NUREG-0739, Forwards Analysis of Comments on 810326 Fr Notice, (46FR18827) & NUREG-0764, Toward Safety Goal:Discussion of Preliminary Policy Considerations
| ML20009C475 | |
| Person / Time | |
|---|---|
| Issue date: | 07/15/1981 |
| From: | Paul Goldberg NRC OFFICE OF POLICY EVALUATIONS (OPE) |
| To: | NRC OFFICE OF POLICY EVALUATIONS (OPE) |
| References | |
| FRN-45FR71023, FRN-46FR18827, RTR-NUREG-0739, RTR-NUREG-0764, RTR-NUREG-739, RTR-NUREG-764, RULE-PR-50 45FR71023, 46FR18827, NUDOCS 8107210128 | |
| Download: ML20009C475 (66) | |
Text
(
p# "'%
Sh a ftcfd[
j g
,,k UNITED STATES (y$
$f01 NUCLEAR REGULATORY COMMISSION j
(
p, WASHINGTON, D, C,20555
- y 9,,,,
+#
July 15, 1981 MEMORANDUM FOR:
File h
Ny FROM:
SUBJECT:
ANALYSIS OF COMMENTS ON DEVELOPMENT OF SAFETY GOAL This paper presents, in some detail, comments received on the Federal Register Notice (46 FR 18827, March 2C,1981) and the supporting report, NUREG-0764, "Toward a Safety Goal:
Discussion of Preliminary Policy Considerations." Table 1 briefly characterizes the comments and Table 2 lists the topics covered by comenters.
I have also prepared an abridged, eight-page version of this analysis.
\\\\*
cp..\\
y l
'*?uh 7ln aas, a L,$ :_ 'b :.'..,!l'7
'\\vj i_ : S:),
I,'
-Q\\ Y ]4 s O
Vj Y ({.
5
,,.( n()1981 " ~3 O
I b'
{
Contact:
E t
Paul Goldberg, OPE
{
(
63-43302 f
8107210128 810715 PDR REVGP I4RGISGDS PDR
~
\\
u ANALYSIS OF PUBLIC COMMENTS ON DEVELOPMENT OF A SAFETY G0AL TABLE OF CONTENTS Abstract of Comments Chart of ResponsesSection I Introduction - commentersSection II AIF Proposed Approach Section III Deficiencies of NRC Approach and Alternative ViewsSection IV Answers to QuestionsSection V Schedule of Safety Goal Development and Sequence of other NRC ActivitiesSection VI Other Comments Paul Goldberg Office of Policy Evaluation July 13,1931 e
4 4
0
I.
INTRODUCTION - COMMENTERS Twenty-five comme ts on the Safety Goal Project have been received as of July 7.
The commenters and the' dates their comments were received are listed below.
List of Commenters Marvin Lewis May 7 Norman Buske May 18 Atomic Industrial Forum May 18 Robert Alexander May 18 Washington Public Power Supply System May 26 Union of Concerned Scientists May 26 (asks for extention ofcommentperiod)
Kerr-McGee May 28 PLENTY May 28 Duke May 29 Cynthia Sharpe May 29 Tennessee Valley Authority June 1 1
Public Service Electric and Gas June 2 (notifies NRC that comraents will be late)
Stone & Webster-June 2 American Mining Congress June 2 General Electric June 11 Northeast Utilities June 16 Bechtel June 19 Dr. Henry Hurwitz June 25 UCS June 26 New York Public Interest Research Group June 26 Catherine Quigg June 29 Pollutio.n and Environmental Problems, Inc.
Eckhard Festag June 29 Combustion Engineering June 29 Lynn Rudmin Chong June 29 Westinghouse Electric Corporation July 1 Under each section, comments are listed in the order in which they were received.
e s
TABLE 1 COMMENTS RECEIVED:
LIST AND
SUMMARY
1.
Marvin Lewis believes that NRC is attempting to use a safety goal to 1.rovide a number of lives which may be traded for profit.
2.
Norman Buske believes that the safety goal diverts attention from a needed safety standard and that states should set their own standards.
3.
The Atomic Indust ial Forum (AIF) responds to the !uestions listed in NUREG-0764 and the Federal Register Notice and proffers its own proposed approach to safety goals.
4.
Robert Alexander believes that one more Class 9 accident will eliminate public acceptance of a nuclear program and that absence of strict verifiability dooms all quantitative goals.
5.
Washington Public Power Supply System (WPPSS) is in general agreement with the /.IF approach and believes that the technical bases for a goal by which it could measure and demonstrate the high level of safety of its plants exist today.
6.
Kerr-McGee believes that two separate goals are needed for low probability, high risk events on the one hand'and accidents or minor events in normal operations on the other.
7.
Albert Bates considers the release of radiation to the environment by any Federal agent or licensee a violation of the natural rights of citizens.
8.
Duke Power Company supports AIF's comments and proposal and' believes that application of probabilistic risk techniques through use of quantitative safety criteria can assure an acceptable level of safety.
9.
Cynthia Sharpe considers the safety coal a debate on the value of life versus the value of nuclear industry and technology.
10.
Tennessee Valley Authority (TVA) believes that the.ACRS proposal is a good starting point for discussion but that a goal should be simpler than the ACRS proposal.
11.
Stone & Webster (S&W) endorses the AIF proposal and believes that NRC must establish a set of rules with a single quantitative, absolute goal and should use the principles in WASH-1400 to analyze plant safety-related design features.
12.
The American Mining Company endorses the concept of a safety goal but believes that NRC should consider the distinctions among the various elements of the fuel cycle in developing the goal.
2-
- 13. General Electric (GE endorses the AIF proposal and advocates establishment of an interim safety goal representing an industry consensus.
- 14. Northeast Utilities endorses the AIF proposal eid believes that a goal is necessary to prioritize and assess the plethora of prop,osed changes to plants and regulations.
- 15. Bechtel endorses the AIF proposal.
16.
Dr. Henry Hurwitz states that the safety goal should avoid tha hypocrisy of establishing radiological standards for nuclear reactors that are not being applied to energy efficient dwelling.
17.
The Union of Concerned Scientists (UCS) feels that a quantitative safety goal cannot be a substitute for conservative deterministic criteria for licensing.
18.
The New York Public Interest Research Group (NYPIRG) expresses concern about the credibility of the NRC and other authorities and about NRC's treatment of public fears.
19.
Catherine Quigg believes that the ALARA (as low as reasonably achievable) approach to radiation exposure is unacceptable and that ALRAA (as low as achievable) would be a more socially responsible goal.
i 20.
Eckhard Festig believes that there should be a moratorium on reactor construction until and unless a research program establishes that reactors with more desirable characteristics can be built. '
21.
Combustion Engineering endorses the AIF comments and proposal.
22.
Lynn Rudmin Chong suggests that the ALARA policy be discontinued I
and that the Union of Concerned Scientists and Physicians for Social Responsibility be included in NRC decisions to allay public fear and distrust.
- 23. Westinghouse endorses the AIF proposal.
_ _. ---+
_,-,y--,_
..-.-,-._,-m
-. ----m.,
1A3LE 2 TOPICS COVERED BY COMMENTERS Critique of NRC AIF Report
_ Questions Other Coninenter Proposal Schedule
(.?J.' LEG-0764 )
1 2
3 4
5 6 7 Coments Marvin Lewis X
X X
Norman Baske X
AIF X*
X X
X X
X X X X Robert Alexander X
S&W X
X X
X X
X X X WPPSS X
X X
X X
X X X Kerr-McGee X
Albert Bates, PLENTY X
Duke X
X Cynthia Sharpe X
TVA X
X X
X X
X X X X PSE&G AMC X
GE X
X X
NU X
X X
X X
X X X Bechtel X
X Dr. llenry liprwitz X
UCS X
X NY PIRG X
X X
X X
X X Catherine Quigg X
Eckhard Festag X
Combustion Engineering X
X Lynn Rudmin Chong X
X.
X
- The AIF proposal is the only detail' d proposal received in response to the Commission's solicitation.
e X's in this column opposite other comentersindicate references to the AIF proposal.
II. AIF PROPOSED APPROACH The most extensive comments are those of the Atomic Industrial Forum (AIF), which commented on NUREG-0764 and also offered "A Proposed Approach to the Establishment and Use of Quantitative Safety Goals in the Nuclear Regulatory Process." Because this is a detailed proposal which has been endorsed by eight of the other commenters -- Stone and Webster (S&W), Duke Power (Duke), Washington Public Power Supply System (WPPSS), General Electric (GE), Northeast Utilities (NU), Bechtel Power Corporation (Bechtel), Combustion Engineering (CE),andWestinghouseElectricCorporation--AIF's approach is outlined here.
1 AIF suggests three governing principles to provide a rational and f
logical framework upon which specific quantitative safety goals can be founded:
l 1.
The goals should provide a level of protection for members of the public such that no individual bears an inordinate risk.
t 2.
The quantitative safety goals for nuclear power plants should j
be consistent with those appifed to other technologies. The r,Jals should endeavor to ensure that incremental societal risks are commensurate with the societal benefits derived from the technology.
