ML20204G175
| ML20204G175 | |
| Person / Time | |
|---|---|
| Issue date: | 07/15/1986 |
| From: | Asselstine J NRC COMMISSION (OCM) |
| To: | Walske C ATOMIC INDUSTRIAL FORUM |
| Shared Package | |
| ML20204G176 | List: |
| References | |
| NUDOCS 8608070094 | |
| Download: ML20204G175 (9) | |
Text
4 hk
- %*o UNITED STATES NUCLEAR REGULATORY COMMISSION o
3 3
WASHINGTON D.C.20666 July 15, 1986 OFFICE OF THE COMMISSIONER Mr. Carl Walske, President Atomic Industrial Forum, Inc.
7101 Wisconsin Avenue Bethesda, Maryland 20814-4805
Dear Mr. Walske:
Thank you for your letter of May 29, 1986, regarding my May 22, 1986 testimony before the Energy Conservation and Power Subcomittee of the House Committee on Energy and Commerce.
In your letter, you expressed concern that my statement may have been misinterpreted in the public arena.
In particular, you point to my statement that "... given the present level of safety being achieved by the operating nuclear power plants in this country, we can expect to see a core meltdown accident within the next 20 years, and it is possible that such an accident could result in off-site releases of radiation which are as large as, or larger than, the releases estimated to have occurred at Chernobyl." You state that the Atomic Industrial Forum does not agree with my characterization of the likelihood of a core meltdown in this country within the next 20 years, and it is the position of the AIF that, even if there were to be a core meltdown, the probability of a substantial release of radioactive materials is very low (i.e., one substantial release in 200 core meltdowns).
I stand by my statement before the Energy Conservation and Power Subcom-mittee.
I believe that it represents an accurate and balanced assessment of the risk posed by the 100 operating nuclear powerplants in this country.
I have provided my rationale for the views contained in my statement before the Subcomittee in various forums in the past. However, since you have taken issue with my statement I want to take this opportunity to explain my position in detail.
I share your concern for accuracy.
I recall reading in the newspapers in recent months statements by senior officials within the nuclear industry that our plants are " perfectly safe" and we "will not have a Chernobyl-type plant accident here." Apparently, such absolute statements are thought to be needed to counter-balance arguments from the other side that there is an immediate threat to the public which requires the shutdown of our nuclear plants.
In my view, neither position is accurate. To convey an impression that Chernobyl-type releases are impossible in this country is as inaccurate as conveying an impression that a similar disaster is a certainty.
I attempted to take the middle road in my opening statement before the Subcommittee. We do not fully understand the risks of nuclear power, and we should not be fearful of saying so.
Your letter contained a number of specific criticisms of my statement.
First, you stated that the NRC staff's 45 percent core meltdown estimate over the next 20 years does not take into account safety inprovements now e600070094 e60715 PDR COMMS NRCC CORRESPONDENCE PDR
4~,
being developed and others which will be forthcoming.
I agree. However, that estimate also does not include all contributors to the probability of a core meltdown.
For example, it does not accurately reflect the contributions to risk from external events such as earthquakes and floods.
More importantly, it does not properly account for human errors or degradation in the material condition of the plant. The performance of existing plants demonstrates that these weaknesses in probabilistic studies may result in a significant underestimate of the actual risk.
Specifi-cally, we are learning that the plants often react in ways we do not expect. As Harold Denton, the Commission's chief safety officer, wrote:
"I believe that the recent Davis-Besse event illustrates that, in the real world, system and component reliabilities can degrade below those we and the industry routinely assume in estimating core melt frequencies." (See, Memorandum from Harold R. Denton to William J. Dircks, dated June 27, 1985.) Thus, my views do not rest only on the 45 percent estimate or on the theoretical calculations of IDCOR which you reference. Neither takes into account the large uncertainties in these theoretical estimates and neither accurately reflects the actual operation of the plants in the real world.
