Cinquantamila.it La storia raccontata da Giorgio Dell'Arti

Robert J. Budnitz
Lawrence Berkeley National Laboratory
University of California, Berkeley, USA


[This paper summarizes a verbal presentation by the author in Erice on 21 August 2012. Like that presentation, this paper is not a comprehensive discussion of the topics covered. Rather, it is an overview summary which contains the author’s personal opinion on each of the topics covered, and is intended to introduce the subject to an audience of technically trained experts in other areas who may or may not have been introduced to these topics before.]

Abstract. There are 104 nuclear power units operating in the U.S. today on 67 sites. The fleet is generally operating well and is far safer than was the case 20–30 years ago. About 30 new nuclear units are in various stages of being proposed for construction in the U.S., although probably not all of them will be built. A number of issues confront nuclear power today in the U.S., and this paper provides the author’s personal opinion about each of them. Among the topics are natural gas prices; Yucca Mountain’s demise and the future of nuclear waste management; Fukushima lessons-learned; license extensions out to 60 years or beyond; safety and reliability enhancements; the advent of more risk-based utility decisions and NRC safety decisions based on PRA; the NRC’s design-certification process and process for new-plant combined operating licenses; and the possible advent of new SMR (small modular reactor) designs.


Figure 1 shows the locations of the 104 nuclear plants operating in the US today. All produce more than 500 MW (electric), and some are as large as about 1300 MW. The US has about one quarter of the nuclear power reactors operating in the world today. The youngest of these was designed in the 1970s, so some of the technology of the operating power reactors is very old. For example, the control rooms generally use analog electronics, although many of them have recently begun to switch over in a major way to digital technology.

Figure 2 shows the locations of the 30 or so proposed new U.S. nuclear plant units as of today, from the U.S. Nuclear Regulatory Commission (NRC) website. They are proposed for 19 sites, 5 of which are new sites. (The others are planned to be built adjacent to existing reactors on existing sites.) Many of these have signed contracts to begin design or to procure major equipment, and four of them have received NRC licenses and are in the midst of actual construction. Whether all of these will eventually be built is not known today. My own “guess” is that perhaps half of them will incur very long delays, or may even never be built, due to a variety of economic and other factors to be discussed below.

There are several major issues confronting nuclear power in the U.S. today and the bulk of this discussion will be to provide the author’s personal views on each of them.

The topics to be covered are as follows:

• Natural gas prices – will they hamper plans for building new nuclear plants?
• Current plants can be viewed as “cash cows”
• Yucca Mountain’s demise
• Fukushima lessons-learned (flooding protection, station blackout protection)
• License extensions out to 60 years and possibly beyond
• Capacity factors and outage management continue to improve
• Safety and reliability enhancements
• The NRC’s design certifications and new-plant combined operating licenses
• The advent of more risk-based utility decisions and NRC safety decisions based on PRA
• SMRs (small modular reactors?)

Natural gas prices – will they hamper plans for building new nuclear plants?
The short answer is “yes.” Recently, natural gas prices in the U.S. have plummeted, and predictions are that they will remain low for a long while, perhaps a decade or more. If sustained, this is very likely to make a huge difference in the current plans by many electric utilities to maintain their current positions vis-à-vis the new nuclear plants now on the drawing board (see Figure 2.) Nobody knows yet how much the impact will be, and to predict it would be folly now, but this is the single largest uncertainty factor today as to whether most or only a few of the proposed new nuclear plants will be built. My own feeling is that a significant fraction of the proposed new plants will end up incurring major delays.

Current plants can be viewed as “cash cows.”
This is certainly the case today. Most of the current plants are old enough that their capital cost has been paid off by now, so the only day-to-day cost “on the books” is their operating costs: labor, fuel, replacement parts, etc. These costs are quite low in aggregate, meaning that the cost to the utility and hence to the ratepayers of operating them is now very low and highly competitive with other options. This fact, which helps the local economy wherever the plants now run, is a major factor in the current U.S. economic picture vis-à-vis these operating plants. It means that even what might seem like sizeable investments to keep them running, such as multi-hundred-million-dollar steam generator replacements, are seen as economically viable. It also means that there is a huge financial incentive to keep these plants running for as long as they are technically able to do so safely.

