ML20080B787
| ML20080B787 | |
| Person / Time | |
|---|---|
| Site: | Indian Point  |
| Issue date: | 02/06/1984 |
| From: | Weiss E UNION OF CONCERNED SCIENTISTS |
| To: | NRC COMMISSION (OCM) |
| References | |
| ISSUANCES-SP, NUDOCS 8402070267 | |
| Download: ML20080B787 (25) | |
Text
.
00CMETED USNRC UNITED STATES OF AMERICA 'g4 Eg _6 p3:j g
. NUCLEAR REGULATORY COMMISSION BEFORE THE COMMISSION
)
In the Matter of )
)
CONSOLIDATED EDIS0N COMPANY OF NEW YORK, INC. ) Docket Nos. 50-247-SP (Indian Point Unit 2) ) 50-286-SP
)
POWER AUTHORITY OF THE STATE OF NEW YORK )
(Indian Point Unit 3) ) 6 February 1984
)
INTERVENORS'COMMENTSONLICENSINGBOARb RECOMMENDATIONS ON INDIAN POINT UNITS 2 AND 3 Introduction The Union of Concerned Scientists (UCS) has neither the inclination nor the ability to retrace in detail the volumes of evidence in the record. Our taste for such a task disappeared when the Commission voted in June,1983, to allow the Indian Point plants to operate despite the plain fact that adequate emergency planning was not in place. When a majority of the Commission voted to authorize continued operation on the basis of more unfulfilled promises, we were forced to conclude that neither the facts nor the regul a tory
" requirements" is of paramount importance wnen the operation of Indian Point is at stake.
The UCS petition which began this proceeding was filed some six months after the accident at Three Mile Island Unit 2. Its basic premise was that the THI-2 accident had shown that serious accidents at nuclear plants were possible (contrary to the AEC and NRC's previously held official positions),
8402070267 840206 PDR ADOCK 05000247 C PDR -'
h
t .
. 3 and that while the probability of such accidents was essentially unknowable, the consequences of a serious accident at the Indian Point site would be intolerable. -
Four years later, and following a remarkably wasteful and frustrating diversion through the haze of probabilistic risk assessment, a diversion required by the Commission as a precondition to participation by the intervenorsN, the conclusions reached by the Board are at the bottom line the same ones that are the premise of UCS's petition; that is: the probability of an accident at Indian Point cannot be shown to be different from any other plant, but the consequences would be much greater.
The tragedy of this case is that the Board devoted enormous attention and effort to trying to choose the "best" probabilistic risk assessment (PRA) numbers and in so doing completely avoided the threshold question, squarely presented by the intervenors, of whether any probability numbers are sufficiently precise to be a useful objective indicator of overall risk. It is UCS's posicion, in which we find support from authorities including members of the NRC's Advisory Committee on Reactor Safeguards, that a state-of-the-art PRA generates probability estimates with uncertainties so great that, while they provide an illusion of precision by virtue of being stated in l
quantitative terms, they are not an reliable basis for decision-making. The Board never fairly addressed this. Instead, it seems to have fel t itsel f compelled to select some numbers and since only the staff and licensees presented numbers, it chose the staff's (which were certainly "better" than the licensees').
I
'-1/
The Commission's heavy-handed intervention caused the Licensing Board Chairman to resign in protest. The new Chairman selected by the l Commission wrote a dissenting opinion to the Board's Recommendations which rais.es serious questions regarding his understanding of the majority opinion and the evidence in this case. -
i
- f. .
r The intervenors did not play the PRA game. As a practical matter, we were barred from the table due to inability to ante up the large su.ns of money
~
- - 'to perform a PRA.' For example, the NRC's recentif-published "PRA - PFoc~edures Guide" (NUREG/CR-2300, January 1983) estimates that it would require a 29-member team a year to perform a complete PRA; personnel costs alone could approach $2-$3 million. Furthennore, this estimate fails to consider the necessity of access to substantial computer resources and detailed infonnation on plant design, operating procedures, and plant logs (access to which we can scarcely imagine being granted by the licensees and their consultants).
However, even if we had access to that kind of money, playing the PRA game would be antithetical to our fundamental position that PRA results are not reliable and should not be used to determine overall risk. The results were sadly predictable: the Board chose between the numbers presented by those who did produce PRA results and intervenors were frozen out without any serious consideration of our basic position. The Board cannot be assigned all the fault for this; it was the Commission which forced this proceeding into a showcase for probabilistic risk assessment.
The fact is that while a life-threatening accident at Indian Foint is a
" low probability" event, no one knows how likely it is to occur. Many previcus attempts to estimate probabilities, while soothing at the time, have not been of notewortny accuracy. Most recently, one can point to the two Salem failures to scram -- an event which wc; confidently said to have a vanishing 1p mdttiscule probability of occurrence.