I l
3.
The goals should promote the rational allocation of societal resources for the purposes of reducing public risk in order to achieve the optimum benefit attainabic for the cost.
l l
l l
li-2 In accord with these principles, AIF proposes two primary criteria and two secondary criteria each of which has an associated quantitative goal which satisfies the criterion in AIF's view. The primary criteria establish goals for limiting radiological health risks to individuals and to society at large.
The secondary criteria relate to allocation of resources in achieving marginal reductions in residual risk and in providing goals for the prevention of accidents which could pose risk.
The primary criteria and associated goals are:
INDIVIDUAL RISK CP.ITERION The maximum incremental risk of radiologically induced adverse health effects to a hypothetical biologically average individual in the vicinity of a nuclear plant site should not result in a significant increase in the individual's annual mortality risk.
t.
Suggested Safety Goal:
-5 10 yr mortality risk to maximum exposed average individual.
1 POPULATION RISK CRITERION The incremental cumulative risk of adverse radiologically induced health effects to the exposed population per 1000 MW(e) of nuclear power capacity, considering the annual frequency and consequences O
,--e g,
c
,,e
II-3 of events integrated over the spectrum of potential accidents, should be no more than a small fraction of the average background incidence of health effects.
4 Suggested Safety Goal:
1 Statistically estimated fatality /yr per 1000 f54(e)
The secondary criteria and associated goals are:
COST-BENEFIT CRITERION Measures proposed to achieve incremerital reductions in residual risk beyond those provided to meet the primary criteria should be 4
evaluated on a cost-benefit basis.
The benefit, in terms of population risk reduction, afforded by a change in plant' design or l
operating procedure should be comparable to that which is generally achievable through alternative investment of the cost of the change in other areas of public risk reduction.
l Suggested Goal:
$100/ man-rem (equivalent to $1 million per statistically estimated life saved).
II-4 LARGE SCALE FUEL MELT CRITERION Although the basic principles established for formulating quantitative safety goals can be satisified through the previous three criteria, an additional decision criterion for industry and regulatory management is proposed as a secondary or subordinate goal involving the probability of large scale fuel melt accidents:
The probability of accidents involving large scale fuel melt should be sufficiently low that, given the expected population of reactors, the recurrence ' interval for such accidents is on the order of one per several decades.
A large scale fuel melt condition is defined by a set of system or equipment states such that the core is uncovered for a sufficient period of time to cause the bulk of the fuel to become molten.
Suggested Goal:
10-4/ reactor-yr fcr Probability of Large Scale Fuel Melt AIF believes that quantitative safety goals are required for decision-making using probabilistic risk assessment (PRA) analyses performed for the purpose of making regulatory judgments and that submittal of comprehensive PRA studies should not be made a specific condition for licensing of individual plants.
Rather the initial application of PRA and quantitative goals in the regulatory process
II-5 should be to examine the existing body of regulations to establish generically the level of safety which they provide and to identify areas where changes are warranted. This.till permit the determination of whether compliance with existing regulations will provide a level of safety that satisfies the quantitative safety goals.
Applicants who satisfy these regulations can thus be presumed to have met'the safety goals unless the plant design or site has unique features.
4 AIF believes that this approach helps ensure stability of the basic licensing review process and timely decision-making on pending or future applications and is consistent with the recommendations of the Risk Assessment Review (Lewis) Group and the Rogovin Inquiry Group to introduce PRA into the licensing process in a gradual and systematic manner.
l i
,-., ~.
,e,,
e--
c,
--m,
,w.
,-.-----------,-n-
~, -
III'. DEFICIENCIES OF NRC APPROACH AND ALTERNATIVE VIEWS Nine commenters take the view that NUREG-0764 leaves out considerations important to developing a safety goal or that an acceptable safety goal cannot be developed.
One of them, Robert Alexander, believes that "one more Class 9 accident" will eliminate public acceptance of a nuclear program and that absence of strict verifiability dooms all quantitative goals.
Cynthia Sharpe finds the notion that it is possible to develop a " safety goal" frightening because " numerous studies have shown that there is no safe exposure to radiation" and she considers the development of a safety goal a debate on the value of life versus the value of nuclear industry and technology.
She believes that no cost-effectiveness comparison applies when w'e talk of people's lives and that regulations should consider that nuclear plants are on their way out of use; that it has been determined that their risk is not worth the electricity they generate.
She believes that solar power, wind power, and other forms of electric generation are available and preferable to nuclear power and that they obviate the i
need for nuclear power.
She feels that safety goals should address only currently operating plants; plants which are not yet operational should be converted to a non-nuclear fuel source.
~, - - -.
y e
7,
_. e -, -
.3
,,, _, _,., --, my
III-2 Marvin Lewis feels that the NRC has had little discussion with groups with the most expertise in risk / benefit comparison: morality has not been discussed with clergy and the willingness to lay our lives on the line has not been discussed with ratepayers. He asserts that probabilistic risk analysis, which is the basis for the safety goal proposals summarized in NUREG-0764, did a very poor job of assessing numerical values to accidents (sic) according to the Risk Assessment Review Group chaired by Dr. Harold Lewis.
According to Marvin Lewis, the proper objective of a safety goal is to save lives and justify turning off all nuclear power plants; the improper use is to provide a number of lives which may be traded for profit. He feels that this latter use is what NRC is attempting to do. He also feels that the entire safety goal program is couched in bureaucratese that is almost indecipherable.
The Union of Concerned Scientists (UCS) questions NRC's a'pproacn to safety goals and the use of probabilistic risk assessment in the following terms: There is no clear statement of the use to which a quantitative safety goal would be put in the context of the licensing and regulatory processes.
If it is to be used in making decisions about the acceptability of specific nuclear plants, far greater attention must be devoted to considering the extent to which it is possiF.e to determine whether the goal has been met.
In UCS's view, the uncertainties in quantitative risk assessment are so great as to preclude an objectively defensible determination that any particular goal has been met.
I I-3 Can quantitative safety goals be used in the regulatory process if quantitative risk assessment is uncapable of yielding technically supportable results within an acceptable limit of uncertainty?
UCS believes that the answer to this question is "no."
Risk estimates which are inherently unreliable cannot legitimately be used as regulatory tools. They give the illusion of precision but can be manipulated to support whatever the predetermined desired objective may be. The safety goal, and the risk assessments used to demonstrate that is has been met, will not be used in a pure effort to advance scientifc understanding. On the contrary, they will be used to make decisions when have great economic and political consequences for the organizations which will sponsor the wcrk.
Unless the public can be assured that the means exist to verify compliance with a safety goal, it will be justifiably concluded that this represents an effort to mask the risks of nuclear power rather than to identify and understand them.
All of the approaches suggested in NUREG-0739 to asse.ssing' risk virtually ignore one extremely important element of risk that is unique to nuclear power: the potential for essentially permanent contamination of hundreds of square miles and decontamination to the extent possible of thousands of square miles, including the risk of permanently losing towns, farms, and industries as well as important natural resources such as equifers. These risks should be assessed in terms of the economic costs involved and also the social trauma and dislocation of such an accident.
+-
III-4 Because of the unique risks associated with a nuclear catastrophe, the risk aversion factor for nuclear generation should be significantly higher than the 1.2 proposed (in NUREG-0739).
Moreover, equivalent deaths is an incomplete and unsatisfactory measure of the nuclear risk.
Compariso'ns of nuclear risks to coal risks are based on the implicit premise that those are the two significant energy choices at present.
In fact, orders for future nuclear plants have declined to a virtual halt in large part because of the drop in growth of demand for electricty, which has now persisted for almost a decade.
Thus, energy conservation is a primary alternative to nuclear construction.
If the risk comparison is to be meaningful, it should compare nuclear with conservation as well as with coal.
i The AIF proposal (NUREG-0739, p. 39-41) is objectionable iri several respects.
First, UCS sees no obvious reason why the safety goals "should be generally applicable to all technologies or risk related activities." The purpose of a safety goal is presumably to establish acceptable level of risk from nuclear power. As noted above, there are unique aspects of that risk which are not associated with either other modes of generating or other risk-related activities.
Efforts to treat all risks as comparable inevitably overlook these important nuclear risks.
III-5 AIF also suggests that the safety goal should " reflect societal benefits." Whether one believes that nuclear power is a benefit at all depends on how one views the alternatives from a variety of standpoints beyond considerations of risk.
It is apparent that a determinaticn of the societal benefit of a nuclear plant is a political determination' well beyond the scope of NRC's expertise.
Moreover, it is fundamentally inconsistent with the Atomic Energy Act and with the principles which the AEC and NRC have espoused since the outset of the nuclear power program.
Finally, AIF suggests that it is ' acceptable to have an accident as serious as TMI-2 every several decades.
UCS suspects that the public is not willing to tolerate an accident of the severity of TMI-2 every 30 years, even if only economic and social consequences are considered.
It is also extremely dubious that application of such a safety goal could assure that the one serious core damage accident every thirty years could be contained without deteriorating into a far more serious accident than TMI-2.
