GRETA JOY DICUS, COMMISSIONER
U.S. NUCLEAR REGULATORY COMMISSION
LUNCHEON KEYNOTE SPEECH
7th INTERNATIONAL SYMPOSIUM
CURRENT ISSUES RELATED TO NUCLEAR
POWER PLANT STRUCTURES,
EQUIPMENT AND PIPING
DECEMBER 2, 1998
TITLE: IN CONSIDERATION OF RISK
Good afternoon. Dr. Gupta, ladies and gentlemen. Before I begin, I'd like to thank our host for his wonderful hospitality, the good weather, and of course, the opportunity to address this audience who specializes in what is viewed to an outsider, the sometimes esoteric art of seismic analysis, and earthquake engineering for structural and piping design.
I am a radiation biologist by training, and my various past and current jobs have dealt with issues such as radiobiological research, administrative issues, waste disposal and emergency planning. But having had many more different issues put before me as Commissioner, I can truly say that engineering issues are one of the more complex issues to understand.
Now at this point, speech writers tell you it is customary to convey a humorous story or joke, and so my staff suggested that I tell a story that would be appropriate for this crowd - something that structural, seismic and piping engineers in the nuclear field would appreciate. They said I should include terms like - dynamic analyses, seismic experience data, soil-structure interaction, response spectra, time-history, and "hclpf" (high-confidence, low probability of failure) just to name a few - things that engineers love, but as a radiation biologist, I had no idea what these terms meant, nor how they could be used. How could I tell a convincing engineering joke? What was I going to do? Well, my better judgment told me, for this instance, I should not listen to my staff, so I nixed the whole idea - and it was a good thing too. It wouldn't have turned out well. My staff has a real "deconvoluted" sense of humor.
Nevertheless, since this is a luncheon speech, I will offer you some food for thought rather than to offer insights or solutions to a particular issue. I'm going to talk about the concept of risk and our perception of it.
As most of you are aware, the NRC is moving toward a risk-informed regulatory structure. By going fully to a risk-informed regulatory structure, we believe that this will result in more scrutable criteria and regulations - that there will be more stability in the regulatory process - that by performing PRAs and other risk assessment processes, and using those results, we will know when we have exceeded some acceptable risk threshold. I have to think about that, because everyone perceives risk differently. What is an acceptable risk to one person, may not be acceptable to another. And what one factors into determining risk, and the order of significance, may not be the same to everyone.
Let's examine what is risk. As a general population, how do we define it, how do we measure it, and how do we determine whether the level of risk is acceptable?
When one's chance of being injured in a car accident is so many orders of magnitude greater than one's chances of being injured by nuclear power, why does the public appear to fear nuclear power so much more? Is it because, in the case of driving, an individual feels they have control of the risk by deciding whether or not to drive, or how to drive and they choose to take that risk; whereas in the case of nuclear power, they don't have that same feeling of controlling the risk and they choose not to take it? Very possible. However, what risks does one really control when driving a car? There are many outside factors or events of which one isn't even aware. Is it because one does not believe the statistics or the numbers? Perhaps one does not care what the numbers are.
How does this perception of risk fit in the context of nuclear regulations?
- How do we use risk information in our decision making process?
- Does use of risk information require quantitative measures, or can we use judgment or qualitative measures.
Let's consider this: If I were to tell an ironworker that I would give him or her $1000.00 if he or she would walk a 24 inch wide steel beam 50 feet off the ground, (like a steel structure under construction), but that there was a 10-3 probability that he or she would fall and die, would the ironworker walk it? Maybe so. If I made that same offer to you, would you walk it? Maybe, but probably not. The point is that many factors went into the decision. What was risk significant, what was not, and what was acceptable risk, and what was not. What could have been some of those considerations the ironworker or you took under consideration to make the decision?
1. How badly do I need the money?
2. I've done this for 10 years.
3. Do I have a family?
4. Will I be wearing a safety belt?
5. How long is the beam?
6. How close am I to retirement? And for the risk analyst,
7. What is the probability of falling and not dying.
In the traditional sense, one is evaluating the risk of the action by considering the likelihood and consequences of failure, against the likelihood and consequences of success. Whichever argument wins will determine the decision.