On the ct.er hand, we do know that if a substantial release occurred at Indian Point, the consequences could be catastrophic in terms of death, illness and property loss. Intervenors also believe that it is a
self-delusion to imagine that the emcrgency response would be adequate in the New York area at the best of times. At the worst of times -- for example, during a winter storm ---it-would be-a nightmare. - -
We are well aware that asking the NRC to close Indian Point for these reasons runs against the grain of 25 years of a system which was premised upon the belief that serious accidents could not happen. As TMI-2 fades in the institutional memory (there is only one remaining Commissioner who remembers the utterly bewildered reaction by the Commissioners to Governor Thornburgh's pleas for advice on whether he should order an evacuation), so do our hopes for a genuine reconsideration of the wisdom of allowing reactors to operate 30 miles from New York City, with 17 million people within a 50-mile radius.
Nonetheless, UCS, along with other Intervenors, has decided to attempt to put this proceeding into perspective, if only for reasons of historical accuracy. Our comments on the risk issues follow, in which we are joined by the undersigned Intervenors and interested States.
A Perspective on the Board's Risk Conclusions (i.e., Understanding Them and Contrasting -Them With Earlier Predictions)
Section D of the Commission's Order of May 30, 1980, directed the General Counsel and the Office of Policy Evaluation to establish a task force to prepare a report to the Commission based on information available at the time so that the Commission could determine whether the plants should be permitted to operate during the pendency of the proceeding. The " Task Force on Interim Operation of Indian Point" (composed of the NRC staf f members), sent its report to the Commission via SECY-80-283 (June 12, 1980). It was later published as NUREG-0715.
e The centerpiece of this effort was a " quick-and-dirty" probabilistic risk assessment of Indian Point based on a brief review of the designs of the
= .
-Indian ~ Point ~ reactors compared with " insights" gained from- the Reactor Safety - -
Study, the Reactor Safety Study Methodology Applications Program,' and the Interim Reliability Evaluation Program. The Commission based its decision allowing interim operation on the Task Force report and the Director's Decision setting out some short term plant modifications agreed to by Con Ed and PASNY. Consolidated Edison Co. of New York and Power Authority of the State of New York (Indian Point Units 2 and 3), 00-80-5,11 NRC 351 (1980).
It should be noted that the Task Force report calculated that the plant improvements agreed to by the licensees and incorporated in the Director's Decision had a negligible effect on risk. Surprisingly, during the proceeding neither the NRC staff nor the licensees were able to quantify the impact on risk of the measures implemented by the Director's Decision. It is obvious, in retrospect, that the Director' Decision provided little in the way of a basis for continued operation.
The results of the Task Force report included annualized frequency estimates for core mel t, early fatali ties , early injuries, latent cancer fatalities, and offsite property damage. These results can be contrasted with I
the results obtained by the Board from their analysis of the record. In order to do this, we have summed the results for Indian Point Units 2 and 3 from l
l SECY-80-283 (which consists of dcabling the values since they were predicted to be the same for both units). In addition, we have applied a Gross National Product (GNP) escalator to the SECY-80-283 offsite property damage results to change them from 1974 dollars to 1982 dollars to make them roughly comparable with the ASLB's figures. In making this comparison, we assume for the purposes f o_f illustration that the Board's analysis of the record is correct.
The comparison follows:
CONSEQUENCE SECY-89-283, 1980 ASLB DECISION, 1983 DIFFERENCE Core Melt Frequency 2.0 x 10-5 7.0 x 10-4 35
- +~-
Early Fatalities -4.4 x 10-4 1.8 x '0-4 to ~ 0,4 -
3.8 x 10-4 0. 9 ' ,
Early Injuries 5.4 x 10-4 1.6 x 10-1 to 296 2.4 x 10-1 444 Later.t Cancer 4.8 x 10-4 2.6 x 10-I to 542 Fatalities 2.9 x 10-1 604 Offsite Property 1.5 x 10+3 4.1 x 10+5 to 273 Damage ($1982) 4.5 x 10+5 300 Thus, if we accept for the sake if argument the results in both SECV-80-283 and the ASLB recommendations, we now "know" that a core melt accident at Indian Point is roughly 35 times more likely than the Commission was originally informed, early fatalities resulting from accidents range from about the same likelihood to about half what the Commission was originally informed, early injuries are roughly 300 to 450 times more likely, latent cancer fatalities are roughly 550 to 600 times more likely, and offsite 2
property damage is roughly 300 times larger on an annualized basis!_/
This entire exercise indicates the following: (a) both the probability and consequences of an accident at Indian Point are much grea ter than the Commission was led to believe in 1980, and/or (b) one can have no confidence
-2/ We wonder what the probability is that a 5. cent washer would disable the electrical system at Indian Point 3, causing a seven-month outage for repairs at a cost of $8 million for parts and labor and $150 million for replacement power. ' J. Cra.ig, "5-Cent Washer Blamed for Nuke Reactor Damages," Reporter Dispatch, Gannett, Westchester, January 31, 1984.