Recent operating experience, including the several serious operating events at U.S. nuclear powerplants in 1985, indicates that inadequate or improper maintenance, surveillance testing errors, equipment failures, design inadequacies, and operator and other personnel errors are occurring at U.S.
plants at an unacceptably high rate and that they are significant contributors to operating events that can lead to severe accidents. This operating experience shows that these contributors are causing the total loss of one or more safety systems and multiple equipment failures at plants that can substantially erode defense-in-depth and lead to accident conditions beyond'the design basis of the plant.
One would hope that we are aggressively pursuing the root causes of these occurrences. Unfortunately, it does not appear that all U.S. nuclear utilities are learning the lessons of experience. Our Office for Analysis and Evaluation of Operational Data (AE00) conducted a survey in the fall of 1984 to determine how well licensees were learning the lessons of experi-ence. AE00 concluded:
"Most plants are making moderate, not extensive, use of their in-house operating experience, and in general are making less j
use of the large body of knowledge associated with events and concerns that originate elsewhere in the industry."
(See, "1985 Annual Report /AE00 5601," April 1986, p. 5.) This reinforces a previous AE00 report which found that our licensees often repeat the same mistakes, even at the same plant. The actual operating experience of our existing plants and the industry's failure to heed the lessons of experience indicate, in my l
judgment, that we can expect to see another serious accident in this j
country during the next 20 years.
l In your letter, you emphasized that risk is not eouivalent to core melt probability.
I agree. You went on to state that it is not technically correct to say that any core melt accident at a U.S. reactor would yield l
l Chernobyl-like consequences, which you said my statement implies. However, t
r t
t
~ - -
__-_-_m.-._-_._--,m__--
you quoted only a part of my statement. What I said just before the statement quoted in your letter was:
Third, although we believe that all of our reactors have some capability to withstand severe core meltoown accidents, the extent to which they can withstand such accidents depends upon the sequence of events during the accident, the individual plant designs and the manner in which each plant is operated and maintained. While we hope that their occurrence is unlikely, there are accident sequences for U.S. plants that can leaa to rupture or bypassing of the containment in U.S. reactors which would result in the off-site release of fission products comparable to or worse than the releases estimated by the NRC staff to have taken place during the Chernobyl accident. That is why the Commission told the Congress recently that it could not rule out a commercial nuclear power plant accident in the United States resulting in tens of billions of dollars in property losses and injuries to the public.
Thus, my statement made the point that not all core meltdown accidents can be expected to result in large offsite releases of radiation which can harm the public and contaminate large areas of land and property. The central questions, of course, are: how likely is such an accident, what are the uncertainties in estimating the probabilities, and how well do we under-stand this risk? Your letter can be interpreted very easily by the uninitiated to say that the reactor risks are well understood and that an accident involving substantial and harmful releases of radioactivity to the environment is all but impossible in this country. That clearly is not an accurate representation of the facts.
Your letter stated that "With our reactors IDCOR does not find any such releases as serious as Chernobyl apparently was."
I question whether there is a sound scientific basis for this conclusion.
The 1975 Reactor Safety Study (WASH-1400), which the industry and the NRC touted as an objective assessment of reactor risk, contains several release categories associated with core meltdowns that are equal to or greater than our estimates of the releases at Chernobyl. The NRC staff has recently advised the Commission that the best available information suggests that some changes in specific radionuclide group releases to the atmosphere are justified; however, the overall consequences are not significantly different from those using source terms contained in the Reactor Safety Study. Thus, the best available information indicates that severe accidents with Chernobyl-type releases, or worse, can occur at U.S. plants.
The question then becomes: how likely are such accidents and what are the uncertainties in estimating their probabilities?
In my view, two con-clusions regarding the Reactor Safety Study are germane to this question.
First, the uncertainties in reactor risks are much larger than estimated in that 1975 report, even with all of the research and analyses that have been completed since then. Second, the bottom-line results of quantitative probabilistic risk assessments are not reliable.
I thought there was a general recognition of these conclusions, but your letter seems to indicate a belief that we can now make sweeping generalizations about the low
i I e i
likelihood of a large-scale radiation release for all U.S. plants.
In doing so, you seem to give no weight to the American Physical Society Study Group on Radionuclide Release from Severe Accidents at Nuclear Power Plants which concluded:
"It is impossible to make the sweeping generalization that the calculated source term for any accident sequence involving any reactor plant would always be a small fraction of the fission product inventory at reactor shutdown."