DISCUSSION AND EXAMPLES
As the NRC embarks on becoming more risk-informed in how it regulates, some have said that the NRC is not moving rapidly enough in this area and others have said we are moving too fast. But there are risks to both the NRC and industry if we move too slowly or too quickly. If we move too quickly without thoroughly developing the framework, then unintended consequences could hamper further development or implementation of what I believe to be a good concept, and to which I am committed. On the other hand, if we move too slowly, we could unintentionally divert attention from areas of greater safety significance to those of lesser safety importance, and/or cause financial and human resources to be spent unnecessarily.
Let me discuss several topics that the staff currently has on its plate, how the concept and use of risk has affected development and implementation of each of these, and the problems that have been introduced by the concept of risk-informed. The topics I will discuss are: a) Potassium Iodide, b) License Renewal, c) Risk-Informed Part 50, and d) Assessment of Licensee Performance. The first subject I'd like to touch upon is the Commission's policy change regarding Potassium Iodide.
Policy Change Regarding Potassium Iodide
When a severe accident occurs, one of the fission products that is of concern is radioactive iodine. In humans, iodine is stored in and utilized by the thyroid gland, which functions in part, in the regulation of the body's metabolism. Uptake of excessive amounts of radioactive iodine by the thyroid can impair its function, increase the development of thyroid cancer, or even render the gland from functioning at all. Use of potassium iodide (KI) as a prophylaxis can saturate the thyroid with non-radioactive iodine to minimize the uptake of radioactive iodine.
Under the current policy, stockpiling or predistribution of KI for members of the general public is not specifically recommended to state and local emergency planners. However, the NRC has under consideration, a recommendation to suggest to state and local emergency planners that the stockpiling or predistribution of KI for the general public as a supplement to evacuation or sheltering should be considered. Guidance and information documents would be revised accordingly. Let me present some questions for your consideration regarding this issue:
Did the public's perception of risk drive the Commission's decision to change its policy? What were the risks of concern - (cancer and/or a belief that local government was not adequately protecting the public)? By stockpiling KI, are we signaling the possibility that an accident at a nuclear plant is more real than we have previously been willing to acknowledge, or are we being prudent by allowing the public greater opportunity to control their own risk to one of the effects of a radioactive release?
What is the likelihood of a severe accident? It's different for each plant. The calculated numbers are quite low. Do the numbers have any bearing on public response because the word "nuclear" creates an aura of complexity, mystique and destruction? They probably do not. How do these numbers compare to the likelihood of a household losing or misplacing its KI, and not having it when needed?
What is the risk of someone not taking it properly (i.e., dose) or at the proper time? Are there risks of liability resulting from such a situation?
Does the public fully understand the limited protection afforded by KI - and does risk from other fission products or plant hazards even matter to the public?
As my vote on this issue has indicated, I believe the nation should have a policy on KI distribution. Regional stockpiles of KI should be available and in sufficient dosage treatments to provide coverage for a large segment of the population if more than one event should occur. The NRC should also provide accurate and balanced guidance to offsite planners, in laymen's terms, to assist them in making appropriate decisions on KI policy at the state and local level.
The second issue is license renewal.
There are risks associated with license renewal. Due to the aging effects of concrete and steel, is there a greater risk of the containment failing during an accident before reaching it's design pressure, during the renewed license period, than during its original 40-year license term? What about the containment's structural response during a design basis earthquake? Do we know enough about these issues to determine whether or not a safety issue exists, to the point that adequate protection to the environment, and health and safety of the public can not be concluded? You are the experts who can help give us the information necessary for us to make appropriate regulatory decisions.
Are there financial risks for the licensee? What has staff done to alleviate some of those concerns? The Commission has recently revised the procedures regarding the adjudicatory process. Issues that can be raised are only those which are relevant to the renewed license period. The Commission also has publically stated that staff would endeavor to meet a 30-36 month review schedule.
NRC staff has also tried to resolve as many issues on a generic basis as possible.