It is difficult to conceive of a use of risk assessment techniques that could reliably ensure that an accident would progress only as far as TMI-2 but not to catastrophe.
In summary, UCS believes that in order for a safety goal to be a useful tool for regulating nuclear power, it must meet the following I
conditions:
i I
l
III-6 1.
Compliance with the goal must be technically verifiable within reasonably small uncertai.nty limits.
Stating the uncertainty, however clearly, is not a substitute for reducing it to acceptable levels when regulatory decisions are to be based upon quantitative assessment.
2.
Establishment and implementation of the goal must fairly account for the unique risks of catastrophic nuclear, accidents, including economic costs.
3.
Establishment and implementation of the goal must not I
require NRC to resolve questions which are policital in nature.
4.
The scientific and technical community outside the nuclear industry and NRC must be involved in establishing th,e goal and reviewing risk assessments. We are, unhappily, confident that risk assessment is destined to be an adversary tool for the nuclear industry to resist safety improvements rather than as a means to build consensus around a safety goal.
5.
Both the safety goal and the quantitative risk assessments must be undsrstood and accepted by the public as being unbiased and technically justified.
O yr-M-m-
--:--,e g
g-9g--e,w
,-g-,
e
---,w.-,,,,m4-n 9--e-or---
,-y, y
yw--
III-7 6.
A quantitative safety goal cannot be a substitute for conservative deterministic criteria for the licensing of nuclear plants.
NYPIRG seeks to place the consideration of safety goals in the historical context of radiation dangers and standards and i:uclear industry development.
In NYPIRG's view, international standards of
" safe" limits of exposure were set and have been revised from time to time -- always in rueful recognition that ionizing radiation is more dangerous than had been previously thought.
Recent, more precise computations' indicate that much less of the energy of the Hiroshima bomb went into neutrons than had been thought and more into gamma rays. itany of the consequent cancers had been attributed to neutrons; NYPIRG feels that the new data open the possible interpretation, urged especially by Edward-Radford (Chairman of the NAS Committee on the Biological Effects of Ionizing Radiation, which was reported in 1980), that gamma radiation is more carcinogenic than had been assumed. Meanwhile all current, figures on medical consequences of exposure to radiation must be considered tentative and subject to rapid change.
l
,-.,__w.
o III-8 The standards that were adopted for workers and for the general public (in tiie new atomic industry following WW II) were derived from the Japanese experience; tM dangers were considered slight encugh so that it would be feasible to have such an industry. The decision-makers apparently believed that the dangers of ionizing radiation were slight enough so that the country should commit itself to a program of aggressively promoting atomic energy.
In accordanc.e with the prevailing promotional atmosphere, the AEC consistently failed to gather data on the actual exposure histories of all workers in every sector of the nuclear industry -- and their medical histories as well.
In light of the state of ignorance about such vital matters, this neglect is very much to be regretted; but even today, the lesson has not been learned. The needed data are still not being gathered, and we are learning much more slowly than should have been necessary that low-level exposures are bad for people.
In retrospect, it was an enormous error for this country to have made the kind of commitment it did to nuclear energy before the dangers were fully understood and long before the technology had been satisfactorily worked out for the entire fuel cycle. An enormous institutional momentum was set up, not only in the utilities, but in the goverment.
After the Three Mile Island accident, public opinion took a sharp turn toward skepticism and distrust of the official reassurances about the safety of nuclear power.
Several observers (e.g.,
Slovic, Fischhoff, & Lichstein,1980) have noted that in the public's mind, the risks of nuclear power are uriique: "Few technological activities are judged so uncontrollable, unknown, potentially catastrophic, or dread."
III-9 People's inability to perceive radiation with their senses means that we are dependent on authorities of unknown trustworthiness and credibility not only to warn us of danger, but to let us know when it has arrived, to tell us what to do, to help us know when the danger has passed, and to inform people whether they have been injured.
Judging by the manner in which new emergency planning for nuclear plant accidents is being handled by industry and the federal review agencies (NRC and FEMA), there is little evidence that NRC has learned to respect and take seriously the importance of public confidence. At a time when it is critical to rebuild the faith and trust of the public if nuclear power is to continue, NRC has failed to strictly enforce its own regulation -- that e tergency plans be fully implemented (except for public notification capability) i by April 1, 1981. The public is looking for signs of genuine change of attitude about public health and safety in the " upgrading" of emergency preparedness, and it finds that its fears coi;tinue l
to be treated as a public relations problem rather than as legitimate concerns.
The administration of the emergency plan review process l
has undoubtedly contributed to further public skepticism about the NRC's credibility and its commitment to safety. We hope that at least in the present effort to set safety standards, the issue of public trust will be more realistically considered.
l l
l
O III-10 If a safety goal for the generating plant is formulated in terms of fractional deaths per megarfatt year of generated electricity, that is likely to look like and be taken as the total health consequence of nuclear power whereas it is in fact but a small part of the integrated sum for the whole fuel cycle.
Catherine Quigg, Research Director of Pollution arid Environmental Control, Inc., includes two articles in her coments: One, "New Trends in Safety Design and Analysis," by Dr. David Okrent, Professor at the University of California and member of the Advisory Comittee on Reactor Safeguards, describes changes in safety-related elements of nuclear plants which the author anticipates as a result of the Three-Mile Island accident.
l The other article,.from the May,1981 issue of Science magazine is entitled "New A-Bomb Studies Alter Radiation Estimates." In Ms.
l Quigg's view, it provides further evidence that the ALARA (as low as reasonably achievable) approach to radiation exposure is not l
acceptable and that ALAA (as low as achievable) would be a more socially responsible goal.
Ms. Quigg also believes that attention shold be given to the recent Supreme Court decision on cost / benefit decisions when safety is at risk.
She feels that this decision may preclude any consideration of cost when setting health and safety goals for workers and the public.
III-11 Eckhard Festag states that he is one of a growing number of individuals who believe that nuclear power plants should not be built at all.
He attaches an article from "American Scientist, May-June 1981, entitled " Nuclear Synergism: An Emerging Framework for Energy Systems," by A. A. Harms and W. Haefele. Mr. Festag believes that the aritcle underscores his contention that the current construction program is premature by pointing to the possibility of building reactors that are much more ecologically sound and fuel-conserving than the present generation.
l In Mr. Festag's opinion, no more' commercial reactors should be built.
Instead, there should be a research program along the lines indicated by Harms and Haefele which, if successful, may give way to the resumption of commercial construction.
e O
y w
III-12 Lynn Rudmin Cho.ng suggests that:
a)
ALARA as policy be discontinued. Health effects "as low as reasonably achievable" is tantamount to your agency committing premeditated murder. What ALARA means is that a certain number of cancers -- numbers estimated beforehand by your agency -- are allowable in exchange for nuclear-generated electrici ty.
Honicker v. NRC is probably but the first' of countless attempts by the public to disagree with A!, ARA.
b)
The Union of Concerned Scientists be always included in engineering decisions that involve NRC. There is never a way for the public to trust that NRC and the nuclear industry are not hiring " rubber stamps" when engineering decisions need criticism.
Including the Union of Concerned Scienti ts will somewhat allay public fears and distrust.
c)
The Physicians for Social Responsibility be always included in EIS decisions and input and in the preliminary decisions that lead to an EIS being done. There is never a way for the public to trust that RNC and the nuclear industry are not hiring " rubber stamps" when health effects decisions need criticism.
Including the Physicians for Social Responsibility will somewhat allay public fears and distrust.
III-13 Albert Bates considers NRC's use of the concept of risk to be an over-simplification which raises problems. One problem is that although overall risk numbers can be derived for events, the risks to individuals differ greatly. To an individual who is injured by an event the risk becomes a certainty and while the risk may have been acceptable, there is a very good possibility that the certainty is not. Technologies which emit untraceable agents of injury, such as radiation, into the environment have rendered obsolete the legal mechanisms which traditionally have established causality of injury and have compensated victims. As causation cannot be proven, compensation cannot be made and it may be said that society itself is injured. Mr. Bates, considers the release of radiation to the environment by any Federal agency or licensee a violation of the natural rights of citizens to which it is beyond our capabillity to assign a value or cost. Mr. Bates considers the statement on p. 21 of NUREG-0764, "the cumulative record of evolutionary processes should be consulted for guidance (in establishing a safety goal) only if such processes properly accommodate social pressures and realities,"
naive.
In his view, social pressures and realities may be subjective, fickle, or not adequately adjusted for the consequences of conduct within the finite physical s'ystem supporting a society; they should be accommodated by redirection within the ongoing evolutionary process.
The cumulative record of evolutionary processes is a more objective l
f yardstick.
l e
-c,-
r
,r w-y.--.-
4 IV. ANSWERS TO QUESTIONS The following questions were published for public comment in the Federal Register (46 FR 24336, April 30, 1981) as part of the Commission Statement on Development of a Safety Goal -- Preliminary Policy Considerations, which was also included in NUREG-0764. The responses.of commenters are listed under the various questions:
Question 1(a)
Among the criteria for selection of an approach to safety-goal formulation that are presented, which ar~e particularly important? Unimportant?
