case that the general emergency is declared I 24 at the height of a blizzard. Now, if Three Mile Island was 25 at the height of a blizzard and had another general emergency
102-
. . .- 184 1
I'm sure Tom Gerusky knows you can wash -- you know, the
/^
2 blizzard washes the stuff out of th e sky very effectively.
3 You sit there and make a balance. That's why you have the 4 authority and we don't. You make the, balance and say look, '
5 this is one case where I'm not 91h'g 5 to implemen the i
6 recommendation, it's dumb to tell people in Goldsboro to get ,
7 out of your house and get out on a slippery highway.
8 So what we are looking at -- this is a case where b 9
it's fortunate we lon't have the power. This is an automat;ic. .{!
10 recommendation that is subject to site-specific situations, i
11 specific interpretation by the state and local tuthority. i 12 It could be enlarged, it could be altered. It could be 13 altered to reilect unique ec .2igurations such as nearby
( '
\
14 towns or nriarby industrial facilities or institutions, ad , .
s I, 15 it can be sitered subject to current meteorology.
s I 16 p<
So it's just a standardized recommendation subject
\
17 to such enanges. You must be careful to do that.
18 MR. GERUFKY:
4 ,
In a general emergency now at the 19 plant site, we ebaquite all non-essential personnel. This is 20 really what we're saying. What yoo're saying is thas~n,on-21 essential -- thi peopic vho can , gat out should get dut. Those 22 i who are needed for essential purposes, shCai.d st y\ And the 23 critical people needed to keep the plant 1 ths' Dow Chemical 24 plant safe should remain, while every non-essential pernon s, 25 could leave, and those people will irgve if, inbeed, there is I
% +4 % ,,
a
103 185 I threat to their health. '
,s*
2 MR. BERNERO: That is something that I think we 3
have to defer to the state authority for that judgment.
4 Whether it's cut them down to the guys who keep the chemical 5
reactors or the chlorine tank from blowing, or leave the 6
operating staff in the plant making chlorine or whatever.
7 That I defer to the local authority to decide, whether they a
would go to skeleton crew under those circumstances or leave 9 the shif t alone and go to skeleton crew later. I don't think
- 10 we want to try to make that decision for you. That is very 11 specific to the plant, and that is a unique site.
12 Most f acilitier, ndght have -- like the odd fellows 13 home nearby or some in'stitution, usually with some static
\..
14 population that's hard to move, as an issue. Whereas in that 15
, particular case you're talking about an institution which of 16
.itself is a major energy source, or potential risk source.
17 And I just think it would be better to defer that.
10
, MR. EURE: On your source term, that 200 rem 19 drop-off there, is that with containment breach or -- ?
20 MR. BERNERO: Yes. Typically, you find that the 21 . thing that dominate s the curve or makes that curve is a 22 sequence where the containment f ails before the core melts 23 or a sequence where the containment fails virtually as soon 24 as the core melts through. Early containment frilure, or n
- 25 unmitigated core melt I think is the way I like to say it.
J b! W
104, 186
.v ..
(
1 MR. JERRETT: Lyle Jerrett from Florida. I'm not 2
so sure I see the full benefit of the 2-mile (vacuation.
3 You're saying it's still site and situationel-dependent. So 4
as we face the judge in the aftermath, you still must explain 5
why you did or did not evacuate at 2 miles, so I'm not so sure 6
you've really done anything 7
MR. BERNERO: Well, within -- what I'm trying to do 8 is push toward an automatic or reflexive decision. You just 9
heard some very good reasons why you don't -- of course, we 10 don' t even have the authority to make the decision. It is 11 hard. We're not going to get an easy solution, but I would '
12 hope that we would spare ourselves the confusion and
- g. 13 indecision. ~
Q ,'
14 I remember when the Commission was passing the 15 emergency planning rule in 1980 it hnd prompt notification, 16 15 minutes, and I objected to the rule as a facade. I said 17 yes, you've got 15 minutes so once someone has decided to do 18 some thing, sure we can do it quickly. Bat the decision process 18 isn' t there. We've got En operator trying to say is that core 20 going to melt. We're trying to translate that into general 21 emergency recommendations. We've got too much uncertainty 22 there.
23 I'm trying to put as little uncertainty in that 24 process as possible. We can 't make it automatic; we don't t
' 25 l even have the legal authority to do it.