The issues that are left are plant-specific issues and "age-related" degradation issues, some of which are within your (the audiences) field of expertise: fatigue of metal components, concrete degradation and seismic response of structures, and thermal aging of stainless steel components are just a few examples. In a related area, the French have recently confirmed a generic issue regarding the aging of seismic absorbers which dampen earthquake-induced vibration in relay cabinets and motor control centers. To what degree is this an issue for U.S. plants? How great are the uncertainties in our understanding of these processes and how are plant PRA results affected by these uncertainties? How much "margin" exists, based upon deterministic calculations? If there is an increase in risk in core damage frequency, however small, will it be acceptable to the public?
Two license renewal applications have been received and are currently under review by the NRC staff. Both reviews are going well and are on or ahead of schedule. No exceptional problems have been identified to date.
The Commission strongly supports the concept of license renewal. If a licensee makes the financial and business decision to pursue license renewal, the NRC has an obligation to respond to that decision in an efficient and effective manner while ensuring adequate protection of health and safety issues and protection of the environment.
Another item I'd like to mention is a risk-informed Part 50.
Risk-Informed Part 50
The Commission is in the process of transitioning 10 CFR Part 50 to become risk-informed and eventually performance-based. What does that mean? Does it mean the same to a regulator as it means to the public. Is there even agreement between colleagues?
Let's first consider criteria. Should regulatory criteria themselves be derived from risk insights? An example of this is the NRC regulation, 10 CFR 50.55a, which endorses the ASME Code. Are the specified criteria based on risk-insights? Should they be - and if so, what would they be? By and large, currently they are not. Yet, the Code is recognized internationally as a comprehensive and workable standard.
Second, let's look at the methods used to show compliance with those criteria. Should the means of meeting regulatory criteria allow the use of risk analyses? For example, in the case of a cracked steel component or a steam generator tube, what should be the role of probabilistic fracture mechanics to demonstrate compliance with Code requirements - what conditions are appropriate for its use? This is an issue that the NRC staff is currently addressing.
What is the role of deterministic criteria and methods in a risk-informed framework? Clearly, there is a place for them in any regulatory framework, whether it be a deterministic one or a risk-informed one.
These are the issues that are being wrestled with, not only in the nuclear reactor arena, but in the nuclear materials (NMSS) arena as well. NRC's policy with regard to a risk-informed regulatory framework has to be consistent agency-wide, and implementation of that policy is very challenging. To illustrate this point, the Commission has been wrestling with what is basically a definitions paper on risk-informed and performance-based concepts. We have been reviewing it since February, and only now are we "close" to coming to a resolution. Most of the comments we have received regarding the agency's PRA Implementation Plan have been from those dealing primarily with reactor issues - the ACRS (Advisory Committee on Reactor Safeguards), and NEI (Nuclear Energy Institute). We have received very little input from the State regulatory bodies, industrial users, or medical practitioners.
The last area I'd like to mention is NRC's processes for assessing licensee performance.
NRC's Assessment of Licensee Performance
NRC is revising its performance assessment processes to be more performance oriented, and utilizing performance indicators that, in some fashion, considers risk elements.
Although the staff will be sending a paper to the Commission in the January time frame that describes the new assessment process in detail, I would suspect that the new approach directs more focus on "risk-significant" events, evolutions, or problems.
I would envision that inspection efforts will be towards more "risk-significant" structures, systems, and components. The focus will still be on safety.
Will there be a threshold at which a structure, system or component will be deemed risk-significant? I doubt that will happen, nor do I think that will necessarily be desirable. As I alluded to at the outset with my example of the ironworker, what one perceives to be risk-significant may not be to another.
These are some of the areas the Commission and the industry are facing today. As you can tell there are a lot more questions than there are answers - and where there are answers, they are not easy to develop. It will take the efforts of the NRC, the industry and individuals such as yourselves, to resolve the issues involving these complex questions. But resolve them we must, to successfully accomplish our risk-informed goals. No matter the difficulty, I have confidence that we will achieve those goals.
I thank you for your attention. You've been a very kind audience.