Illus-trative criteria are that the goal be comprehensive, logical, verifiable, practical, and publicly acceptable, flesponses S&W believes that an approach to the formulation of selection criteria should recognize that the process of evaluations is more important than whether or not the goal is ideal. The objective should be to select some-thing reasonable and make sure everyone understands it in principle rather than to quibble about whether cr net the goal pleases everyone.
WPPSS considers all of the listed criteria important and finds it difficult to rank them in order of importance.
WPPSS has the following comments on the criteria listed:
Comprehensive -- The safety goals should lead to nuclear plant safety standards for health risks to individuals and the population.
______________J
IV-2 Verifiable -- Compliance with the safety goal should be verified by a two-step process:
(1) Review of the structure of the fault trees and event trees used in the PRA in accordance with the PRA guidelines being formulated now by the Technical Writing Group.
(2) Review of the probabilities ar.d uncertainties assigned to basic events in the fault trees and to event tree branch points.
i New data may require re-evaluation of parts of the PRA.
Practical -- The implementation of the safety standards should result in more effective use of resources to reduce significant risks.
WPPSS agrees with the NUREG-0764 statement that the safety goal should "... lea' to the d
resolution of real problems by real people within the real resource constraints.
l Acceptance to the Public -- The technical community should recommend com-prehensive, logical, verifiable, and practical safety goals and implement standards based on these goals.
If this is done with effective communications with the public, then the result should be acceptable to the public. WPPSS I
cannot support the incorporation of non-technical factors based on current l
perceptions of _ a segment of the population in the name of "public acceptance."
l We feel the resulting distortion of the technicai credibility of the goal l
and evaluation process would undermine long tern acceptance.
l l
l i
__ ~.
IV-3 TVA believes that, while each of the five illustrative criteria may be necessary, simplicity should be the primary consideration, because unneces-sarily complex goals are not censistent with the degree of precision in the results obtained form risk assessments.
General Electric feels that, the following aspects of the criteria are of particular significance:
The safety goal form should be quantitative in nature.
This would allow an objective assurance that the public health and safety is being maintained. A quantitative goal would also provide a firm basis for evaluating PRAs and determining their compliance with the safety goals.
s The quantitative safety goal should be simply stated and readily applicable.
This would ensure a good degree of understanding and will contribute to the acceptability of the safety goal selected.
Ease in application should also allow the designer maximum flexibility in reducing risk to the public through either preventive or mitigative means.
-- The procedure for determining compliance with the quantitative safety goal should not incorporate conservatism in it methodology.
Proba-bilistic Risk Assessments should treat uncertainty through a prescribed methodology for modeling and calculations.
This would ensure the most realistic results are obtained allowing a reasonable comparison to the risk studies with the safety goal.
IV-4 Northeast Utilities states that it will be difficult to formulate an approach to a safety goal that provides for all ideal criteria on any short timeframe.
Thus, they submit that the most important criteria should be emphasized:
The approach should be comprehensive and logical.
If these two are. net, then the other criteria will follow automatically.
NYPIRG states that the criteria listed are general enough to seem largely unobjectionable.
Yet they feel that the criteria fail to direct attention to the one matter of overric ing importance:
survival.
They approach this matter by making explicit what they consider to be basic assumptions and concerns about nuclear power safety.
The primary and overriding value of our society is the survival of the a.
human species, and, more narrowly, national security or the survival of the American population and its central social institutions.
b.
When an activity threatens the survival of a substantial fraction of the American population and of basic American institutions, the public and its elected representatives wish to take all available' preventive measures, regardless of the associated probability of disaster and with only secondary concern for cost.
It i:-
t easy to achieve consensus on just how large a fraction of the population would have to lose their lives to constitute an unacceptable threat to the survival of the nation, 3
but it seems wise to err on the side of conservatism.
3 The prevailing experts on national defense believe that there is an c.
l l
.n
4 IV-5 appreciable probability of a nuclear war in which many people would be killed outright and huge numbers more slowly. Therefore, there is s
considerable controversy about the ability of the nation to survive or recover.
It follows, further, that in considering any other social policy, we must examine its implications (if any) for national sur-vivability after an attack by atomic bombs.
d.
In a number of ways, the more we rely on the central generation of electricity by means of nuclear reactors as an important component of a national et..cgy policy,.the more seriously we undermine that suvivability, for the following reasons:
1.
lants for the central generation of electricity are specially vulnerable to attack, whether by saboteurs, terrorists, or by enemies in a formal war.
II. The recently discovered vulnerability of electric power grids to dis-ruption by tte EMP (brief intense pulse of electricity) produced by extra-atmospheric detonation of a nuclear bomb adds another urgent security reason not to prolong our dependence on large central electrical generators.
III. In addition to the crippling effect upon industry, domestic consumers, and everyone else who is dependent upon a reliable supply of electricity, a direct hit upon a nuclear plant by a nuclear weapon could greatly increase the amount of lethal fallout over an enormous area.
IV-6 e.
Of all the methods of obtaining energy currently used or near to com-merical application, only the nuclear generation of electricity itself possesses the capability of generating accidents (in peacetime) so catastrophic as to threaten society itself.
NYPIRG strongly urges that the first priority for a consideration of safety goals is to proceed with the following steps:
(1) Determine a threshold level of unacceptable threat to the survival of the United States, so serious that the benefits of generating electrical power could not possibly counterbal'ance it.
(2) Explicitly recognize the fact that this country has always taken the prudent position that it will not willingly imperil its survival, no matter how small the calculated probability, if the risk can be avoided without even more aversive consequences.
(3) Examine the several suggested ways that existing nuclear reactors of various desi'gns may cause unacceptable disasters in war and in peace, and establish a list of all possible scenarios by which such disasters might take place.
(4) Fix as the first safety goal the elimination of the possibility of each such scenario.
(5) Accept the implication that if unacceptable disasters cannot in
IV-7 principle be absolutely avoided at any existing or proposed comercial nuclear reactor, it must be abandoned.
i NYPIRG asserts that if the fears of the anti-nuclear scientists and engineers are unrealistic, it is of the utmo'st importance that detailed proof be given to lay these fears to rest. Af ter this major issue of safety has been considered, most of the more minor and less pressing issues raised in the materials generated by the NRC to date can be given consideration.
NYPIRG feels that a reasonable safety criterion, which would force the nuclear power industry to prevent the possibility of catastrophe, would be that no nuclear plant should be licensed unless it is fully covered by commercial insurance.
For the present situation is, in their considered judgment, quite intolerable:
the victims of future radiation damage must pay the costs and must bear all the losses uncompensated.
Many ]f the safety goals elaborated in NUREG-0764 and many of the questions raised in the Federal Register are meaningful primarily in relation to the dangers of routine operation and sub-catastrophic accidants, which are serious enough so that they must be addressed.
Oddly enough, however, nothing is said about cancer and other illnesses or genetic effects of the radionuclides spread around the environment as a result of drilling for i
uranium, mining it, milling and fabricating and transporting it; or as a result of failure to keep the wastes from the whole cycle (especially the
{
spent fuel and decommissioned reactors) totally segregated from the biosphere l
for the next half million years.
IV-U Question 1(b)
Should additional or different criteria be considered? What criteria and with how much emphasis?
1 Responses AIF cites the three bas': principles listed in the summary of its proposal above and adds that the goal should have the attributes of simplicity of expression and practicality of application.
These are important because unduly complex goals are not consistent with the degree of orecision in the results obtained from risk assessments'and because simplicity will make the goals more understandable and acceptable to the technical community and the public.
S&W considers the various presentations to date more than sufficient bases i
for selection of criteria. TVA cites three basic principles which are very similar to AIF's:
l 1.
The safety goals (which imply risk) as applied to nuclear power facilities should be commensurate with the safety goals of other competing technologies.
2.
The safety gcals should take into account optimum allocation of the nation's resources over the long term.
3.
No individual should bear an unreasonable risk.
-,----~-----,,--,-----,.,-,,-----c y--
IV-9 NU maintains that even though the goal will have social issue consequences,-
it is not a social issue goal. The safety goal will be a technological goal.
Its technological success or failure will establish the social per-ceptions, and it should not be vice versa.
This perspective can be main-tained by allowing the social and political concerns to be dealt with on a public and Congressional level which will reflect the technological efforts of the industry, NRC staff, ACRS and Commissioners.
Question 2 Which of the following are particularly important to include in a safety goal.
(a) Some general approach to risk acceptability?
(b) Quantitative safety goals?
(c) Qualitative -- even subjective -- standards?
(d) Approach to safety-cost trade-offs?
(e)
Goals for future safety improvements?
(f) Standards for determinipg when new requirements should be applied retroactivsly?
Responses AIF's answer is as follows:
(a) This is emphasized in the three basic principles.
(b) Quantitative safety goals are essential to improving the rationality of the regulatory process and to derive the full benefits of applying.TA techniques to regulatory decision-making.
O IV-10 (c) The current regulatory process is based on a qualitative approach to ensure an acceptable level of risk to the public.