188
..s I I could be 30 times too low or 200 times too high. And we would r- 2 like to know where we stand in those particular areas before i 8
we come out with such a dichotemous approach, a recommendation' 4
from the government saying something from one viewpoint and i 5
from another -- .- And before it's all over, let's get out 6 :
there and measure it to see whether our estimates were too 7
high or whether they weren't. It doesn't make any difference 8
whether they're too high or whether they weren't. When you 8
take the advanced course all you need to know is the wind 10 .
direction, what's coming out, and evacuate everybody.
11 MR. BERNERO: As far as releasing it to the press, 12 you're preaching to the choir. That draft rule got out of 18 this agency over my objections to the congressional staff.
I' And I can't control what the congressional staff does. It's 15 a very unfortunate thing. I was really upset by it.
16 MR. OTTO: Harry Otto from Delaware. We're 17 spending a great deal cf time sharpening up our triggers here.
18 We're spending also a great deal of time sharpening up our 18 planning tools.
But it seems to me that one area where very 20 little has hapoened is standardization of some of our accident 21 scenarios, which in my experience have tended to be extremely 22 shallow, unrealistic and technically inaccurate.
23 It seems to me that a lot of the scenarios could 24 be easily developed. It could be we're spending a great deal 25 of consulting time in terms of developing the scenarios
7 189 I
that which under present regulation are required through the 2
.f licensing process. 1 3
But in my experience, I think a great deal of 4
ef fort needs to be spent be, tween NRC and possibly FEMA 5
jointly to develop appropriata scenarios that really test our 6
ability to perform both as state agencies, local governments !
+
7 and also, the federal agancies.
8 Right now, we seem to be spending a great deal of 8
time and effort on the part of all agency people and within 10 the states, and I'm not sure that that's entirely fruitful.
11 MR. BERNERO: I referred in passing to the IDCOR 12 effort, and they're doing a lot of work on accident sequences 13 characteristic of dif ferent types of reactors. And one of
.- 14 the things in a separate agenda -- we're doing a lot of work 15 on this, and we have a trial program working with Region I 16 with Tom Murley.
He went up there to Region I recently as 17 the new administrator and he brought this interest with him.
- 18 And that is, from the work we're doing, we can identify for 18 each type of reactor certain accident. sequences that really are characteristic.
! 21 What we're trying to do is develop that understandi ng 22 so that people in our own regional office, the people at the 23 j
plant site, the people in the local authorities, can under-24 l
! stand that and can follow or even practice in a realistic 25 way.
It's a very sorely needed thing because so many of the 1
t
Aug ,
190 ;,
.. " o I i. ,
times now, you know, they pick out a sequence that may not g 2-even be characteristic for that plant, you know. I agree !
3 with you.
4 MR. AAROE:
5 In what form will this information come out? As a NUREG?
6 MR. BERNERO: Oh, the NUREG-0654 will come out 7
through the conventional review process before it becomes 8
a NUREG.
You know, there is a conventional review with the states.
The draft regulation, you know, that has the clean-up ,
10 features I spoke-of, would come out as a proposed regulation 11 and we would circulate it to you people as well as the '
12 l general public for comment before it's adopted. '
I3 .
MR. KERR: Thank you, Bob. I learned one new term I4 today that I'm going to have to figure out how to put in my 15 own vocabulary, and that's legal robustness. That was a new
!* 16 one.
17 Imt me just mention two or three things. Tomorrow 18 we start at 8:30 here in this room. When we adjourn, I would l'
ask you to clear this room aspromptly as possible so that we can move some chairs and get set up for the reception.
21 With that, I'll adjourn the meeting for today.
22 Thank you very much.
23 (Whereupon, at 4 :45 p.m. , the meeting was recessed, 24 to reconvene at 8:30 a.m. the following day, Wednesday, 25 s December 7, 1983.)
t '
gg TECHNICAL CONSULTANTS oN ENER oa'e G Brucenbaugh f 723 Hamaton Avenue-Svete K R<chara B Hubbard San Jose. Cat,fornia 95125 Gregory c. M>nor Phone'(408) 266 2716 30 October 1987 Mr. Donnie H. Grimsley, Director DIEEDOM OF INFORMATION' Division of Rules and Records ACI REQUESI Office of Administration and Resources Management U.S. Nuclear Re Washington, D.h.ulatory 20555 Commission
/' -[ %
_ _ _t h //='d **
RE: Documents Related to NUREG 1082
Dear Mr. Grimsley:
Pursuant to the Freedom of information Act and 10 CFR Part 9, Subpart A, "Freedom of !nformation Act Regulations *, please make availablo at the Commission's Washington, D.C., Public Document Room single copies of records 1/ in the following categones:
A. Copies of all versions of NUREG 1082, Techn/ cal Bases for a Graded Resoonse in Emeroencv Plannina and Preoaredness.
~
I am aware of a final review draft cornpleted in November 1984; this request includes that draft as well as all earlier and subsequent drafts.