in the accuracy of the numbers generated by risk assessments; they are almost laughably imprecise. It must also be kept in mind that the numbers presented w_.--- above are "best' estimates" and-do not' include' the very-large uncertainty -bands - -- -- -
that surround them. ',
When one considers the range of risk values found between the licensees' lower bound estimate of risk and the Board's upper bound estimate ("best estimate" values multiplied by an .NRC staff witness' intuitive guess on uncertainty bounds),3_,/ the numbers literally span more than five orders of magnitude. If what we "know", based on the PRA effort expended on the Indian
! Point proceeding, is that the probability of a serious accident (i.e., one causing large numbers of fatalities) eis somewhere between one chance in a few thousand and one chance in several .hundred million, we do not really "know" anything at all. .-
Ccmparisons of Indian Point Risks with Other Risks 1
The NRC staff compared its risk estimates with existing risks from a variety of causes. The value of thisi comparison is open to question,-even if j the results are academically appealing. Perhaps a more relevant risk question would be a comparison of the NRC staff's risk estimates with those estimated to result from a different type of ? electric generating facility at the same
( site (i.e.. coal, oil, or natural gas') or conservation to replace Indian Point. Even this would be of quesitionable Jalue due to the uncertainties 1
involved, the differences in the typsls of impact (i.e., chronic impacts versus low probability /high consequence impacts), and the lack of a technically viable assessment methodology.
l 3_/ Rowsome, ff. tr. 8778, p. IV.C-19; Recommendation (, pp. I'01-102.
t
In addition, despite its clear irrelevance to the issues in this proceeding, Dr. Bernard Cohen's testimony was admitted by the Board over the
' - ~
f nterventors' ~ obj ections.4/ Dr. Cohen expressed risk in terms of loss of life expectancy for the average person living "near" the plant, spending a life time "in that area." Neither "near" nor " area" were defined by Dr. Cohen.5./
The Board was also " unable to confirm" Cohen's figures from the sources he gave and found other discrepancies in his testimony.5./ Dr. Cohen compared Indian Point to risks from such things as cars, alcohol, stroke, etc.
The overriding objection to Dr. Cohen's testimony, in addition to his looseness with the facts, is philosophical. Comparing the risk of nuclear accidents to all individualized risk over a lifetime completely obscures the risk that makes Indian Point unique and occasions this proceeding. The societal risk inherent in the fact that an accident at Indian Point would kill and injure far more people than if it occurred anywhere else makes his comparisons grossly incomplete and misleading.
Moreover, Cohen assumes that if an equivalent amount of money as that -
needed to reduce the Indian Point risk would have a greater impact -on life expectancy by being devoted to other risks (such as reducing auto accidents),
one should for that ..ason not reduce the risk at Indian Point. Based on that logic, NRC could site a plant in the middle of Central Park without a containment building. Cohen's testimony does little beyond illustrating the self-evident fact that the high consequence accidents of concern here are of relatively low probability.
~~
4/ Id. at 344.
5/ Id. at 347.
6/ Id.
On a philosophical level, one might reasonably question whether either set of values presented in the preceding section are useful as they are
~
_. -expressed -- -i .e.+,~1n-annualized -terms. ' -Consequences - from'the operation - of = - - --
nuclear power plants are not manifested in the same way as conseque'nces from the operation of other types of generating facilities.
For? instance, a coal-fired plant would produce a chronic source of air pollutants (e.g., sulfur dioxides, etc.), the magni'.ude of which would depend upon the efficacy of- pollution controls and the level of contaminants in the input fuel . On the other hand, a nuclear power plant might operate for its Jentire lifetime without significant incident, or it might suffer a catastrophic accident at any time which can cause many thousands of early and ifelayed deaths and injuries, and substantial offsite property lossas (in the ter$s of billions of 'dollaFs or more). Such consequence levels are enonnously greater than any hypothetical coal-related accident consequences.
A good il10stration~of this dilemma is provided by the Bcard's estimates o f' the cumulative societal risks posed by operation of the Indian Point reactors until about the year 2006. The Board estimates that, under- varying assumptions , from 1.1 to 2.4 persons might die from "early fatali ties" assuming no supportive treatment, and that 0.45 to 0.91 persons might die from "early fatalities" assumirig~ supportive treatment.7_/
These results are statistically elegant but totally misleading. Unless there i s an accident with a substantial fission product release during the lifetime of the Indian Point reactors, there will be no "early fatalities" --
i.e., no one will receive whole body doses in the range of 300-600 Rem and die of radiation sickness with'ir. 60-90 days. On the other hand, if there is such
- 7/ Id., Table VIII, p. 94.
h A
e - - -~ , - , - - - . -, . .
v a large release, it -is far ,more likely that there will be hun'dreds to thousands of early. fatal _ities than there will be only one or two. In sum,
. _% n ~
~__-_.--there will either be no early fatalities at all- (in which case -the population surrounding Indian Point will be eternally grateful for their good fo'rtune) or there will be a disaster involving hundreds to thousands of early deaths and similar numbers of latent cancer fatalities. Expressing the fatality risk on an ~ arD1ualized, per-year basis totally obscures the nature of this risk.
s
- Similar comments could be made concerning other relevant impacts such as the financi i consequences of accidents.