While we believe that this approach has resulted in acceptably small risk, it is often difficult to demonstrate that the NUREG-0764 criteria of logic, veri-fiability, practicality and public acceptance are met.
We therefore believe that the final result in safety goal development should be quantitative values based on underlying qualitative principles.
(d) Safety-cost trade-offs should be an important secondary goal.
(See response to 5(a).
(e) The safety goals should include a cost-benefit criterion for determining
' need for future safety improvements once the primary safety goals are met.
(f) The cost-benefit criterion should also be used to determine the need for app. lying new requirements retroactively.
WPPSS considers items (a), (b), (d), and (f) important. Their recommen-dations on these items appear in responses to other questions. They recommend that qualitative standards (c) not be included in goal (though the evaluation process for goal implementation will necessarily include criteria for qualitative judgments) and that goals for future improvements be postponed until a standard is subjected ta a trial period.
TVA's responses are as follows:
7y
-.. ~,, - -..-
. - ~, - -.
IV-11 (a) Yes, the concept of risk and safety goals should be approached in view of the three basfc orinciples described.in 1(b) above.
(b) Yes, attention should be focused on the development of quantitative safety goals.
Quantitative safety goals provide a common basis for r
evaluation and comparison.
(c) A qualified yes.
Care must be exercised, however, since qualitative goals frequently attempt to define an individual (s) perception of risk and as such cannot be measured on a common basis.
(d) Safety-cost trade-off criteria should be available in order to provide a measure of incremental risk reduction.
Its role should be in con-sideration of additional risk reduction after the primary safety goals l
are met or are exceeded.
As such, safety-cost trade-off cr,iteria should l
l be considered a secondary goal.
(e and f) The nuclear safety goals should be commensurate with the safety goals, cost and benefits of other technologies as much as practicable.
If and when changes occur in other technologies which result in changes i
l in their risk, cost, benefits, nuclear plant risk should change accord-ingly.
The safety goals should be consistent with optimum allocation of the nation's resources over the long term.
S&W believes that a general opproach to isk acceptability should be combined with qualitative standards as statements of intent.
Safety /cest trade-off
IV-12 values should be used in conjunction with decision criteria for retroactive application, and as a balancing mechanism to keep requirements in perspective with other risks.
Of particular importance in the establishment of a safety goal are quantitative emphases, both in the goal itself and in the analyses that verify that the goal is met.
Implementation of proposed safety improve-ments in existing and future plants can and should be dependent on the results of safety / cost trade-offs that take into consideration the safety goal.
In the view of NU the safety goal should specifically define acceptable risks. The goal should be quantitative' and the evaluations to assess compliance should be quantitative.
The goal should focus on safety, not on cost / benefit.
Cost / benefit should be subsequently considered by the licensee in evaluating alternative corrections to identified deficiences.
New requirements should not be automatically imposed on plants that meet the safety goal.
They should be identified as optional modifications which may be judged by the licensee as to whether their implementation can more effectively or efficiently meet the safety goals.
Plants that do not meet the safety goal should have the same information made available to them along with guidance from the NRC.
(a) NYPIRG urges a dual approach:
first, the consideration of truly disastrous consequences of the operation of nuclear power plants, where probabilistic considerations play no part; and second, all other accidents and undesirable side effects of nuclear power, which we will for con-venience, call nondisastrous.
IV-13 (b) NYPIRG rejects the use of a probabilistic framework for risks once the threshold of socially unacceptable disaster has been passed.
They cannot think of an example outside tne realm of nuclear power where equally severe, potentially catastrophic consequences of a public policy have been tolerated on the argument that the computed probab;Tity of the disaster is small. Probabilistic risk assessment has the non-obvious flaw of pseudoprecision:
since it gives rise to highly specific numbers, at least as used to date, it can give rise to a wholly specious sense of security.
No advocate of such methods as those used in WASH-1400 is willing to defend the proposition that every possible cause of an accident has been taken into account; it is completely obvious that to do so is impossible.
For that reason alone, the resulting numbers are not nearly as precise as they look. Worse, there is logically no way to know just how imprecise they are.
No one as yet has assigned a probability to sabotage or the chance that a grave accident may be caused by deliberate malevolence.
t (c) NYPIRG urges the use of cumulative life-shortening in place of separate i
consequence figures on " prompt" and " delayed" or latent deaths. We are strongly opposed to attempts to set a dollar value on human lives, l
convenient though that is for the risk assessor.
Instead, quantitative risk assessments must include, at a minimum, not only cumulative life-shortening but total monetary loss.
(d) NYPIRG considers safety-cost trade-offs relevant only non-disastrous consequences.
w w
v-um
--m
,-v
1 IV-14 (e) We do not consider it particularly important to include deferred safety imrcovements in a statement of goals.
(f) Allegedly, great strides toward safety have been made in the design and construction of nuclear plants.
It follows, therefore, that the greatest danger is posed by the ones that currently exist, and the least by those that are being constructed.
There is obviously a need to reverse the present set of priorities, and to give first consideration to the safety of operating reactors, particularly with respect to their potentiality for disaster.
Question 3 (a) Among the approaches to safety goal formulation that are discussed, what approach or combination of approaches is particularly appropriate?
Inappropriate? Why?
(b) Should any other approach be considered? What approach?
Responses AIF considers its proposal an appropriate approach. WPPSS feels that the approach should involve a consensus of informed pro essionals committed to improving safety and that AIF and IEEE goals should be adopted by NRC following review and resolution of concerns.
S&W judges the AIF approach the optimum of those proposed.
1
-r-
4 O
IV-15 TVA emphasizes that the seven safety goal characteristics must be ultimately considered as a whole.
However, in address.ing the seven characteristics listed, the TVA concept of a safety goal would reflect the following:
1.
Both individual and societal goals 2.
See response 2(b) and 2(c) 3.
Ends-oriented 4.
Goals commensurate with other technologies 5.
Both individual and societal goals 6.
Site (or region) independent 7.
Atemporal They believe these characteristics, taken as a whole, are consistent and can provide for a simple, yet effective safety goal.
NYPIRG believes that the approach they outline should take prio'rity over those discussed in NUREG-0764 and NUREG-0739.
Question 4 (a) Among the approaches to dealing with uncertainty that are discussed, what approach or combination of approaches is particularly appropriate?
Inappropriate?
(b) Should any other approach be considered? What approach?
I
e IV-16
esponses AIF thinks that the most appropriate approaches to dealing with uncertainty are the use of a prescribed methodology for modeling and calculation and a restrained use of quantitative criteria.
Use of a prescribed method or choice of methods assures an acceptably advanced approach to risk analysis and consistent treatment of uncertainty.
More importantly, it allows for a meaningful comparison of risk analyses with each other and with the safety goals.
However, although uncertainty can be reduced by development of better analytical methods and expansion of the data base, the probabil.istic nature of risk assessment dictates that uncertainty can never be completely eliminated.
In many cases, the estimated risk value, even with uncertainties, may fall well above or below the relevant safety goals.
In such cases, regulatory decisions may be based on.the PRA studies and safety goals with great confidence.
However, where the best estimate results of PRA studies are near the safety goal value, additional sound engint3 ring judgment must support the regulatory decision process.
WPPSS believes that the PRA for a plant should be carried through using best estimate point estimates for probabilities in order to establish dominant cut sets and dominant sequences in event trees.
Estimates of uncertainties, based on best available data can then be propagated through fault trees and event trees using only the risk outlines and the dominant sequences with appropriate methodology.
~
IV-17 TVA believes that the most appropriate way to control uncertainty is to know where the uncertainty is present.
It can then be controlled by the consistent use of a prescribed methodology for calculation. The use of a prescribed methodology provides a consistent treatment of the uncertainty and allows for a meaningful comparison between risk analyses and compliance with the safety,'als.
Also, by knowing where the uncertainty is present, attention can be focused on its improvement.
The nuclear industry is presently defining acceptable methods of modeling and calculation.
S&W feels that sensitivity analysis should be used in deaiing with uncer-tainty.
The primary concentration in all analyses and data fumulation should be in ensuring' realistic estimates.
Conservatism and use of worst-case scenarios must be avoided if the analyses are to be valid.
NU considers the methodology for establishing a safety goal anct assessing compliance to it very important and a necessarily technological effort which would be more amenable to assessment by the industry than a socio-logical effort. Methodology should'be realistic and not arbitrary, judg-mental, or needlessly conservative.
For instance, an arbitrary assumption that core uncovery equals core degradation or core melt would totally invalidate the technical validity of the safety goal and assessments of compliance.
Conservatisms must be clearly justified, identified, made realistic, and addressed with further work leading to their quantification.
Of the approaches discussed, modeling seems to NYPIRG most untenable (for reasons discussed in the above material on quantitative safety goals); and
,,~
c
k IV-18 the one NYPIRG advocates is evidently the " restrained use of quantitative criteria." Their suggestion to work with confidence intervals is not to be confused with the "high confidence level" discussed under " conservatism":
the latter amounts to substituting an extremely conservative figure wherever the uncertainty can be calculated. The listed objections are overcome if two sets of computations are carried through -- one " conservative" and one
" permissive," between them encompassing a specific confidence interval.