B. All records which evaluate, critique, or otherwise comment upon any draft (s) of NUREG 1082.
C. All records which discuss the concept of ' graded response'.
D. All records which document the transmittal of of any of the records identified in items 'A', 'B', and 'C' above to anyone who it not employed by the U.S. Nuclear Regulatory Commission.
If you intend to deny any of these records as 'predecis/onar, you should specify with 3recision to what decision (s) these documents are predecisional to. Blanket assertions o' 'predecisionar exemptions are insufficient to meet your burden under NRC regulations, FOIA case law, and the Freedom of Information Act.
1/ b) of the As used here, Commission's ' records' regulations. has the definition Furthermore, provided "records' are considerein 10 CFR 9.3(d to in the possession of the NRC, its contractors, its subcontractors, or others as provided for in 10 CFR 9.4.
I l
-4wwT927Hjy: ~
t
2 If you or any members of your staff have any questions concerning this request, please contact the undersigned directly by telephone at (408) 266 2716. Your prompt attention to this request will be appreclated.
Sincerey, Steven C. Sholly Associate Consultant i
l L
_c
~
, .so ' ' , . , ,
. b' ! .
$ ===
g C O M Tts5 0[tijt E n[!CD 5 tE.its
.reCus:of negesentatibes 5_
E5tfb!ngt:n, D.C. 20515 5 n=
UI June 1,.1983 g""
==
1ne Eonorable Kunzio Palladino b
C'nairman E hi:ed States Nucle.ar Regulatory Cc issio: "g
%cshington, D. C. 20555 =
U ha: Chair an Palladino: .
=
b As the Nuclear Regulatory Co d.ssion's (NRC) June 9 deadline opproaches for for=ulation of an acceptable emergency preparedness plan for h the Indian Poi =: nuclear plant, it is increasingly obvious that current g regulatory procedures are f ailing to acco=plish thet: legitimate goal of E providing reasonable assurance to the public that their health and saf sty
' vill not be endangered in the event of a radiologieti e=ergency requiring -
=
- otective ac
- ion. Exis
- ing statutes and regula:1 ens of the NRC and the E TCderal !=ergency P.anagement Agency (TDM) have inadve.rtently givan 5' er. tensive leverage to States and localities who have reservations about
' ' E tucitar f acilities which are' unrelated to e=e:gency planning issues.
As E o :esult, States and localities which choose not to participate in the E a=ergency plan:ing proc E is suan c e of nev lic ens se,es s , f ocause or even wh'atever epera reasons, ing pla. zay beshut
- s to abledern.
to, stopTnis the 5 d.s met consis t en: vith the legitinate goal of nuclear e ergency preptredness y and va.s, not the intent of Congress when draf ting the enabling legislation. g
=
Ve in Congress have no: alvays ren: clear signals to the Government'. E agencies i:velvei and are,in pt:t, responsible f or the current situation. .
E
- is possible tha additional legislation may be required to devise a E sensible long ter= solution to the developing situation. Bovever , in the E chert term, we urge the NRC along with TDR to quickly reviev and revise El those inter'nal regulations which are being si'sused by some local of ficials. El Suspension of cperating' licenses or denial of nev licenses vill cause [i cdverse and unnecessary econo =.ic consequences to utilities and their i ratepayers. Ve are confiden: that these consequences can be avoided I' th:cush fley.ibili:y en the part of the NRC and TDM, coupled v'.th a l continued resolu: ion towards working with all parties in devising odequate emergency plans.. Tnis is consis tent with Congress' intent that
!l j) ulti=ately every nuclear power plant vill have an approved nucita: emergency ;
plan. .
- l j \, qS ,
i 4godj) 1 ft i.
