|
CBO Statement of Costs of Reprocessing before the November 14, 2007
Mr. Chairman, Senator Domenici, and Members of the Committee, thank you for the invitation to discuss the Congressional Budget Office's (CBO's) analysis of the costs of two alternatives for the use and disposal of nuclear fuel. For the past 50 years, the nuclear waste produced at reactors across the United States has largely been stored at the reactor sites. That practice, however, has been deemed untenable for the long run. CBO's analysis compares the cost of two fuel-cycle alternatives for the current generation of thermal reactors. One alternative is direct disposal (as stipulated by current law), which involves using nuclear fuel once, cooling it at an interim storage site, and then disposing of it in a long-term repository. The second alternative is reprocessing, in which spent nuclear fuel is cooled and then reprocessed for one additional use in a reactor, and the wastes from reprocessing are stored in a long-term repository. My testimony makes the following key points:
Background on Nuclear Fuel-Cycle Alternatives As of 2006, 104 nuclear reactors were operating in the United States, with a collective generating capacity of about 100 gigawatts of electricity. Those reactors account for nearly 20 percent of the electricity produced in this country.1 All of the commercial nuclear power plants in the United States generate electricity by relying on the uranium-235 isotope to sustain a nuclear reaction. Uranium-235 is relatively scarce and typically makes up less than 1 percent of mined uranium ore. The bulk of that ore consists of uranium-238, which cannot be used directly to sustain a nuclear fission chain reaction. For a sustained reaction to occur, the uranium must be enriched—that is, the proportion of uranium-235 must be increased, generally to between 3 percent and 5 percent in the case of fuel for civilian reactors. After approximately four years in a reactor, too little uranium-235 remains in the fuel to generate electricity. The spent fuel can be handled in one of two ways: Under direct disposal, it is placed in interim storage for cooling, with the goal of eventually storing it in a stable geologic formation over the long term. Under reprocessing—which is done in a few countries but not the United States—a reprocessing facility recovers the useful components of the spent fuel (uranium and certain forms of plutonium) and returns them to the fuel cycle, where they are combined with newly mined uranium to produce more reactor fuel (see Figure 1). Any waste remaining from the spent nuclear fuel after the uranium and plutonium are removed is intended to be stored in a long-term repository. Thus, under either option, some form of long-term storage facility is necessary. Nuclear Fuel Cycles  Source: Congressional Budget Office based on Ian Hore-Lacy, Nuclear Energy in the 21st Century (London: World Nuclear University Press, 2006). Note: U3O8 = uranium oxide concentrate; UF6 = uranium hexafloride; U-235 = uranium-235. No long-term repository for storing commercial nuclear waste is currently operating anywhere in the world. The Department of Energy (DOE) is planning to build and operate such a repository at Yucca Mountain in Nevada. That facility, originally scheduled to open in 1998, is now intended to start operating in 2017, although a later opening date—2020 or 2021—is more likely.2 That date would be nearly 40 years after lawmakers directed DOE to begin studying potential sites for a deep underground repository for spent nuclear fuel.3 With such delays, the accumulated stock of nuclear waste is expected to exceed Yucca Mountain's mandated capacity before the facility begins accepting waste for storage. One approach to that problem is to expand the repository's capacity, either physically or by lifting the mandated limit on how much waste Yucca Mountain can accept (an option that many observers believe could be undertaken without compromising safety). Another approach is to reprocess spent nuclear fuel for reuse in reactors. That option could increase the effective capacity of the repository by allowing more nuclear waste to be stored in a given amount of space. The main factor that determines the overall storage capacity of a long-term repository is the heat content of nuclear waste, not its volume. The waste that results from reprocessing spent fuel from thermal reactors has a lower heat content (after a period of cooling) than the spent fuel itself does. Thus, it can be stored more densely.4 The extent of that densification directly affects the relative costs of direct disposal and reprocessing. However, unlike waste from the reprocessing process, spent fuel that has been reprocessed and used again has a higher heat content than spent fuel that has been used only once. Storing