The result would be an envelope within which most cases would be contained, reminding the decision-maker of the uncertainty attaching to probabilistic risk-assessment figures even when their simplifying assumptions are accepted.
Finally, they have many reservations about the " arbitration" methods described, and see more merit in some of the "non-quantitative approaches."
(
Question 5 i
What should be some of the characteristics of safety requirements:
(a) What should be the role of safety-cost trade-offs?
Responses AIF's view, which is described in its proposed approach, is that safety-cost trade-offs should be used in evaluating incremental reductions in risk. A cost benefit ratio should be a secondary % iety goal.
Its role should be in consideration of additional risk reduction after tF r imary safety goals are met.
i l
-4 o
IV-19 WPPSS takes the position that each plant should be required to meet the safety standard.
Safety-cost trade-offs should be used as a decision-making tool to effectively allocate resources to further reduce the risk below the mandatory standard. The cost criterion should be reasonable and commensurate with exemplary criteria applied in other industries.
TVA's answer is the same one given to question 2(d).
S&W holds that safety / cost trade-offs should not be necessary if it is established that a design meets the stated safety goal, except in the case of proposed changes to an existing or already approved design. Any change that would either increase risk without commensurate benefit or reduce risk but only at great cost would need an extensive justification.
The burden for this justification, in the form of safety / cost trade-offs, will fall on the party that proposed the change, whether that be the owner or the staff.
In any safety / cost trade-off, the law of diminishing returns must be kept in mind.
In any incremental analysis, one reaches the point where the added increment of safety or other benefit cannot be justified.
In NU's view, the requirements of a safety goal should be to assess plant safety in a quantitative manner.
Safety / cost trade-offs should be considered by the licensee in selecting the optimum of plant improvements if they are found necessary.
NYPIRG believes the safety-cost trade-offs should be strictly limited to minor, certainly non-disastrous events.
Once the possibility of disaster
IV-20 is securely ruled out, an argument can be made for balancing marginal gains of safety against cost. They warn, however, that if any such guideline as
$1,000 per expected man-rem is adopted, it be indexed to the inflation rate; that particular number was suggested several years ago and has already been considerably devalued.
Question 5(b)
To what extent should benefits of nuclear power, absolute and relative to alternatives, enter safety-requirements decisions?
Responses AIF believes that, as discussed in its position paper, these should be one of the underlying bases for establishing the safety goals.
WPPSS considers the benefits of nuclear-generated electricity the same as those of electricity generated by other sources.
The favorable cost of nuclear generation and the desirability of independence from foreign oil may affect the overall societal decision process of nuclear versus its alternatives, but WPPSS does not feel that these considerations could contribute effectively to either the objectivity or the utility of a safety goal based process.
The safety requirements on nuclear power as they affect the risk to the public can and should be consistent with those in other sectors of the economy.
.e-.
r,.---r m--
+.,,
IV-21 As TVA's second principle suggests, TVA believes that risk criteria must be examined in relation to the risks, cost, and benefits of other technologies now present.
S&W feels that safety requirements must be reasonable in the sense that making them too stringent will result in electricity generation by alternative types of fuel that may very well be less safe than nuclear power. The NRC is empowered only to regulate the nuclear industry, not the entire electrical utility industry.
NU maintains that the benefits of nuclear power should not be of consideration in establishing a safety goal or assessing compliance to it. The benefits of nuclear power will be a result of the establishment of and adherence to a safety goal.
The benefits similarly have no role in the selection of the optimum of plant modifications.if modifications are necessary..This is because the requirement for modification will be based on non-adherence to a safety objective. Adherence to a goal will thus ensure long term b.enefi,ts of nuclear power.
NYPIRG feels that the benefits of nuclear power should be considered only in relation to non-disastrous consequences.
Then, they must be considered in relation to the costs and benefits of alternatives such as investments in energy efficiency.
Ques. ion 5(c)
To what extent is it appropriate for requirements for new and previously approved plants to differ?
o IV-22 Responses AIF believes that the goal should be generally applicable to all plants.
TVA als-feels that the goals should be applied equally.
WPPSS thinks that all plants, new and previously approved, should meet the mandatory safety standards, though not necessarily by utilizing the same design features.
The cost / benefit criteria would be expected to give appro-priate weight to retrofit complications and must account for size, age, operating experience, etc., of currently operating plants.
S&W feels that as effective means to improve safety become known they should be adopted on future plants.
Plants approved prior to that knowledge should not be affected by the safety improvements unless it can be shown through safety /value analysis that the benefits outweigh the costs.
NU affirms that minimum safety goals should be established.and applied to each plant.
In addition, a single comprehensive safety goal should be established and collectively met by all nuclear power plants.
Each plant would have to meet the minimum safety goal.
In order to assure the compre-hensive safety goal is not violated, the aggregate risk contribution from all plants would have to be summed.
As new -lants are added the aggregate risk would increase. Thus, newer plants may have to meet greater safety levels to assure the comprehensive safety goal is not violated.
" Older" plants should not be required to upgrade solely to allow more new plants to be added to the aggregate risk.
i 1
,r.-
IV-23 In NYPIRG's view, it is appropriate for requirements for new and previously approved plants to differ only with respect-to non-disastrous dangers.
In this respect, we notice some pro-industry bias in the comment that, " safety benefits are likely to be... perhaps less than for future plants, in view of lesser remaining operating life and possible side effects..." Nothing is said about the fact that any expected benefits diminish more than pro-portionately to remaining operating life (since capacity factors go down with age) while the increased risk from embrittlement-induced radiaticn, and other effects of aging components increases more than proportionately.
Question 5(d)
Should a safety goal be applied directly to cases in order to attain a similar degree of safety from case to case (even though that may result in specific design and operation requirements differing according to circum-stances)? Or should the goal be applied generically and have r'equirements, rather than estimated degee-of-safety results, be uniform?
Responses AIF cites the discussion in Section IV of its pre ased approach which sug-gests that the initial application of PRA and cuantitative goals in the regulatory process should be to examine the existing body of regulations to establish generically the level of safety which they provide and to identify areas where changes are warranted.
If compi.loce with existing regulations satisfies the safety goal, applicants and licensees who meet existing regulations can be presumed to have met the safety goal.
IV-24 WPPSS maintains that every plant should meat a national safety standard.
The means by which the plant owner implements the safety standard should be left to the owner provided that he is able to show compliance.
Generic cases should be admissible in demonstrating compliance to the NRC, generally at the owner's discretion, recognizing that economics of scale will favor 1
some collaboration.
i The overriding benefit expected from this approach is the owner's internali-zation of the process and the corresponding increase in safety accountability.
The potential impact of plant specific solution on the NRC review role can be controlled by standardizing the process and focussing NRC review in that area.
In S&W's view, the safety goal should be applied separately to each nuclear plant that enters the design process.
Plant design and safety,featurec should only be required to make a plant as safe as another plant. The" should not be required to make plant designs identical from one plant to the next, particularly if it is not warranted from a safety' viewpoint.
~
NU believes that a ninimum, uniform safety goal for cach plant and a compre-hensive aggregate safety goal for all plants should be generically established.
Such goals would undoubtedly be implemented through the differing design and operational circumstances that exist from plant to plant.
- However, there should be no sub-tier safety goal requirements for each specific design or operational variation.
i l
v.
IV-25 NYPIRG's approach to s7fety requires site-specific calculations to determine the disaster potential of each reactor separately.
Question 5(e)
To what extent should the goal reflect protection of individuals regardless of numbers of persons affected, and to what extent should it reflect total, integrated population or societal effects?
Responses AIF's proposed approach reflects its position that there should be dual primary goals, one for the individual (maximum exposed average individual) and one fnr the exposed population.
WPPSS feels that individuals and populations should be protected to a comparable extent.
If it is possible to simplify the safety standard to one quantitative criterion which protects the public and can be shown to also protect (" umbrella") the ind v dual to a comparable extent (or vice i i versa), then such a simplifying approach could be desirable.
4 TVA believes that there should be dual goals, one for the maximum exposed i
average individual and one for the exposed population.
The resources of society over the long term should be optimized, but no individual should bear an unreasonable risk.
S&W takes the position that regardless of whether a goal is stated in individual terms or societal terms, the goal is a societal goal.
The
,,.-,---n
.,-v, nn,-,
--e,,-w.-
--~,--ea,
,,,n,
, -,,,c
,---,n, n,..nw
, -.. ~
.,c n
~
s IV-26 difficulties of predicting dose effects, particularly of low level, must be dealt with in either case.
NU answers questions (e) and (f) by affirming that a minimum, uniform safety goal for each plant and a comprehensive aggregate safety goal for all plants can be established independent of the number of plants affected. This would be a technological goal based on what the nuclear community perceives is necessary to assure safe nuclear plant operation.
The impact of the total integrated population, the equity of distribution of benefits, and the equity of ciistribution of adverse impacts are sociological issues that should not be resolved by the nuclear community, but by the public and Congress.
NYPIRG believes that the distinction between individual and societal risk is widely misunderstood and misused, as NYPIRG feels it was, for example, by former Chairman Hendrie.