If f the accident probabilities and release magnitudes posed by Indian Point could be transferred to anywhere in the country, the' risk posed to a hypothetical individual located "near" the plant wo'uld be identical -- i.e.,
the individual risk would not change. The individual- risk issue is a red herring in this proceeding. This case is of importance only because of the very large societal risks associated with an accident at Indian Point.
. While PRA may, if one believes its supporters, provide some " perspective" on the' estimated likelihood of a catastrophic accident, PRA is- utterly incapable of providing an answer to the question: Will such an accident occur during the remaining 26 years that one of the Indian Point reactors will be in operation? One might just as easily (and at less cost) rely on Las Vegas oddsmakers as on a PRA for an answer.
None of this is surprising (nor should it be). The answer to a shutdown / continued operation question does not magically fall out of a PRA.
The Commission is thus, not unexpectedly, at roughly the same point it was nearly four years ago. The same bottomline question remains: Is the risk of a catastrophic accident at Indian T aint' worth the benefits of having the reactors continue to operate for the remainder of their useful lifetimes? Our answer is still "no."
e
f A related consideration is that, given a large release" of radioactivity at the Indian Point site, the conditional likelihood of having a thousand i---deaths - is - nearlyi-identical - to the-conditional -likelihood -of ten -thousand 1
' ~
deaths, especially when one considers the uncertainties inheren't in the analysis. Indeed, one could justifiably conclude that there is no statistically significant difference in probability between the two outcomes for densely populatea sites.
The Board concedes as much in its brief discussion of this problem,0I - but then sidesteps the issue and argues that such considerations argue for the implementation of their proposed fixes. These fixes consist of: (a) a " Safety Assurance Program" which is necessary to ensure that , the risk estimates presented "come true", (b) a requirement -for " appropriate action" if a tornado watch or tornado warning is issued for the Indian Point area, and (c) installation of a loose parts monitoring system at Indian Point Unit 3.
We do not quarrel with the concept of a " Safety Assurance Program";
indeed, it makes good sense to implement such a program. In fact, it is more than a li ttl e surprising that such a program does not already exist.
Similarly, taking " appropriate action" for tornado watches and warnings, and installing a loose parts monitoring system also make good sense. However, it is clear that none of these measures address the problems posed by the location of Indian Point with respect to a large population center.
Further, none of these measures were shown on this record to contribute significantly to a reduction of risk for the Indian Point plants. The lack of a " Safety Assurance Program" might itself invalidate the PRA. There is little evidence in the record of the efficacy of precautionary shutdowns (we presume 8/ Id. at 103-105.
this is the thrust of the Board's recommendations on tor 6adoes)S! in terms of reducing risk, nor the accuracy of tornado predictions. There is also no evidence that the loose parts monitoring system- at- Indian- Point Unit 2 contributes significantly to reducing risk, and there is nothing unique about Unit 3 which suggests that such a system would make much difference there. In addition, the loose parts monitoring system addresses primarily steam generator tube rupture accidents which were not found to be risk dominant for the Indian Point reactors. We conclude that these measures, although desirable, have not been demonstrated to significantly reduce the risk posed by the operation of Indian Point Units 2 and 3. In that regard, they are similar to the short-term actions required in the Director's Decision in 1980 which likewise were sho.en to have little apparent effect on reducing risk.
Indian Point Compared With An " Average" Site Returning to the SECY-80-283 evaluation urging interim operation of the plants, the Staff .therein concluded that the site is about an order of magnitude more risky than a typical site and the design is an equivalent degree less risky, so that these two factors cancel out and Indian Point is about as risky as a typical reactor at a typical site.$I When we revisit the
! NRC Staff's testimony in the hearing, the following picture emerges.
NRC witnesses Rowsome and Blond testified that there are about ten times l
as many people at the Indian Point site as there are at an average site --
the same as the first part of the conclusion from SECY-80-283. However, 9/ Recommendations, pp. 141-143.
p/ SECY-80-283, p. 38.
g / Rowsome and Blond, ff. tr. 12834, at 28 and B-6.
5 . .
. .' . .n z , . . .. ..