that previously recycled spent fuel in the long-term repository immediately would eliminate all of the densification benefits of reprocessing. Consequently, for reprocessing to reduce the need for—and cost requirements of—long-term storage, previously recycled spent fuel would have to be allowed to accumulate at some location outside the repository. Besides potentially lowering long-term disposal costs, reprocessing spent nuclear fuel has the advantage of reducing expenditures for freshly mined and enriched uranium. In effect, recovering unused uranium from spent fuel extends the life of unmined uranium resources. Furthermore, recovered plutonium is not subject to many of the fuel-preparation costs (such as for conversion and enrichment) that are necessary with uranium (see Figure 1). That potential for front-end savings was especially appealing when the U.S. commercial nuclear program began in the 1950s. It became less pronounced by the 1960s, as uranium prices declined and uranium preparation techniques matured. Spot prices for uranium have recently reached historical highs (adjusted for inflation), but high prices would have to persist for decades to increase the economic viability of reprocessing. Reprocessing and direct disposal differ not only in potential costs but also in possible risks for the proliferation of nuclear weapons. Spent nuclear fuel itself poses little risk of proliferation because the plutonium it contains is mixed with highly radioactive elements and can be recovered only in dedicated reprocessing facilities.5 But the most widely used method of reprocessing—called plutonium and uranium recovery by extraction, or PUREX—yields pure plutonium, which has relatively low radioactivity and can be handled directly. Thus, the PUREX method recovers plutonium in a form that poses risks for theft and proliferation. Other reprocessing methods being considered by policymakers try to reduce those risks by not separating pure plutonium from spent fuel. Reprocessing Facilities The United States has limited experience with commercial reprocessing. Three reprocessing plants were built for commercial use, but only one—a plant in West Valley, New York, that opened in 1966—achieved any level of operation. The need for costly upgrades caused it to close in 1976, having handled only spent fuel from national defense operations.6 Today, five nations—France, the United Kingdom, Japan, Russia, and India—have or are developing reprocessing facilities. The world's largest reprocessing plant is located in La Hague, France, and has a gross capacity of 1,700 metric tons per year. The United Kingdom has two reprocessing centers at its Sellafield Nuclear Site: a 900-metric-ton thermal oxide reprocessing plant (THORP) and a facility that specializes in reprocessing waste for two specific British nuclear facilities (Oldbury and Wylfa, both of which are expected to cease operations by 2010). Another reprocessing facility has been under construction in Rokkasho, Japan, since the late 1980s. The start of commercial operations there has been delayed several times but is now expected to occur later this year. Thermal Reactors Versus Fast-Neutron Reactors CBO's analysis compares the cost of reprocessing nuclear fuel from thermal reactors—the type of commercial reactor used now in the United States—with the cost of using uranium fuel a single time and then putting all of it in a geologic repository. However, some policy initiatives, such as the Global Nuclear Energy Partnership, have focused on another type of reactor: an advanced burner reactor, which is a type of fast-neutron reactor. Whereas thermal reactors rely on less energetic, or modulated, neutrons to sustain a nuclear chain reaction, fast-neutron reactors rely on unmodulated (and hence more energetic) neutrons for a reaction. Fast-neutron reactors use plutonium as a fuel source rather than uranium because plutonium maintains a reaction with unmodulated neutrons more readily than commercial-grade enriched uranium does. Fast-neutron reactors offer several advantages. They can convert plentiful uranium-238 (which is not usable for nuclear chain reactions) into plutonium in such a way as to produce (or breed) more plutonium than the reactor itself uses. In that way, a fast-neutron breeder reactor can extend uranium resources by accessing 60 to 100 times more of the energy content of uranium than thermal reactors can.7 Fast-neutron reactors also generate less spent fuel than thermal reactors do. Besides uranium and plutonium, spent nuclear fuel includes two other types of waste: fission products and minor actinides. Minor actinides decay less rapidly than fission products do. Because the capacity of a geologic repository depends to a significant degree on the long-term radioactivity of waste, it is greatly