Yucca Mountain’s demise
The proposed Yucca Mountain deep geological repository for disposing of the country’s nuclear waste is currently “dead in the water,” having been stopped by the current Federal administration. Efforts by some to revive it may or may not succeed, and this is a highly political matter beyond my ability to predict. In 2010, the President appointed a “Blue Ribbon Commission” to recommend what to do next, and that commission’s report was issued a half year ago. It contains a number of very useful policy recommendations, most of which are important to do whether or not the Yucca Mountain repository is revived. If it is, it will incur a delay of perhaps a decade compared to the “old” (Yucca Mountain) schedule; if it is not, finding another site and developing it, technically and politically, probably means a delay of 20 years or more compared to the “old” schedule, in my opinion. In the meantime, all of our reactor waste is still on the reactor sites where it always has been, although much of it is now in dry cask storage (or soon will be), which is both adequately safe and adequately secure in my opinion for a period of several decades.

Fukushima lessons-learned (flooding protection, station blackout protection)
After the Fukushima accident in Japan, both the industry (through its industry organizations -- NEI, INPO, and EPRI) and the NRC embarked on a series of lessons-learned studies. These have already resulted in some backfits and other changes, and more is to come. The major short-term insight on the U.S. side is the conclusion that the root cause of the accident at Fukushima was the result of a very poor siting decision, and we have tentatively concluded that none of our sites is vulnerable in the same way. Nevertheless, several other potential vulnerabilities have been identified, and each is being worked on by both the industry and the NRC. The NRC has ordered each plant to do a re-evaluation of their site’s vulnerability to flooding and to large earthquakes, and to study the capability of the design itself to withstand potentially larger floods and larger earthquakes than previously contemplated. All of the plants are also reviewing their ability to withstand an extended “station blackout” (loss of all electricity) and to identify ways to improve their capabilities. The stations are also reviewing their operating procedures and plans to cope with a major beyond-design-basis event, and their abilities to obtain offsite assistance in the event it is needed. A few studies and changes related to other issues are also underway now. In my opinion, there is excellent cooperation between the NRC and our industry on all of these issues, which is very nice to report.

All of the above is going on now. As mentioned, there are several longer-term initiatives too, whose study and implementation will take a few years. In the meantime, everyone (in the U.S. a around the world) is watching as the Japanese learn their own “lessons.” Which lessons will apply where, and how, is still for the future, and many of them are likely to be highly targeted to one or another specific reactor design feature or site.

License extensions out to 60 years and possibly beyond
Each U.S. reactor license was issued originally for 40 years. Based on extensive technical studies (measurements, analysis, inspections, etc.), 71 of our 104 operating reactor units have already been granted an extension to 60 years by the NRC, and essentially all of the others have applied to the NRC for these extensions, or soon will. The technical issue now confronting everyone is whether there is a valid technical basis for an even longer extension. The studies are just beginning. There is still time to do these studies, because the first of our operating units has recently passed its 40th-year mark, and all the others are “younger.” Of course, much of the rest of the world is working on this issue too and is watching the U.S.

Capacity factors and outage management continue to improve
The long-term trend of vastly improved capacity factors and improved management of refueling outages has made a remarkable difference to our U.S. plants. (See Figures 3 and 4.) Their economic viability is far better, and the enhanced reliability of operation has also been a major benefit in improving their safety performance a lot too. This is a source of pride in the industry, and in the NRC also, and it is justified. There is no one single factor (or even 2 or 3 factors) that accounts for this long-term improvement. It is instead the result of a myriad of small incremental improvements in every facet of operations. It is surely a sign of how much can be done by long-term concerted effort on everyone’s part.

Safety and reliability enhancements
There is no doubt, on a statistical average basis, that the U.S. nuclear plant fleet is much safer and much more reliable than it was two or three decades ago. (This does not mean that accidents can’t or won’t happen --- a single outlier or poor performer could have an accident any time, as we saw in Japan last year.) But the statistics and their interpretation speak for themselves.

The gains have been impressive, as shown by numerous measures of reactor performance over the last two decades. The number of so-called “significant events” per reactor year for US reactors since 1988 is shown in Figure 5. Figure 5 also shows the NRC’s definition of the term “significant event.” Back in 1979, at the time of the Three Mile Island accident, there were about 4 such events per nuclear unit per year. By 1988 this had been reduced to less than one (0.9). By 2006, that number had dropped to 0.01, that is, only one event that year throughout the US for our 104 operating units. For the 5-year period 2006-2010, the average was about 3 per year country-wide. These data show a safety improvement in this figure-of-merit by about two orders of magnitude.
Another indicator of safe operation is the number of scrams while the reactor is critical, shown in Figure 6: If any of a large number of abnormalities occurs, a signal is generated automatically to shut down the plant, or “scram” it, by inserting the control rods. The operators can also scram the reactor if a signal is not automatically generated. In 1985, there were about 4 scrams per nuclear unit per year. By 2006 that number had dropped by an order of magnitude to 0.32, or only about one scram per year for every three units. Note that a scram does not necessarily mean that the reactor is in trouble. It is often due to a minor issue that needs to be fixed.