Rowsome and Blond also testified that the Indian Point reactors are roughly ~
average with respect to the estimated frequency of severe releases of radioactivity 12/ -- this is significantly different from the second part of the conclusion of SECY-80-283. Nonetheless, Rowsome and Blond conclude that the societal rists posed by Indian Point are "probably average to above C average"comparedtoothersites.13/
4 How can this be the case if tiie frequency of severe releases is roughly average and the population is ten times larger? Obviously, the NRC staff's overall conclusion conflicts with the two subsidiary conclusions upon which it rests. UCS addressed this in our proposed findings.EI The Board disregarded this issue. The societal risk posed by Indian Point m.ust be larger than average if one accepts the conclusions that the frequency of severe releases is about average and the population is ten times larger. Certainly none of the Board's proposed " fixes" addrc5s this problem.
The Board Failed Completely to Consider Whether the PRA Results Were So Uncertain as to be an Inappropriate Tool for Decision-Making What is really at issue in terms of PRAs in this proceeding is the extent i to which PRA results are sufficiently reliable to aid in deciding whether Indian Point Units 2 and 3 should be :, hut down, remain in operation unchanged, or remain in operation after certain " fixes". There were two approaches to PRA presented to the Boa--d: . the IPPSS, prepared for the licensees, relied upon a Bayesian approach, while the NRC staff /Sandia estimates relied on a MAXIMUS ' statistical approach. The Board judged that the NRC staff /Sandia 12/ Id. at 33.
13/ Id.
M/ Intervenors' Proposed Findings, Question One, p. 41.
. ___)
_. . . .s . . ,...
Rowsome and Blond also testified that the In'dian Point reactors are roughly average with respect to the estimated frequency of severe releases of radioactivitySI -- thi; is significantly different from the second part of the conclusion of SECY-80-283. Nonetheless, Rowsome and Blond conclude that the societal risks posed by Indian Point are "probably average to above average" compared to other sites.I3,/-
How can this be the case if the frequency of severe releases is roughly average and the population is ten times larger? Obviously, the NRC staff's overall conclusion conflicts with the two subsidiary conciusions upon which it rests. UCS addressed this in our proposed findings.El The Board disregarded this issue. ,
The societal risk posed by Indian Point ntust be larger than average if one accepts the conclusions that the frequency of severe releases is about average and the population is ten times larger. Certainly none of the Board's proposed " fixes" address this problem.
-The Board Failea Completely to Consider Whether the PRA Results Were So Uncertain as to be an Inappropriate Tool for Decision-Making What is really at issue in terms of PRAs in this proceeding is the extent to which PRA results are sufficiently reliable to aid in deciding whether Indian Point Units 2 and 3 should be shut down, remain in operation unchanged, or remain in operation after certain " fixes". There were two approaches to PRA presented to the Board: the IPPSS, prepared for the licensees, relied upon a Bayesian approach, while the NRC staff /Sandia esti; nates relied on a MAXIMUS statistical approach. The Board judged that the NRC staff /Sandia 12/ Id. at 33.
13/ Id.
M / Intervenors' Proposed Findings, Question One, p. 41.
i
recommendations, but this scarcely does' our findings justice. In brief, as our findings detail, it is acknowledged that the uncertainties in the risk
-~
estimates are large, but no party put -forth a credible estimate of the range
' ~
of- uncertainties. Indeed, the NRC staff and Sandia witnesses 'took the position that a comprehensive uncertainty analysis cannot yet be performed.
The NRC staff presented only one witness' guess as to the possible range of uncertainties. The licensees' witnesses presented an " uncertainty" estimate, but the estimate was completely grounded in the sub]ective judgments of the analysts, and was rightfully rejected by both the intervenors and the Board.EI The limited uncertainty analysis performed by the KRC staff and Sandia witnesses accounted for only statistical sources of uncertainty, and did not acocunt for the very large uncertainties in data, completeness, and modeling.
More importantly, however. Sandia's analysis was thoroughly grounded in and dependent upon the IPPSS resul ts. Sandia examined in detail only those sequences found by the IPPSS to be risk dominant, and even then SarMia kept track of only those differences which caused differences in risk estimates by more than a factor of two. There was no exploration of the issue of whether l
l the IPPSS assessment of risk dominant accidents was correct. Thus, the NRC staff and Sandia risk estimates are in reality a mixture of IPPSS Bayesian results and " fudge factors" applied by Sandia to account for their differences with the IPPSS analysis. In sum, the NRC staff /Sandia risk estimates are fundamentally grounded in the very results from IPPSS that the Board rejected en methodological and statistical grounds.18_/
1_7/
7 Intervenors' Proposed Findings, Question One, pp. 44-47.
18/ Intervenors' Proposed Findings, Question One, pp. 6-9.
(
~
j; . . .
Moreover, the risk estimat'es are incomplete. Neither the NRC staff /Sandia nor the licensees' risk estimates (IPPSS) accounted for sabotage,
~ ~
desig'n/ construction errors.