influenced by the amount of minor actinides present in spent fuel. Advanced burner reactors can potentially burn all of the actinides in nuclear fuel, so waste from those reactors requires less geologic storage space than does either spent nuclear fuel from thermal reactors or the waste from reprocessing thermal reactors' spent fuel. Whereas reprocessing spent fuel is merely an option with thermal reactors (to extend uranium resources or to potentially expand long-term storage capacity), it is an integral part of the fuel cycle for advanced burner reactors. The fuel needed to power advanced burner reactors can be collected by reprocessing spent fuel from thermal reactors or from burner reactors. Furthermore, if burner reactors are used to reduce thermal-reactor waste, spent nuclear fuel must be reprocessed. This testimony does not consider reprocessing in the context of fast-neutron reactors, for three reasons. First, no commercial fast-neutron reactors exist in the United States and none are planned. Second, the 60-year-old PUREX process is essentially the only reprocessing method now used for thermal reactors, and given its long history, the cost of PUREX is better known than the costs of more-recent reprocessing technologies that are being considered for fast-neutron reactors. Third, reprocessing fuel for advanced burner reactors would probably require reprocessing nuclear waste from thermal reactors as a first step to create the fuel for the burner reactors and to manage any existing thermal-reactor waste. Thus, reprocessing thermal-reactor waste can be thought of as a transitional element to a burner-reactor program. Cost Comparisons for Direct Disposal and Reprocessing Reprocessing nuclear fuel could have several economic advantages over direct disposal. It could reduce spending on newly produced uranium fuel and extend the useful life of uranium resources. In addition, it could save money on long-term storage by reducing the size of the repository necessary to handle spent nuclear fuel or by delaying the need to expand such a facility in the future. With current reactor technology, reprocessing would also have economic disadvantages. First, it would require building dedicated facilities to recover the useful components of spent nuclear fuel and then to combine them into a form usable in a nuclear reactor. Second, previously recycled spent fuel would also need some form of long-term storage. To quantify the relative costs of reprocessing and direct disposal, CBO's analysis focuses on the costs of handling nuclear fuel after it is discharged from a reactor. In the case of reprocessing, those costs include the costs of reprocessing services (both recovering uranium and plutonium and fabricating them into usable nuclear fuel), transportation, and long-term disposal of wastes, partially offset by "fuel credits," which various models use to reflect the value of the reprocessed fuel (in the form of savings on the costs of newly purchased fuel). In the case of direct disposal, the costs in this analysis include costs for interim storage to cool the spent fuel, transportation, and long-term disposal. CBO reviewed a number of studies that shed light on the costs of nuclear fuel-cycle alternatives, including reports by the National Research Council, the Massachusetts Institute of Technology, the Idaho National Laboratory, and the Nuclear Energy Agency of the Organisation for Economic Co-operation and Development (OECD).8 However, CBO's analysis focused on two studies in particular: a 2006 report by the Boston Consulting Group (BCG) and a 2003 report by researchers at Harvard University's Kennedy School of Government.9 Those two studies are the only recent analyses available that investigate the costs of all facets of both reprocessing and direct disposal (including transportation, interim storage, and credits for recycled fuel). Other studies consider only the costs of reprocessing or do not examine the various components of total costs. In addition, the two studies' estimates of the cost of reprocessing services—one of the largest cost elements—bound the range of estimates provided in, or implied by, the other studies. The Kennedy study's estimate of the cost of reprocessing services is about twice the size of the BCG study's estimate. Other studies that CBO examined had cost estimates for reprocessing services that fell within the range defined by those two reports. Evaluating the Boston Consulting Group and Kennedy Studies on a Common Ground The BCG study concludes that reprocessing spent fuel costs about $30 more per kilogram than direct disposal (which the study estimates at $555 per kilogram). To directly compare that estimate with the results of the Kennedy study, CBO modified the Kennedy study to reflect a similar initial framework as in the BCG study. In