/ f Q
~
b// U
1 e-1 6
- 9. . ., = '
- '.*ne Eo'.e cM e 1:ar.:io Pai.edino
, June 1, :.fi3 0.__
t=
E if ;
5* 1 k's appreciate your continued attection to these r.agters. g
=l ,
. . E' E
Sincerely, E I
E a
F &
L.,, u fr w a r - y C
F.a fnef5jan\ Jr. J e ts T. Broyhill
[,
. 'fn ' g F.inority Me=be: Panhing F.ineri:y Me=ber Co _ tee on Inte:1o: and Co=ittee on Energy and 3 l
Insular Affaits Co=e:ce 5 '
~. Q
'- , t 4 iL k
Carlos . F.oorhead j Ranking F.inori:y Menber . ;
Subco=ittee en Energy Conservation ;
ud Power i
.- - - Co=ittee on Inergy and Ccmerce I s
- 4 t
e
- I r
4
.. I I
e e
e 9
e
l ,
Qoopg ,
4 J,, UNITED STA1 Es
! ?., , 'a NUCLEAR REGULATORY COMMISSION
. E .'e j w AssiNoTON, D, C. 20665 ' h .W k..u... / APR 111985
.) Ln 2
~
y Harold R. Denton, Director
!!EMORANDUM FOR:
Office of Nuclear Reactor Regulation 3 h Darrell G. Eisenhut, Deputy Director h Office of Nuclear Reactor Pegulation g James M. Taylor, Director Office of Inspection & Enforcement 6 /
Edward L. Jordan, Director 7 Division of Emergency Preparedness and Engineering Response 6 N Robert M. Bernero, Director Divisien of Systems Integration E.ry]*
Thenis P. Speis, Director Division of Safety Technology Office of Nuclear Reactor Pegulation Joseph F. Scinto, Deputy Director of the Hearing Division Office of the Executive Legal Director h FFN:: Frank H. Rowsome, Assistant Director i for Technology Division of Safety Technology Office of Nuclear Reactor Regulation
SUBJECT:
EMERGENCY PREPAREDNESS There are a number of reasons to look at refom of the troublesome emergency l preparedness (EP) regulations that de not depend upon source-term reductions.
l In my view, the case shapes up as follows:
l
- 1. The emergency preparedness rules and implementation unter the NRC/ FEMA s agreement of NJREG-0654 have failed to reduce offsite radiological risk.
l This was first documented in the Irdian Point hearing. Many of the redundant ree. sons for this conclusion are generic. Reexaninetion of the beses for this conclusion suggests that it can be extended to all plants. The technical basis for this argument is sketched in the i
'ttachment. As a result, emergency preparedness fails as a final layer l
een p e
~ ~
- APR 111985 of defense-in-depth; it does not enhance the safety of the public arcurd nuclear power plants.
- 2. The emergency preperedness regulations are fatally flawed as exPMples of sound regulation. They place responsibility with those who have no authority insofar es they mandate participation by offsite state and local governments. It offers a tempting legal gambit for local authorities to balk in the face of rate shock, something it was never intended to do. The "graded emergency response" proposals initially developed by ASTP0 a year ago (and advocated by IDCOR today) do nothing to cure this vicious "Catch 22" aspect.
- 3. The vagueness as well as the "Catch 22" features of the regulations constitute a fertilizer for litigation, delay, controversy, and bad public relations that has a legal nexus - thanks to the regulations -
but no technical relevance to public health and safety. The result unnecessarily involves the courts, attenuates NRC control, and damages the irstitutional fabric of reactor safety regulation, as well as costing massive anounts of money.
- 4. Even if emergency preperedness could eliminate early casualties, as projected with WASH-1400 source tertns, the value of doing so would be nany orders of magnitude less than the costs of compliance, under the
. regulatory analysic conventions employed in generic standards development, backfit policy, in the proposed safety goals, or those suggested in the Indian Point Hearings on risk. Since we intend to use the same conventions in our effort to scrub the rules of unproductive feetures, emer9ency preparedness is a natural application of this '
b policy.
Note that none of these arguments that the emergency preparedness :
regulations terms.
warrant overhaul depend upon reductions in W.SH-1400 standards development, the reductions would constitute a fifth reeson for refern of the EP regulatfors.