(See, R. Wilson et al., Reviews of Modern Physics, Vol. 57 No. 3, Part 11, July 1985--p. S128.) The funda-mental issues raised in that report have not been resolved in a scien-tifically defensible way. Those issues involve factors such as the chemical form of iodine during a severe accident, variations in containment performance due to design and construction differences, and the potential for steam explosion, both within the reactor vessel and within the containment. The resolution of each of these issues has a direct bearing on the potential for a large-scale early release of fission products in the event of a severe accident.
With regard to the chemical form of iodine, the industry has argued that
~
during a severe accident iodine can be expected to join with cesium to form cesium iodide, which reduces the potential for harmful releases of volatile iodine.
Yet, recent experiments have resulted in the production of volatile iodine and have not shown extensive cesium iodide formation.
For this reason, the NRC staff has recently written:
" Based on the experi-mental evidence available today, a definitive position regarding the chemical form of iodine would be premature. At this time, it it not obvious what phenomena are causing specific experimental results." (See, Letter from R.B. Minogue to John J. Taylor of EPRI dated May 22,1986.)
With regard to containment performance, there is considerable evidence that containment strength may vary substantially from plant to plant based upon differences in design and construction. These differences effectively rule out broad generalizations regarding containment performance.
As senior members of the NRC's Office of Nuclear Regulatory Research put it in describing a recent series of tests:
"The lessons learned from the steel tests was that even minor details of stiffening ring attachment made a large difference in ultimate capacity. This means that individual construction details could lead to a large variation, site-to-site, in ultimate capacity (of the containments).
' Rules of Thumb' are probably out of the question."
(See, Trip Report from D.F. Ross, G.H. Marcus, and C.N.
Kelber to Robert B. Minogue dated February 3, 1986.)
With regard to steam explosions, the industry has argued that there is little potential that such explosions could lead to substantial releases based on predictions of fragmentation of the molten core upon emersion in water. However, our researchers at Sandia National Laboratories have not agreed with these predictions, noting, among other things:
"A detailed examination of FIST data to date shows no match between the fragmentation predictions of Fauske, Corradini, or Saito-Theofanus, with regard to debris size or distribution. Mismatch is at least an order of magnitude, showing these theories are missing some key ingredients."
(Id.) For these and other reasons, our researchers have not ruled out steam explosions as a potentially significant phenomenon which could lead to substantial
radiation releases. Given the best available scientific information to date and recognizing the substantial uncertainties which still exist regarding these issues, I believe we are still a long way from making defensible generalizations about releases from core meltdowns.
You stated that the implication of my statement " exaggerates the risk from U.S. reactors by at least a factor of two hundred." However, your assertion fails to take into account all potentially significant.
contributors to risk, all potentially significant core meltdown phenomena, a reasonable range of technically defensible parametric values for calculating containment performance during core meltdowns, all potentially significant accident sequences, the effects of human error or design and construction errors, the effects of materials degradation with age, and significant operating events, including so-called precursors to core meltdown accidents.
According to WASH-1400, there are many accident scenarios that can lead to substantial releases, including a small break loss of coolant with failure of the containment sprays, an interfacing systems loss of coolant (i.e., an accident involving overpressurization of low pressure piping that is outsida of the containment but is connected to the high pressure primary cooling piping such that the loss of coolant occurs outside of the con-tainment rather than the design basis loss of coolant inside containment),
anticipated transients without scram, station blackout, and loss of coolant accidents with failure of emergency core cooling injection. The specific release category that results from these scenarios is dependent on core meltdown phenomena and containment response thereto. While much progress has been made in understanding these accident scenarios since WASH-1400 was published in 1975, there remain very substantial uncertainties in evalu-ating them.