~
aging effects, cognitive human errors, or Equipment failure due to environmental causes (environmental quali'fication) wasscarcely_consideredintheriskanalyses.El Cominents on the Bayesian Methodology in IPPSS On the matter of the Bayesian approach, we can add the following footnote to the Board's appropriate rejection of that methodology. Inasmuch as in an early order in this proceeding the Commission declared tha.t the ex parte rules would not apply and they would receive information on this case from anyone
[CLI-81-1, 13 NRC 1, 5 n.4 (1981), a revised, CLI-81-23,14 NRC 610 (1981)],
the Commission should be interested in the following. -
Robert G. Easterling of Sandia National Laboratories (who testified for the NRC Staff in the Indian Point proceeding and based on whose testimony the Board rejected the Bayesian methodology) sent the NRC a comment letter on June 2, 1983 (after the close of the record in this proceeding; copy attached),
regarding the NRC staff's " Safety Goal Evaluation Plan". In this comment, Easterling noted the following about Bayesian analysis, and about IPPSS and the Zion Probabilistic Safety Study (ZPSS), which were conducted by essentially the same analysts:
19/ Intervenors' Proposed Findings, 0uestion One, pp. 17-19, 28, and 34-38.
"In the Bayesian analysis performed for the Zion Probabilistic Safety Study, the reported data pertaining to the occurrence of a large LOCA were zero occurrences in 131 reactor-years (U.S. PWRs). The assessment that emerged, from the Bayesian analysis corresponded to zero occurrences in 845 reactor years. The prior distribution effectively contributed 714 LOCA-free reactor-years to the analysis.
The same analysts, when addressing Indian Point, chose a prior distribution that effectively contributed 344 LOCA-free years, even though the world's state-of-knowledge had not changd at all . Whimsy. Another example: At Zion, the motor-driven auxiliary feedwater pump data were four failures in 462 demands. The assumed prior was relatively optimistic and led to a posterior assessment based effectively on 2.7 failures in 542 demands; the prior substracted 1.3 failures and added 80 demands!" (Easterling letter, pages 2-3)
Dr. Easterling also addressed the Bayesian approach this way (Letter dated 6/2/83 from R. G. Easterling to S. J. Chilk, pages 3 and 4):
"Sometimes (as in WASH-1400), it is argued that tne prior distribution, f(p), represents physical variability such as amony plants, or manufacturers, or environments. While it is true that titere may in fact be such differences (and one role of statistical methods is to investigate these factors and account for them in estimates and in measures of uncertainty), there is no reason to assume their effects can be expressed by a known probability distribution.
Such an assumption is unwarranted. All that is lef t is to regard f(p) as some analyst's arbitrarily quantified degree-of-belief at some time. Does NRC really want to base important deci.51ons on such a foundation?"
L
" Suppose an analyst is uncertain about the correct
. mitigating system success criterion for a given accident sequence. He might entertain two or three or more possibilities and really not know how to choose among them.
He could assign subjective probabilities to these possibilities and- propogate away, but it seems to me that to do so would create articifial information. Ignorance is being replaced by numerical precision (and ignorance is not a uniform distribution)." (Emphasis added)
, , _ _ .~ _
One wonders'what else Easterling (and others) might have found had a truly 'in-depth evaluation of the IPPSS been performed. The superficiality of
~
-- the S'anilii review
~ ' of T the -IPPS5 ~was d e subject- ofT considerable proposed l --
~-
findings by the Intervenors.EI The Board itself noted that all Sandia did was to examine the twenty-eight accident sequences which IPPSS fcund to dominate risk.21/ Sandia (nor the NRC Staff, for that matter) did not, however, examine any sequences which IPPSS found to be non-dominant to see if revised frequency estiamtes might make them dominant; in other words, IPPSS could have made gross underestimation errors in the purportedly non-dominant sequences and neither Sandia nor the NRC Staff would have any way of knowing this. Even so, Sandia did find an important omission .in IPPSS (i.e., the failure to consider an accident c6used by a pipe break in the component cooling water system) which Sandia found to be an important contributor to core melt frequency.2_2/
Certainly it cannot be said that the IPPSS is so inherently reliable
.that it is unlikely that other errors are present in the portions not reviewed by the NRC staff and Sandia. On the contrary, the IPPSS is an advocacy document prepared to put the best light on the Indian Point plants.
20/ Intervenors' Proposed Findings, Question One, pp. 6-26.
22/. Recommendations, p. 45.
22/ Id., p. 45.
1 I
l l
~
The GA0 Review of IPPSS s
~ ~
~
The General l ccounting Office (GAO) did' a review' off the IDPSS anhlyssis. J - -
It concluded (GA0/RCED-83-158, 5/24/83, " Response to Specific Question On the Indian Point Probabilistic Safety Study"):
"The Indian Point PRA can be summarized by noting that al though the study analysts were able to identi fy the dominant contributors to risk at the plants, they were unable to identify the precise level of risk - associated with the plants. Thus, it appears that it would be extremely difficult to compare the risk of Indian Point to that of other plants, since PRAs performed on those plants would also be subject to considerable uncertainty in their risk estimates." GAO-RCED-83-158, page 5 .