I see a route out of this dilemme that might be called the "Good Neighbor" policy. Since evacuation planning is virtually worthless as a risk reouction tactic I propose that we abandon the EP rules in favor of a recoenition that nuclear power plants can and rnust be made safe enough to be "good neighbors" in the absence of offsite emergency planning. The new rule would not require any state or local participation in emergency planning or drills, although these local authorities would be kept infomed. The rule would pair the abandonment of NUREG-0654 planning with the imposition of the further severe accident safety analysis, called for in the draft Severe Accident Policy, to debug 'the operating plants of outliers, and with the We accident ma core damage into the rerlm of managing core melt accidents on site.
could truly portray the new rule as providing better limitation of offsite
l- 3- AF.R31?9CI radiological risk than the current regulations do, thus bluntirg the argument that we appear to be relaxing the rules. The attachnent sketches some alternatives for the new rule.
I hope you give this proposal serious consideration and pass it on - as food for thought - to Vic Stello and Bill Dircks.
k __
Frank H. Rowsome, Assistant Director for Technology Division of Safety Technology Office of Nuclear Reactor Regulation
Attachment:
Risk Reduction fron Emergency Preparedness cc: F. Gillespie S. Schwartz A. Thadani D. Muller W. Minners P. Jamgethian S. Acharya b
i.
l 1
I e
9
4 AITACHMENT RISK REDUCTION FR0ft EMERGENCY PREPARED!!ESS Realistic analyses of the consequences of severe radiological releases from nuclear power plants have always shown that the nunter of latent casualties are largely unaffected by the speed or tining of evacuation or relocation.
Thus, it is widely acknowledged that short-tem emergency response is notivated by the desire to limit early casualties. I will not go into the reasons why short-tem emergency response does not influence latent casualties here, but the reasons are well known.
In the Indian Point hearings on risk, a number of pieces of evidence demonstrated that energency planning - as currently conceived - has little or no effect on early casualty projections.
- 1. First, FRA always employs data in preference to judgment in assessing the reliability or perfomance of safety functions. There is some data h
on the influence of emergency planning upon'the effectiveness of emergency response. It shows that unplanned evacuations proceed as rapidly as planned evacuations. The data originate in chemical spills, floods, stoms, and other non-nuclear contexts. The data are not so extensive as one might prefer. Nonetheless, the message of the data is clear: spontaneous, unplanned evacuations work very well. The historical record supports the conclusion that once a hazard is l recognized and a decision to evacuate is rade, state and local authorities can carry it out quickly and effectively without elaborate prior planning. This, alone, would have sufficed to support e PRA assessment that energency planning doesn't make a difference to risk.
Rete that this conclusion is generic; it is not limited to Indian Point.
However, the testimony went on to assess the effect of emergency l
response variables on risk.
i l
I
i
= ;
e
- 2. For each releese category event, it was found that early casvelties are no more numerous when evacuation fails than they are when evacuation l works as planned. The default scenario, developed to model cases in l i
which anticipatory evacuation fails to take place, was defined as >
relocation from hot spots of offsite radiological contemination eight to twelve hours after plure passage. The "successful" evacuation scenarios with which it was compared assumed e delay time, which was a function of t
the in-plant accident scenario, followed by anticipatory evacuation at a pace chosen to mateb the evacuation time estimates developed by specialist f consultants on the speed of evacuation. The surprising result that anticipetory evacuation works no better than relocation was traced to its ;
The origins in the calculation of reactor accident consequences.
results of this detective work reveal why we obtained this surprising result. Accidents that take long enough to evolve into a release to enable the people to get out ahead of the plume would not have yielded early casualties even if the people had steyed home. Slowly developing accidents simply pose negligible threats of early casualties. Note that WASH-1400 source tems were used in this assessment. For the rapidly developing accidents that give rise to virtually all the projected early b casualties, many evacuees fail to best the plune. Houses and work t Therefore pieces tend to provide better shielding than cars or busses.
the attendant risks of exposing evacuees to the plume while in transit eenerally compensated for any advantage to those escaping the plume or Some specifics of getting farther away before the plume reached them.
this analysis are site dependent. For example, the expert estimates of evecuation rates for the Indian point site may be longer than those for some other sites.
However, the basic finding is generic to any appli-accidents that develop slowly enough
. cation of WASH-1400 source terrs:
for evacuation to clear the EPZ ahead of the plume do not yield appre-ciable early casualties in any event. The attendant risks of exposure 1n-transit tend to counterbalance the berefits of anticipstory evacu<
a' tion for the rapidly developing accidents that give rise to the bulk of anticipatory the early c,asualties. The conclusion is inescapable:
t 9
evacuation carrot be expected to lower early casualties reliably or effectively fron severe reacter accidents. Since this conclusion applies to the full spectrun of WASP-1400 release categories, it can be expected to remain true if source tems increase a little er decrease a good deal.