For example, during a core meltdown, theoretical source term calculations include models for plating out of significant quantities of y
fission products within the primary system. However, the models do not evaluate, or poorly evaluate, the effects of the heating of the primary system by the plated-out fission products to determine whether this phenomenon alters the sequence of events and the release category. As I mentioned before, steam explosions and their effects on containment and resuspension of fission products are still in dispute. These are just two examples of the many uncertainties and unknowns regarding the release categories which could result from various core meltdown sequences. With regard to the likelihood of the various sequences, for the reasons given above, I would say that none of the sequences can be ruled out. A number of precursor events have occurred at U.S. reactors for each of the above scenarios.
The broad conclusions in your letter seem to be based substantially, if not exclusively, on the IDCOR program. Unfortunately, that program examined only a few plants. The nuclear industry eschewed standardization in such areas as plant design, construction, operations, maintenance and surveillance testing. Thus, each operating plant has its own unique vulnerabilities to core neltdown accidents and to substantial releases of radioactivity. This fact, together with the substantial uncertainties inherent in these types of theoretical analyses and the limited number of
. 9 accident sequences considered, make extrapolation of the IOCOR results to all plants premature at best.
Given the limits of our understanding of severe accident phenomena and the large uncertainties inherent in attempting to predict the likelihood that a core meltdown will proceed to a large and catastrophic radiation release, I reach the same conclusion as did the President's Commission on the Accident at Three Mile Island.
In the words of the Kemeny Comission:
Whether in this particular case we came close to a catastrophic accident or not, this accident was too serious. Accidents as serious as TMI should not be allowed to occur in the future.
The accident got sufficiently out of hand so that those attempting to control it were operating somewhat in the dark. While today the causes are well understood, 6 months after the accident it is still difficult to know the precise state of the core and what the conditions are inside the reactor building. Once an accident reaches this stage, one that goes beyond well-understood principles, and puts those controlling the accident into an experimental mode (this happened during the first day), the uncertainty of whether an accident could result in major releases of radioactivity is too high. Adding to this encrmous damage to the plant, the expensive and potentially dangerous cleanup process that remains, and the great cost of the accident, we must conclude that -- whatever worse could have happened -- the accident had already gone too far to make it tolerable.
While throughout this entire document we emphasize that fundamental changes are necessary to prevent accidents as serious as TMI, we must not assume that an accident of this or greater seriousness cannot happen again, even if the changes we recommend are made. Therefore, in addition to doing everything to prevent such accidents, we must be fully prepared to minimize the potential impact of such an accident on public health and safety, should one occur in the future.
" Report of the President's Commission on the Accident At Three Mile Island," p. 15.
That is why I have advocated a program of new initiatives aimed at both reducing the likelihood of core meltdown accidents and minimizing the l
potential for a large offsite release should such an accident occur. These new initiatives would build upon, but would go beyond the NRC's existing regulatory programs and the self-improvement programs undertaken by the industry in recent years. My proposal consists of three new initiatives for the current generation of plants.
First, we should undertake a detailed reexamination of each U.S. plant to identify and correct design weaknesses and vulnerabilities which can initiate or complicate serious accidents. To be effective, this effort must go beyond the surrogate plant approach advocated by the industry in the 10COR program. What is needed is a thorough and independent review of the
design of each plant, including a verification of the adequacy of the existing design basis for the plant and a review of all changes made to the plant after the approval of the plant's original design basis. Given the absence of standardization in the U.S. nuclear program and the lack of good configuration control at some plants, this step is necessary to assure that all significant design problems are identified and corrected.
Second, we should undertake improvement programs in areas of demonstrated weakness in U.S. nuclear powerplant operations, including management, human performance, equipment reliability, and maintenance and surveillance testing. Despite the existence of voluntary industry efforts in several of these areas, we are still seeing U.S. plant performance that is substan-tially below the levels of safety and reliability being achieved in other countries such as Japan, Sweden, and West Germany.
U.S. operating experi-ence demonstrates that existinn voluntary efforts simply are not doing the job. We need expanded efforts in each of these areas sufficient to ensure a level of performance at U.S. plants which is equal to that now being achieved in these other countries. Of these areas, it appears that management is the dominant factor in achieving excellence in performance.
We need to focus particular attention on those plants with a history of poor operating performance and reliability.