" Employing the Bayesian methodology in nuclear powerplant risk assessments is relatively new and still controversial within the field of PRA. Experts differ as to whether an analyst can precisely translate his or her knowledge and beliefs to probability. Further, experts do not always share the same set of beliefs and consequently disagree on the probabilities." Id., page 40 The GAO review also concluded that the IPPSS aridlysis suffers from the same limitations in data, modeling, and completeness as all other PRAs. _Id.,
- p. 37. Thus, there is nothing essentially new about the IPPSS that makes it fundamentally more reliable than any other PRA.
i
Conclusion s
- ~ The Board-summarized -its first - two findings on risk- as--follows: (1) - A --
severe release at Indian Point could have more serious consequences 'than that same release at virtually any other site licensed by the Commission; and (2)
The chance of a severe release here is probably no greater, and may be less, than elsewhere.E In sum, we have come full circle. We "know" precious little more today about the probability of a severe accident at Indian Point than we knew at the time the UCS petition was filed, although the purses of the probabilistic risk analysts have been considerably enriched in the process,. and more than four years have passed.
The Intervenors urge the Commission to focus on what we do know: that a serious release at Indian Point would be a catastrophe of far greater dimensions than at any other site, and that the current state .of emergency preparedness is grossly inadequate to ensure that effective protective measures would be taken in the event of an accident. The Commission surely cannot continue to act as though adequate emergency preparedness is a reality
! at Indian Point. It is folly to operate two large nuclear plants in the most densely populated area of the country.
l I
-23/ Recommendations, p. 350 We are at a loss to understdnd the basis for a speculation that the chance of a severe release "may be less" at Indian .
Point. Having appropriately rejected the licensees' rosy figures and concluded that there is no basis for comparisons, there is certainly no ,
basis in the record for such a comforting hope. f 1
l l
This record establishes that the Indian Point plants could be shut down for far less' than the price of an insurance policy. We urge the Comission to
~
- - J take that course.- -- - --
Respectfully submitted, f
yn R. Weiss e '
General Counsel Union of Concerned Scientists b
M l- Stev'en C. Sholly Technical Research Associate Union of Concerned Scientists The following parties join in these coments:
Union of Concerned Scientists New York Public Interest Research Group Parents Concerned About Indian Point West Branch Conservation Association i
Rockland Citizens for Safe Energy Greater-New York Council on Energy-Friends of the Earth New York City Audubon Society Westchester People's Action Coalition Honorable Richard L. Brodsky l- Rockland County, New York l
New York City Council l 6 February 1984 I - - _ . . . _ _ . _ , . , _ . - . - _ - - , , _ . .
CCLMETEC USNR:
'84 FEB -6 P3 :18 UNITED STATES OF AMERICA LFF!h DF 5b,1Tm -
NUCLEAR REGULATORY COMMISSION 00Crli t{u{g ' EPvic BEFORE THE ATOMIC SAFETY AND LICENSING BOARD
'r
)
In the matter of )
)
CONSOLIDATED EDIS0N COMPANY OF NEW YORK, INC. ) Docket Nos.
(Indian Point, Unit No. 2) ) 50-247 SP
) 50-286 SP l POWER AUTHORITY OF THE STATE OF NEW YORK )
(Indian Point, Unit 90. 3) ) 6 February 1983
) -
CERTIFICATE OF SERVICE I hereby certify that a single copy of INTERVEN0iS' COMMENTS ON LICENSING BOARD RECOMMENDATIONS ON INDIAN POINT UNITS 2 and 3 was served upon the following by deposit in the U.S. mail, first class postage prepaid, this
'6th day of February 1984, except where noted otherwi e by asterisks.
($.
Steven G. 5 holly
'
- Nunzio Palladino, Chairman
- Victor Gilinsky, Commissioner U.S. Nuclear Regulatory Commission U.S. Nuclear Regulatory Commission
( Washington, D.C. 20555 Washington, D.C. 20555
- Frederick Bernthal, Commissioner
- Thomas Roberts, Commissioner U.S. Nuclear Regulatory Commission U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Washing ton , D.C. 20555 l e
- James Asselstine, Commissioner James P. Gleason, Esq. , Chairman U.S. Nuclear Regulatory Commission Aominstrative Judge Washington, D.C. 20555 Atomic Safety and Licensing Board 513 Gilmoure Drive '
Silver Spring, MD 20901 i.
I J
9 Dr. Oscar H. Paris '
Frederick J. Shon Administrative Judge Administrative Judge Atomic Safety and Licensing Atomic Safety and Licensing Board Board U.S. Nuclear Regulatory U.S. Nuclear Regulatory Commission Commission Washington, DC 20555 Washington, DC 20555
. David Lewis, Esq.