In the transcript of the Comission's meeting of July 25, 1980, Mr. Hendrie treated these as alternatives, for example:
~
"You would not ca~re to go with just the societal one and say we don't care about the individuals." Surely there is no contradiction between individual and societal H the latter is defined as the total risk to individuals!
Indeed, with that definition, the individual risk criterion contains no information not contained in the societal, which in no way is an antithesis to it but rather a needed supplement.
The primary safety goals must be formulated in societal terms.
Once that has been done, there may be value to some additional, individually
IV-27 formulated goals; for. example, limits on allowable radiation for workers.
We urge also that limits on allowable emissions of radionuclides be set so as to allow women to conceive and bear healthy children without anxiety about miscarriage or induced defects in communities located near electrical generating plants.
Since the growing embryo is far more vulnerable to radiation damage than an " average individual," the adoption of such an individual risk safety goal may require more remote siting than is presently allowed.
Question 5(f)
To what extent should equities of distribution of benefits and adverse impacts influence requirements?
Responses AIF believes that providing a safety goal which limits risk to the maximum exposed individual to a small increment of existing risk should eliminate the need to consider distribution of benefits and adverse impacts.
In the view of WPPSS, the equities of distribution of benefits and adverse impacts are partly accommodated by having an individual goal as well as a population goal.
Any other accommodation should be external to the safety goal, i.e., special rates for plant neighbors, utilities' purchase of nearby real estate and resale to individuals who understand and accept the incremental risk.
i.
IV-28 TVA feels that attempts to creatt equity will be difficult due to the complex 4
inters.ctions as well as changing conditions and attitudes; easing of clear.ly defined inequities should be attempted, i
i
.S&W considers it virtually impossible to balance benefits and costs to
-individuals from a regulatory point of view. Whatever pressures drive people to live where they live, work where they work, or invest as they invest, are far too complex to be dealt with in this framework.
NYPIRG is not impressed by the argument that persons living near nuclear plants should be subjected to greater r'isk because they derive some benefits.
l The presumed benefit is the claim that nuclear electricity costs less than that generated in other ways -- at best, a small, marginal economic benefit I
hardly to be weighed against years of suffering from a radiation-induced 4
illness, even if non-fatal and wholly compensated -- and no one who has contracted such an illness has yet had his medical expenses paid by a utility.
1 In fact, however, the alleged economic benefit is highly controversial and most likely illusory:
nuclear power has been so extensively. subsidized, and in so many indirect ways, that if there was truly unbiased cost accounting, the claimed benefit would disappear and be replaced by its opposite.
- Arguably, those who receive the use of_the electricity (another meaning of the word i
" benefit") should pay the costs, which should not be imposed on others.
If these customers should in addition be required to run greater risk of illness, ear.ly death, or having defective descendants, then the stockholders and managers of the utilities should be required to live within a 10-mile radius of any nuclear plant from the operation of which they gain financially.
1
IV-29 Question 5(g)
_Should the safety goal reflect increased aversion to risk of high consequences even at low probability?
Responses In AIF's view, decision rules should be based on an objective estimate of risk, and the use of risk aversion factors should be avoided. Although our society may ultimately require some alteration of the risk criteria to account for risk aversion associated with nuclear power, it is premature and inappropriate to judge the extent of this alteration now.
In fact, society's perception of risk is influenced by many factors, is different from different types of risks, and is constantly changing.
Other established societal mechanisms should be used to decide on the appropriate application of risk aversion, and any attempt on the part of technologists,to do so will merely add to the difficulty in obtainirig public understanding.
Risk aversion factors other than unity will bias cost-benefit evaluations towards emphasis on low probability-high consequence events, a bias which the Kemeny and Rogovin reports stressed should be avoided.
Such risk aversion factors will also bias comparisons with other technological risks.
WPPSS feels that the goal should definitely not reflect increased av2rsion to risk of high consequences at low probability. WPPSS considers risk aversion a subjective and uncertain factor which would detract from the technical basis and overall objectivity of the goal process.
Risk aversion has no application in any other regulated societal risk.
_ _ _ _ _ _ _ _ ~
i IV-30 TVA thinks risk aversion is an extremely difficult concept to define and should not be attempted in formulation of these safety goals. Aversion factors would provide unnecessary emphasis on low probability, high conse-quence events and would further bias comparisons with other technologies.
In S&W's view, risk aversion is basically an added conservatism created by the perceptions of whoever creates the risk aversion factors a.'d, as such, should not be included in the safety goal.
NU-believes that the safety goal should limit the product of probabilities times consequences of accidents that can result in public risk.
- Thus, events that have higher probability or higher consequences than others, will obviously contribute a greater percentage to the total risk.
Such higher level events should not be excessively weighted, since that would distort the proportional significance of lower level events (which may aggregate a substantial portion of the total risk).
NYPIRG does not consider the ACRS compromise proposal (to weight conse-quences exponentially) at all satisfactory as a substitute for the categorical exclusion of threats to the survival of society.
By failing to make this critical distinction, considering separately those aversive consequences that do and that do not imperil security or survivability (here called disastrous and non-disastrous), the ACRS confuses the issues.
If " dams or large quantities of hazardous chemicals" do in fact pose dangers of damage from which a society might not be able to recover, and if that hazard were generally known, they would be prohibited.
,---,-,.--<.e
,_m--.,w-c--
,.,_c,,
,y-w -,
y..
j IV-31
\\
Question 5(h) i What is the proper balance between stability of requirements and flexibility J
for modification as knowledge develops and insights change?
l Responses AIF argues that the safety goal formulations should provide a stable frame-work for ensuring adequate protection of the public health and safety while providing flexibility in meeting these goals by the most cost effective approach to designing and operating the plant. More rigid interpretations of the safety goals than those provided in the attached position paper would not allow sufficient flexibility to incorporate modifications as knowledge develops and insights change.
WPPSS holds that the ability of American utilities to provide customers with economical electric power is influenced significantly by stability in the licensing process.
Therefore, stability of the requirements is essential.
The safety goal should remain stable.
The goal based process WPPSS envisions would include flexibility, principally stentning from improvements in the evaluation technology, assessment of operating experience lessons, and application of the cost / benefit criteria.
The objectivity of the goal process and the consideration of cost / benefit should ensure that any required changes are appropriate and accepted.
TVA takes the position that the goals should be based on the effects on man and society which are constant and should not be based on particular designs or other factors which are subject to change.
~~
IV-32 S&W maintains that once an acceptable level of safety has been achieved, regulatory stability is essential.
Only new knowledge of clear and significant safety problems should be cause for imposition of new requirements.
AU asserts that once a safety goal is established, it should not become self-ratcheting if it is discovered that plants can meet it.
Further improvements beyond those necessary to meet a safety goal should be viewed as meeting other objectives, such as protection of investment. This is 1
because a safety goal will not be established below that level necessary to assure proper protection of the health and safety of the public.
1 In NYPIRG's view, as a general rule, the balance is tipped heavily in the direction of stable rules. They, therefore, urge that special provisions be included for the modification of all applicable rules and the up' dating of their application to specific plants as new knowledge is developed.
In a way, the present crisis of nuclear power came about because of insufficient adaptiveness to emerging information about its dangers. William Lowe made the first public presentation of the results of his and Edgar Mendelsohn's recomputation of the yield of,the Hiroshima and Nagasaki bombs on May 31, 1981. The recency of this event, which may bring about a doubling or quadrupling of risk figures for cancer from gamma irradiation, reminds us that our ignorance about low-level radiation effects is vast, and such We informed pessimists as John Gofman may yet prove to have been correct.
must be prepared to adjust standards promptly, erring on the side of caution.
1 r
IV-33 Question 6 (a) How should the stringency of nuclear-power-plant safety requirements co.:. pare with current practice?
(b) How should stringency of the safety goal compare with risks accepted from other (non-nuclear) electrical energy sources and with risks arising in various other contexts?
Responses AIF contends that since the purpose of safety goals is to provide a level of protection for individuals and society as a whole, the stringency asso-ciated with nuclear power plant safety goals should be comparable to risks from other non-nuclear electrical energy sources.
Once safe., goals are established, current requirements should be evaluated against the safety Ocels to determine need for changes.
t WPPSS believes that the stringency of present nuclear power plant safety requirements is now largely directed at " process-oriented" goals, but should be focussed on "end-oriented" goals, specifically, quantit3tive safety goals.
Methods and procedures for achieving and demonstrating com-pliance with a national safety standard would then receive proper attention.
Safety goals and the evaluation of plant designs with respect to the goals constitute a new technology relative to the basis for current practice.
Accordingly, WPPSS would expect that some new design requirements would
^
i IV-34 follow from application of the technology and that some current safety requirements would not be justifiable with respect to the goal.. These perspectives should be given significant weight in developing future safety requirements.
Overall, WPPSS would not expect the result to be particularly more or less stringent requirements; rather, we would expect the requirements to be more effective in ensuring plant safety. WPPSS cites its response to question 5(b) as applying also to 6(b).