The industry's Institute of Nuclear Power Operations (INPO) has been in operation now for more than six years. Although INP0 has had a positive effect in improving overall industry performance, there are still far too many plants that fail to meet acceptable standards of performance. This indicates either that strong peer pressure within the industry is still not being applied to the poor performers or that peer pressure alone is not sufficient to bring about effective and lasting improvement.
In either case, further regulatory initiatives are needed, especially for the weak performers.
In addition, those members of the industry with more expertise and better performance should provide more help to the weaker performers. The industry itself must become more aggressive in ensuring exemplary performance of all nuclear utilities.
After all, the future of the best managed facility may rest in the hands of the worst managed.
I want to emphasize that I am not seeking perfection in U.S. nuclear power plant operations. What I g seeking is a level of operational performance by the U.S. plants that equals the level of performance being routinely achieved by the plants in such countries as Japan, Sweden and West Germany.
I am convinced that this is an achievable objective, and we in government and you in the industry should dedicate ourselves to meeting this goal within the next three years.
1 Third, we should undertake a detailed study of additional design features, such as a dedicated decay heat removal system and a filtering / venting system for containments which have the ability to reduce substantially the likelihood of a core meltdown and the potential for a large off-site release of radioactivity.
Such design features have already been installed or are being actively pursued by several European countries with aggressive nuclear programs. These design improvements for existing, as well as for future plants, are being accomplished in a disciplined manner at reasonable cost. We should, therefore, give specific attention to those designs which already exist or are under active development in other countries. Any such
o.
features would not necessarily have to satisfy all of the Commission's requirements such as the single failure criterion since they would serve as a final backup for existing plant safety systems.
In my view, these three initiatives would bring about fundamental improve-ments in the safety of U.S. nuclear powerplants that would enable the optimistic safety performance projections expressed in your letter to be realized.
I believe that we both share a common objective:
to assure a safe and reliable nuclear power program in this country.
I therefore invite you and the other leaders of the industry to join with me in a new commitment to safety -- a commitment that will ensure the successful, long-term operation of the plants we now have and the continued avail-ability of the nuclear option for the future. That commitment can best be achieved by a cooperative safety approach which takes advantage of the industry's knowledge and experience but which also recognizes the need for, and legitimate functions of, regulation.
I propose an approach whereby the-NRC and the industry would work together to define the specific objectives of each of the three safety initiatives I have outlined and the detailed requirements needed to achieve those objectives. Under this approach, the industry would be free to take the initiative in proposing for discussion creative solutions in each of the areas I have identified. However, this joint effort would result in a binding comitment by the industry to meet specific new requirements. Those commitments would then be subject to NRC inspection and enforcement. The approach I am suggesting is quite similar to that used in many foreign countries with successful nuclear programs and builds upon the voluntary self-regulation approach advocated by the Nuclear Utilities Management and Human Resource Committee (NUMARC).
In the wake of the Chernobyl accident, I believe that nuclear power is at a crossroads in this and other countries. We have the opportunity to learn the lessons of experience, to correct the mistakes of the past, and to bring about lasting improvement that will ensure the accident-free operation of our plants over their remaining operating lives. We had that opportunity following the Three Mile Island accident but we failed to follow through.
I sincerely hope that we do not have to wait for another nuclear accident before we come to grips with the root causes of nuclear power risks. The failure to do this during the past twenty years of commercial nuclear experience involving large power reactors is, in my view, the fundamental reason why nuclear power is controversial and will remain controversial until a systematic approach to safety is taken. And, the failure over the last twenty years to come to grips with the root causes of the risks is why I hold the views I expressed at the Congressional hearing. An essential first step toward correcting the mistakes of the past is to acknowledge the obvious:
that the public and the Congress will not tolerate, and the industry and the NRC cannot allow, another severe accident as serious as the Three Mile Island accident or worse. The second step is to undertake the new initiatives needed to make this objective a reality.
I have attempted in this letter to outline what more we need to do and why I believe we need to do it.
I suggest that we use this exchange as a foundation on which to build a truly effective safety improvement program,
a program that will assure the long-term protection of the public and that will restore public confidence in the NRC and in nuclear power.
Sincerely, A..
foa-,x.,,seistine O
l
-