- 00cketing and Service Section Atomic Safety and Licensing Board Office of the Secretary i
U.S. Nuclear Regulatory Commmission U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Washington, D.C. 20555 Janice E. Moore, Esq.
Donald F. Hassell, Esq.
Henry J. McGurren, Esq.
Office of the Executive Legal Director U.S. Nuclear Regulatory Commission .
Washington, D.C. 20555 Brent L. Brandenburg, Esq.
Assistant General Counsel Paul F. Colarulli, Esq.
Consolidated Edison Company of Joseph J. Levin, Jr., Esq.
New York, Inc. Pamela S. Horowitz, Esq.
4 Irving Place Charles Morgan, Jr., Esq.
New York, NY 10003 Morgan Associates, Chartered 1899 L Street, N.W.
I Charles M. Pratt, Esq. Washington, D.C. 20036 Stephen L. Baum, Esq.
Power Authority of the State Mayor F. Webster Pierce of New York Village of Buchanan 10 Columbus Circle 236 Tate Avenue New York, NY 10019 Buchanan, NY 10511 Jonathon D. Feinberg - Stanley B. Klimberg, Esq.
- New York State Public Service General Counsel Commission New York State Energy Office 1
Three Empire State Plaza 2 Rockefeller State Plaza Albany, NY 12223 Albany, NY 12223 Charles J. Maikish, Esq. Marc L. Parris, Esq.
Litigation Division Eric Thorsen, Esq.
The Port Authority of New County Attorney York and New Jersey County of Rockland One World Trade Center 11 New Hempstead Road New York NY 10048 New City, NY 10956 Honorable Ruth Messinger Member of the Louncil of the Westchester County Executive City of New-York Care of: Laurie Vetere, Esq.
District 44 148 Martine Avenue City Hall White Plains, NY 10601
- New York, NY 10007
l Ezra I. Bialik, Esq. Andrew S. Roffe, Esq.
Steve Leipsiz, Esq. '
New York State Assembly Environmental Protection Bureau Albany, NY 12248
.New York State Attorney
--General's Office Honorable Richard L. Brodsky
. Two World Trade Center Member of' the County Legislature New York, NY 10047 Westchester County County Office Building I' Donald Davidoff,- Director White Plains, NY 10601 New York State Radiological I n
Emergency Preparedness Group Spence W. Perry, Esq.
Erpire State Plaza Office of General Counsel Tower Building, Room 1750 Federal Emergency Management Agency Albany, NY 12237 500 C Street, S.W.
Washington, D.C. 20472 ,
David H. Pikus. Esq. l Richard F. Czaja, Esq. Stewart M. Glass, Esq.
Shea and Gould Regional Counsel '
330 Madison Avenue Federal Emergency Management Agency New York, NY 10017 Room 1349 26 Federal Plaza .
Phyllis Rodriguez, Spokesperson New York, NY 10278
-Parents Concerned About Indian Point P.O. Box 125 Charles A. Scheiner, Co-Chairperson Croton-on-Hudson, NY 10520 Westchester People's Action Coalition, Inc.
Richard M. Hartzman, Esq. P.O. Box 488 Lorna Salzman White Plains, NY 10602 Friends of the Earth, Inc.
208 West 13th Street Alan Latman, Esq.
New York, NY 10011 44 Sunset Drive Croton-on-Hudson, NY 10520 Judith Kessler, Coordinator Rockland Citizens for Safe Energy Zipporah S. Fleisher 300 New Hempstead Road West Branch Conservation Association 6 New City, NY 10956 443 Buena Vista Road New City, NY 10956 l Renee Schwartz, Esq.
l Paul Chessin, Esq.
l Laurens R. Schwartz, Esq. Joan Hclt Project Director l Margaret Oppel, Esq. New York Public Interest Botein, Hays, Sklar & Hertzberg Research Group, Inc.
200 Park Avenue 9 Murray Street New York, NY 10166 New York, NY 10007 David B. Duboff Craig Kaplan, Esq.
Westchester People's Action National Emergency Civil Coali tior., Inc. Liberties Committee 255 Grove Street 175 Fifth Avenue, Suite 712 White Plains, NY 10601 New York, NY 10010 i
l
,' . . .o ..
- Jeffrey M. Blum, Esq.
Apartment 80 One Stuyvesant Oval
'New York, NY ~10009 - - -
1 Joan Miles Indian Point Coordinator Greater New York Council on Ene'rgy New York City Audubon' Society c/o Dean R. Corren, Director
. 71 West 23rd Street, Suite 1828 New York University New York, NY' 10010 26 Stuyvesant Street New York, NY 10003 l
l
- Served by messenger to indicated addressee at 1717 11 Street, NW, Washington, DC ,
e I
{
i s
l' - i