TVA holds that since the purpose of the safety goals is to provide a level of protection for individuals and society as a whole, the stringency asso-ciated with the risk from nuclear power generation should be comparable to the stringency associated with the risk from other sources.
In S&W's view, the stringency of requiremants is no guarantee of performance, and may, in fact, be counterproductive of real safet, A better un'derstanding of real processes, and ultimately of risk, pluc more consistent and intelligent enforcement where the risk is evident to all, can be more productive.
There is no codpelling reason for nuclear power to be safer than other human activities of comparable benefit. The extent to which nuclear power is, in f'act, safer than other activities is evidence that current requirements on nuclear power are too stringent.
In this regard, two issues not explicitly addressed above are important.
First, goals should be related to the fact of safety, and not to the perception of safety.
The extent to which nuclear power is perceived as more dangerous or less beneficial than it is, is an education problem; and this problem of education -- of owners, regulato:s, and the public -- cannot be over emphasized.
Second, protection of an
-v--
- O IV-35 investment in plant and equipment is an effective driving force for improve-ments in safety. An intelligent owner will.not risk his investment by ignoring other features or improvements that could mean the future financial existence of his company.
NU feels that current safety requirements should be supplemented with safety goals, and each plant should be required to meet c.uinimum uniform safety goal.
If a plant meets the minimum safety goal then any decisions for further safety improvements should be up to the l$ t,see and may be based on cost / benefit considerations.
If a plant does not meet the minimum safety goal, then the time frame for required upgrz. ding should be based on the degree of deviation.
If a plant deviates by less than an order of magnitude
~
from the minimum safety goal, then considerable time should be allowed for improvement such as months or years.
Larger deviations should be corrected on a shorter time frame, Although the nuclear community has not and should not proclaim what risk is acceptable to the public individual, there is a responsibility for the nuclear community to establish what it believes is safe. This has been done in a qualitative manner thus far, by establishing what level of safety requirements, safety features, and redundancy is necessary to license a plant for operation.
It now remains necessary to quantify a safety goal and to quantify the achieved safety level for each plant.
This should be done based on what the nuclear community perceives is necessary to ensure the continued long term safety of nucle v power plant operation.
If nuclear
~
power is to endure in the eyes of the public, then safety goals and m
IV-36 adherence to them should establish nuclear plants as being an insignificant societal risk contributor compared to other societal risks.
(a) NYPIRG believes that current practice is far too permissive and that much more stringent safety requirements are urgently needed.
(b) NYPIRG affirms that in principle, one might say that safety goals for nuclear power should be no more and no less stringent than for other ways of generating power, but with the following understanding: that the relevant comparisons must take into account not what is popularly believed, but, rather, the best available information about the dangerousness of an activity. But stringency is not really the issue, when one attempts to evaluate different energy sources, since the dangers of radiation cannot be directly compared with those of breathing high levels of S02 or of being struck by a windmill blade.
It is necessary to bear in mind the fact that nuclear dangers are particularly aversive to the public because of the unique nature of these perils, and they are accordingly much more likely to give rise to chronic anxiety and stress, itself a deleterious and most undersirable side-effect of nuclear power to which too little attention is' paid.
Question 7 The Commission invites comments on the ACRS proposal and on the other specific proposals described in the paper, and would welcome any alternative proposals or suggestions.
A
IV-37 Responses AIF's comments include an attached positica' paper which presents its proposal.
TVA considers the ACRS proposal a good starting point for further discussion and feels. that it contains most of the essential principles for a safety goal. TVA believes that removing some of the complexity from the ACRS proposal and establishing a simpler safety goal would help public perception and understanding of the goal.
bCSconsiderstheproposals~inNUREG-0739thoughtfulandprovocative.
d e
r-+-
= ' -
- 8"r""N'-"-
--~* --
+=e"*'"--
9w'v>~
e rv-v' v'--y
-w-f-r~WT-r9 7
w-W7='r-"'
T v'-'
V.
SCHEDULE OF SAFETY GOAL DEVELOPMENT AND SEQUENCE OF OTHER NRC ACTIVITIES Most of the commenters had some view on the NRC schedule for development of a safety goal and the need for a goal to be available in time to influence other NRC activities. Marvin Lewis considers the period for comment on NUREG-0764 inadequate.
Ten commenters state that a safety goal should be developed promptly; eight of the ten also believe that it should precede and guide other NRC rulemakings, especially the degraded core rulemaking.
The Atomic Industrial Forum considers it imperative that the Commission adopt a specific set of quantitative safety goals by the end of 1981 in order to provide timely resolution of several pending issues such as the degraded core rulemaking, minimum engineered safety features rulemaking and siting criteria rulemaking.
CE, Bechtel and Westinghouse concur with AIF. WPPSS, Duke Power and Northeast Utilities express similar views citing only the degraded core rulemaking.
General Electric advocates prompt safety goal development to guide the various rulemakings by establishing a logical basis on which rulemaking conclusions could be founded.
TVA thinks a safety goal should be developed in a timely manner to ensure that plants meet it and that the health and safety of the public are adequately protected.
e m
w v --
V-2 Kerr-McGee feels that safety goal development should be accomplished as expeditiously as possible and not be constrained by spiralling levels of buraucracy.
The American Mining Congress, whose special concern is uranium milling, believes that, if the Commission intends to apply the safety goal to elements of the fuel cycle other than reactors, it should define those elements and consider the distinctions among them at this time so that affected parties are afforded due process and can participate in development of the goal.
I e
.. ~,. -, -
VI. OTHER COMMENTS Norman Buske of Search Technical Services believe< that the objective of a safety goal diverts attention from a needed safety " standard." He also thinks that states should set their own overall safety standards to obviate NRC's " potential conflict of interest between the Congressional mandate to maintain a viable nuclear industry" and the cost of safety requirements NRC's lack of responsiveness to the benefits and the risk costs of nuclear power which are felt locally.
He believes that the standards may have to be boiled down to a single number to be amenable to public consideration and he proposes two possible ways of ex. pressing a goal: the first, along the lines suggested by Robert Bernero of NRC, an acceptable probability of severe core damage per gigawatt-year and the second, an upper limit on the expected cost of all accidents per kilovatt-hours generated.
Kerr-McGee Corporation holds that separate safety goals are needed for low probability, high risk events, on the one hand and accidents or minor events in normal ~cperations on the other.
Kerr-McGee also suggests ways of using risk assessment in development of a safety goal.
Westinghouse believes that goals must be high level quantitative goals and address both individual and ' actual risks.
Cost-benefit considerations should apply even though the individual and social limits are satisfied.
Efforts to partition or allocate the top level goals to features, sequences, or systems must be left to the flexibility of the regulated industry if innovative approaches to achieving safety rre to be encouraged.
In this regard Westinghouse believes the conditional goals (large scale fuel melt)
VI-2 of the NUREG-0739 reference discussion goals are not warranted.
Regarding single versus multiple goal forms, it appears appropriate to con-sider multiple goal forms such as addressing individual, societal, economic, or melt frequency risk limits to address the various damage characteris.'cs under consideration.
However, multiple levels of a particular goal form such as the two levels (goal level, and upper limit) of the NUREG-0739 proposal are unnecessary and detract from the desirable characteristics of simplicity and understandability of a safety goal.
Westinghouse believes that numerical levels for safety goals should be established at levels which are reasonable from the standpoints of being achievable and consistent with other activities involving risks, and defendable from the standpoint that risks to individuals and society as a whole are controlled to be small.
The goals proposed by AIF in "A Proposed Approach to the Establishment and Use of Quantitative Safety Goals in the Nuclear Regulatory Process" (May, 1981) achieve these criteria, and Westinghouse supports that document as a viable industry position which NRC should care-fully consider in its deliberations to arrive at a safety goal.
Westinghouse also believes that the numerical limits should be as simple as possible to demonstrate compliance, and therefore encourages number limits as opposed i
l to lines or curves.
Finally, Westinghouse considers the NUREG-0739 treat-ments of risk aversion factors and the distinction between early and latent l
fatality risk goal levels as unnecessary numerical complicatons to the l
formation of risk-based safety goals.
a.
VI-3 With regard to approaches to dealing with uncertainty, Westinghouse believes that a combination of modelling and data stardardization can aid by leading to consistency and predictability. Analyses to show compliance with a safety goal should be performed using realistic assumptions for data and phenomenology.
Explicit treatment of urcertainties may be a worthwhile requircment after confidence in the basic application of the techniques is gained.
NUREG-0764 raises a number of interesting considerations with respect to criteria for assessing an approach to a safety goal, characteristics, verifi-ability, considerations and approaches 'to safety-goal formulation. Westinghouse believes it is worthwhile that these have been identified'for consideration.
However, many of these areas appear to be of a nature that perfect resolution would never be achieved.
This should not be permitted te dilute and retard the NRC objective of arriving at a reasoned quantita'ive stfety goal in a timely fashion.
Dr. Henry Hurwitz states that an NRC safety goal should avoid the hyprocrisy of establishing radiological standards for nuclear reactors that are not being applied to energy efficient dwelling.
,n.,---
,,.--.m,,
,