The forced outage rate, or the percent of time that a plant was shut down as a result of repairs required in the wake of an event, is shown in Figure 7. The average outage rate for the U.S. fleet has dropped from about 10% in 1998 to less than 1% in 2010. The smaller outage rate is largely because the number of forced outages is fewer, and also because the repairs are less complicated, so the plant can come back online more quickly.

The overall safety improvements in the U.S. reactor fleet can be attributed to a number of factors. In my opinion, almost all reflect the development and importance of a culture of safety:

• Learning from experience: an industry-wide reporting system and no-fault reporting
• Analysis: major efforts to analyze each event for its causes and implications
• Maintenance: concentrating on the important things and designing for easier maintenance
• Operator errors: simulator training and better procedures
• Industry-wide peer-to-peer inspection visits, task forces
• Design changes: eliminating design flaws and working toward a “forgiving” design
• NRC: Risk-informed enforcement actions (less attention to minor events).

The NRC’s design certifications and new-plant combined operating licenses
The NRC staff is diligently working on the review of several new reactor designs, to provide a so-called “design certification.” Several designs have been through the NRC process, and the NRC is now reviewing several other design submittals, or on certifying modifications to designs that have already been certified. All of these are for what we call “large LWR” reactors, which represent evolutionary improvements over the existing fleet. Some of the new design features, however, are almost “revolutionary,” such as the features that provide much more passive means of reacting to accidental upsets and bringing the plant to a safe shutdown state without active systems or active operator intervention.

Essentially all of the new reactors being ordered or discussed in the U.S. (discussed above --- see Figure 2) plan to use these new “design certification” plant designs, and will go through a new NRC one-stop licensing process to obtain a so-called “combined operating license.” This new NRC process, which is still being tested for the first time now, is intended to provide the same full NRC safety review and the same full opportunity for public involvement, but in a fashion that expedites the procedures and decreases the risk of unnecessary delays. In my opinion, this change in NRC’s regulator procedures is one of the major new factors in making possible an expansion of nuclear power in the U.S. of other factors also fall into place.

The advent of more risk-based utility decisions and NRC safety decisions based on PRA
Over the last decade or so, both the NRC and the U.S. operating plants have increasingly used PRA (probabilistic risk assessment) methods to analyze plant performance and to improve both safety and reliability. This has been a major initiative, and some experts credit this as the single largest technical contributor to the increased plant reliability and safety that has occurred. (I am among those who hold to this view.)

The NRC itself has been using “risk-informed” and “performance-based” approaches to its regulations for a long while. Among the most successful areas where this has been used is the NRC’s reactor oversight process, used by the resident inspectors and the regulatory staff to track and understand how the plants are performing. It supports a so-called “graded approach” to NRC’s regulatory oversight and their enforcement actions. In this approach, those areas that contribute most to safety are given most weight in terms of both time spent and importance attached to them. Areas judged to represent lesser risk are correspondingly given somewhat less attention. PRA methods and results are used to support the technical judgments. This has been remarkably successful by all accounts. This same approach is now being used in many other areas of NRC regulation, and the benefits are apparent to all.

Recently, a two-year effort by a specially appointed NRC “task force” produced an important report, NUREG-2150, “A Proposed Risk Management Regulatory Framework.” This report is intended to provide a decade-long blueprint for how the NRC might move even more extensively into using a risk-type philosophy to modify how the agency regulates, not only in the reactor area but across its many other responsibilities too. How this will play out in the future is hard to predict in detail, but in my opinion the NRC is sure to be relying more and more on this philosophical approach throughout its work.

The industry on its side is also using the same ideas more and more. Each operating nuclear plant now has an up-to-date PRA which is used to help the plant decide how to carry out a whole host of different activities. On-line maintenance is deployed using PRA insights; outages are managed using PRA-based analysis; operator training and maintenance schedules are affected by risk-informed ideas; engineering analyses of the more traditional kind, such as thermal-hydraulic, metallurgical, fuel-performance, and reactivity analyses, are now routinely informed by PRA analyses.