- 3. At Indian Foint, the bulk of the risk was attributed by the staff to accidents triggered by earthquakes or hurricanes. Initiators of this kind will only trigger reactor releases if they are of extraordinary severity. They can be expected to preclude ground transportation and interfere with comunication throughout the EPZ. Many people might also be deprived of shelter by such an initiator. This is a plant-specific conclusion which is probably not unique to Indian Point, but is far from l a universel pattern among nuclear plants.
PRA has been used as a basis for the development of the emergency prepared-For example, NUREG-0396 employed sensitivity l ness rules and FUREG 0654 studies en the WASH-1400 consequence model to provide e technical basis for Therefore, it is the developnent of energency preparedness guidelines.
., useful to exanine why such insights as those developed at the Indian Point e
Hearing did not surface earlier.
The histord eel data suggesting that unplanned evacuations work as well as planned evacuetions has been available since the early 70's and is con-sistent with U.S. experience since then. This infonnation is well knosm in the pRA comunity. It appears to have had little influence on the development of the rules and regulations, though.
The sensitivity studies in NUREG-0396 and temparable, more recent work are They show, as ene might expect, that techni,cally correct but misleading. They tend to fast ev)cuation yields lower casualties than slow evacuation.
assume that the alternative to evacuation is no evacuetion at all or very slow evacuation.
They have not used relocation as the default cese for m
9
4 those situations in which anticipatory evacuation fails and have not compared the potential to improve relocation times compared with evacuation rates.
The accident at Three Mile Island demonstrates the viability of relocation as Within hours of the time that THI was recognized to the "def ault" scenario.
have the potential to be radiologically. severe, the environs of the plant were crawling vith health physicists renitoring the environment. They came
' from the licensee, fron the NRC, DOE, EPA, FEMA, and the Defense Department, It can be assumed that if there had been hot spots of residual i I an told.
4 ground contaminatien in the surrounding counties after a severe release, these het spots would have been mapped and the residents relocated within a
- ew hours of plume passage. Thus, the mnst accurate model of pre.1980 er.ergency response and a good "default" model to portray the failure of evacuation plans today, is to assume that the population is relocated fron l
i hot spots of residual contanination within a few hours of plume passage.
When such models of relocatien after plume passage are compared with the spectrum of evacuation models, it is found that the default cause yields
' fewer early casualties than the models of slow or no anticipatory evacuation.
I They yield roughly the same casualty estimates as do the models of antici-patory evacuation with realistic evacuation rates, as the Indian Point record l
suggests. Very fast anticipatory evacuation looks better.than relocation.
l Powever, fast relocation looks nearly as good as fast evacuation, and the l
difference in the logistical problems between the two approaches to emergency l
response may well indicate that the speed of selective relocation is more f The singular advantage
' perfectible than the speed of bulk population movement.
held by anticipatory evacuation ever relocation is that it supports an ALARA l In the sinwly develeping accidents that would not principle for exposures.
tause early casualties in any event, successful anticipatory evacuatien does l
evoid modest doses for those beating the plune.
l I
The logistics of relocation after plume passage are quite different fror l
those of anticipatory evacuation. In anticipatory e.vacuation large numbers l ir relocation, cf people froe broad areas are moved considerable distances.
l 1
I t
- i 5-only the small numbers of people in het spots of residual contamination warrant prompt relocation. This entails moving only small numbers of people ,
High levels of residual contam-short distances in order to be successful.
ination - at a level that could threaten early casualties in a few days of exposure - cannot cover much territory. If the contamination were more dilute, prompt relocation would not be necessary to avert early cesuelties.
Relocation presunes that the residual contamination has been measured and mapped within hours of plume passage. Thus it must be guided by a coord-inated group of bealth physicists who infom state and local authorities on what areas to reloc)te first, which second, and so forth. Should one wish to mandate preplanning for speedy relocation after plume passage, the essential element would be to coordinate the roles of licensee and federal health physicists in mapping contamination and prioritizing locales for relocation, since speedy diagnosis of the situation requires a high level of team work anong these federal and utility personnel. As both a legal and practical matter, state and local authorities should physically conduct the l
relocation, as planned by the health physicist team. In light of the good record of unplanned evacuations, the only essential element of state and local perticipation in reactor emergency preparedness is a cle,;r understanding t: of the source from whom they can expect to cet authoritative information on the hazard and relocation priorities. The utility and federal health physicists (PP) who map the contaminetion and set relocation priorities need more knowledge of severe reactor accident risk than is comon among certified fps. If their effort to prioritize a relocation sequence is to be most efficient, they need to be able to project accident sequences, source tems, understand the effect of weather upon consecuences and appreciate the thresholds of contamination, to be expected of accident releases, at which
- early casualties might be expected. They should also be familiar with the shielding factors to be expected of structures and vehicles. Puch of this information is not well known to either utility or federal HPs. Therefore, if we mean to take emergency preparedness seriously as a risk reduction tactic', whether in its current form or in modified fom, there is much to be done to accomplish the relevant training.
e
-6 The foregoing assessmert of emergency planning for relocation presures that the rish warrants some planning. The avidence suggests otherwise. The prefected number of early casualties from reactor accidents are always dwarfed by the projected number of latent casualties, with any plausible source-term model, not to mention the preponderance of cffsite and onsite If a plant property danage in the risk profile of nuclear power plants.
poses an acceptable risk with respect to property demage and latent casualties, a case can be made that early casualties will be a non-problem, All of the ways of conducting even with WASH-la00 source tems.
benefit / cost analysis that have been suggested would indicate that the benefit of even something as radical as the total elimination of early resualties from the risk profile of a nuclear power plant would be worth very small expenditures. An assessinent for Indian Point suggested that the costs of actually conducting evacuations, even if they took place only for genuine core melt accidents, would be larger than their value, not to mention additione.1 costs for preplanning. I conclude that if a plent has been adequately debugged of proninert severe accident vulnerebilities to be acceptable with respect to property damage risks, then the early casualty risk will be acceptable with ample margin. The plan suggested by the draft b
T Severe Accident Policy calls for further decision making on the adequacy of containment perfomance of reference plants and further severe accident safety analysis for each operating plant in order to meet two objectives:
to verify that the generic decision rade on the reference plant applies, and If this is done, to debug the operating plants of outlier vulnerabilities.
a convincing case can be made, I believe, that nothing further eeeds to be done to limit early casualty risk.
It is worth noting here that the proposed Comission safety goals are ill suited to make this case. One characteristic of the safety goals in FUREG-0880, Rev.1, is that they are effectively very stringent for early fateli. ties and very lenient for latent casualties. Therefore, a plant can trip the early fatality guideline and pass the cancer fatality guideline,
I l
even though the projected latent casur.1 ties per unit year outnumber the This remains true of the projected early casualties by thousands to one.
safety goal recomendations of the staff Safety Goal Evaluation Steering Comittee.
In addition, the algorithm for assigning a value to risk reduction for use in benefit / cost analysis does not value the reduction of The stringent early early casualties apart fron societal dose reduction.
fatality guideline will suggest that enough has been done without crediting On the other hand, the perspective on the comperative energency planning.
importance of early and latent casualties ano a more discriminating valuation of worth of early casualty risk reduction wil) not be found in the safety geals. As a result, the safety goals will be of much less value in putting emergency preparedness issues into perspective than one might hope.
Thus, it may be desirable to make the safety goals more conprehensive and discriminating. :
A. case that current energency planning accomplishes no risk reduction could, be made today. It would require a little research to flesh out the documen-tatier, but it would require no assumptions about source-term changes or the ultimate implementation of forthcoming policy initittives such as those on b It would support dropping the emergency i
severe accidents or safety goals.
preparedness rule, or replacing evacuation planning with federal-licensee relocation planning, and so eliminate the need for state or local participa-tion in planning and drills. That would solve the ' Catch 22' problems in the existing regulatiers.
The abandonment of energency plenning, beyerd the imposition of a requirement for onsite accident management planning, is, I think, the preferable answer.
- It more accurately reflects the realities of severe accident risk and deals However, coupling with all of the defects of the current EP regulations.
EP refom with the debugging of operating reactors of their outlier vulner-Such a rulemaking abili. ties in a
- good neighoor" policy would take longer.
would require that our thinking be more mature than it is now about severe '
accident management and the further severe accident safety analysis to debug
- operating reactors. Still, these are tractable problems. Such a rule could 1 recomend it.
take shape in about e year or less if we put our rind to it.