The industry and the NRC have worked cooperatively to develop a methodology standard for performing PRA, published as an ANS-ASME standard. This standard is now used extensively by the industry and the NRC to help assure that PRA analyses relied on for decision-making have acceptable quality.

This is all to the good, and the trend will only continue.

SMRs (small modular reactors)
There is a lot of industrial activity today, both in the U.S. and around the world, to develop so-called SMR (small modular reactor) designs. There are many different design concepts, some of which are smaller versions of the latest large LWRs, some of which are LWRs but of very different designs, and some of which use different concepts such as gas-cooled designs or liquid-metal-cooled fast-spectrum designs. It is beyond my scope here to discuss these in detail. However, it is important to note that the NRC is actively engaged now in trying to adapt its technical regulatory positions, many of which were developed explicitly to regulate large LWRs, to enable licensing of these several new SMR design ideas.

Some of these designs are for individual nuclear units in the range of 200 to 300 MW-electric, and a few are larger, but some of them are far smaller, even as small as 20 to 30 MW-electric in size. Of course, the idea is typically to deploy several on a single site, which is why the term “modular” is used.

Why is there so much industrial interest in developing smaller reactors? There are several reasons. First, it’s much easier to keep a smaller core cool, largely thanks to its higher surface-to-volume ratio and certain passive safety features. Some small cores can almost cool themselves without any intervention. Because the cooling problem is less severe, the plant would not require as much auxiliary equipment, so the design could be simpler, there would be fewer things that could go wrong, and the plant would be easier to understand and operate. The general opinion industry-wide, which I share, is that many of these features will make these designs far safer, and also far easier to analyze to provide high assurance that they are safer.

There are other reasons to favor small plants. They would not require the utilities to raise as much initial capital for their construction. Additional reactors could be added more easily for increased capacity, and the waste inventory per plant would be reduced. Also, when one unit goes off-line for any reason, planned or unplanned, the impact on the electrical grid would be far less.

The main reason that smaller reactors may not be widely deployed soon would be the economics. The economies of scale that are gained in the capital cost with a bigger reactor are a major factor. LWR plants of half the size of the new ones we are starting to build are estimated to cost 75% as much to build if built the same way (with mostly field fabrication.) Although many manufacturers and constructors would prefer to design and build standardized and modular nuclear plants about one fifth to one tenth the size of the current generation, and many utilities would prefer to deploy them, the cost per megawatt of doing so using the traditional approach could be two or more times higher. This has led to some innovative design and manufacturing ideas that go well beyond the “traditional approach,” with a major emphasis on factory fabrication and transporting major modular components from factory to site, to bring the costs down. Anther major cost driver is staffing, and some of these innovative SMR designs claim to be able to run with much less in the way of operating crews and maintenance staffs, and even of security staffs. Whether these cost savings will be sufficient to allow them to penetrate into the marketplace is still not known.

I predict that this area will be one of great activity, both technical and regulatory, over the next few years. Where it will come out in the end is anybody’s guess.

ACKNOWLEDGMENT

The author is an employee at the University of California’s Lawrence Berkeley National Laboratory, but none of his funded projects has supported the development of this paper, which was written on his own time. However, broad general support from LBNL is gratefully acknowledged.

An earlier version of this paper by the same author, covering some of the same scope, was published recently by the American Institute of Physics as a chapter whose full citation is as follows:

“Topics in Nuclear Power,” by R. J. Budnitz, chapter in Physics of Sustainable Energy II: Using Energy Efficiently and Producing it Renewably, D. Hafemeister, D. Kammen, B. Goss Levi, and P. Schwartz (editors), published by the American Institute of Physics (conference proceedings, 5-6 March 2011, published 2012)

The sections of this paper on safety and reliability enhancements and on small modular reactors have borrowed significantly from the AIP paper. Acknowledgment is hereby made to the AIP for the use of this material.




Figure 1: Nuclear reactors operating in the U.S.




Figure 2: Proposed new U.S. nuclear reactors






Figure 3
US Nuclear Industry Capacity Factors 1971 - 2011






Figure 4
US Nuclear Industry Refueling Outage Days, 1990 – 2011




Figure 5: Significant Events at U.S. Nuclear Plants




Figure 6: Automatic Scrams While Critical, 1980 - 2006




Figure7: Forced Outage Rate, 1998 - 2007

ROBERT BULDNITZ