that framework, the Kennedy study implies that reprocessing costs about $700 more per kilogram than direct disposal. Given the volume of waste expected to be generated over the lives of the plants evaluated, those estimates suggest that the present-value cost of reprocessing exceeds that of direct disposal by about $2 billion for the BCG study and by about $26 billion for the Kennedy study, as modified by CBO. (Present-value calculations use a discounted cost framework that describes the amount of funds that would be necessary in 2007 to pay all of the costs of a waste-management option over the assumed lifetime of a reprocessing plant.) Several differing assumptions account for much of the gap between those two present-value estimates. Such assumptions include the interest rate used to estimate the present value of future costs (the discount rate), the relationship between a reprocessing plant's yearly operating costs and total capacity costs, the time horizon over which the plant operates, the cost of a long-term repository, and the degree to which waste from reprocessing can be stored more densely than spent nuclear fuel in the repository. Changes to any of those assumptions will affect the relative costs of the two waste-handling alternatives. To control for those differences, CBO's analysis imposed a common set of cost assumptions on the estimates from the BCG study and from the modified Kennedy study. In particular, CBO assumed the following:
As those common assumptions are applied successively, the two present-value estimates of the difference between reprocessing and direct-disposal costs narrow from a range of $2 billion to $26 billion to a range of $5 billion to $11 billion (see Figure 2). Discounted Cost Differences Between Reprocessing and Direct Disposal (Present-value difference, in billions of 2007 dollars)  Source: Congressional Budget Office based on Boston Consulting Group (BCG), Economic Assessment of Used Nuclear Fuel Management in the United States (study prepared for AREVA Inc., July 2006); and Matthew Bunn and others, The Economics of Reprocessing vs. Direct Disposal of Spent Nuclear Fuel (Cambridge, Mass.: Belfer Center for Science and International Affairs, John F. Kennedy School of Government, Harvard University, December 2003). Note: The numbers shown here represent the extent to which the future costs of reprocessing spent nuclear fuel exceed the future costs of direct disposal on a present-value basis. (Present-value figures convert a stream of future costs into an equivalent lump sum today.) The first set of bars shows the values implied by the BCG and Kennedy School studies under their different assumptions. The successive sets of bars show the cumulative effects of applying common assumptions to the two baseline estimates. The common assumptions that CBO used for this analysis are described in detail in the text. a. Densification is the extent to which waste from reprocessing can be stored more densely than unreprocessed spent fuel in a long-term repository. Most of that remaining gap is attributable to the two studies' different assumptions about the costs of building and operating a reprocessing plant. The BCG study estimates construction costs at about $17 billion for a plant with a capacity of 2,500 metric tons per year. A meaningful comparable estimate cannot be derived from the Kennedy study because that analysis does not explicitly differentiate between capital and operating costs. The likelihood that a newly built U.S. plant would match either study's cost assumptions is difficult to judge; the historical record provides scant evidence about the overall cost of a reprocessing facility and its component parts. Not only are there few large-scale commercial reprocessing plants, but only limited information is available about their construction and operating costs. Neither the 900-metric-ton THORP facility in the United Kingdom nor the 1,700-metric-ton La Hague facility in France has enough capacity to handle the 2,200 metric tons of nuclear waste generated in the United States each year—the amount considered in this analysis. Thus, a facility larger than any past or current example would be necessary if a single reprocessing plant was to handle the United States' entire annual output of spent nuclear fuel. A larger facility would be more costly than existing plants, although to what degree is unknown. The limited information available suggests that the THORP plant cost around $6.3 billion to build (in 2007 dollars). The BCG study indicates that the construction cost of the La Hague facility was around $18 billion (unlike the THORP estimate, however, that total includes a fabrication facility for recycled fuel, which increases the overall cost). The nearly complete 800-metric-ton Rokkasho facility will reportedly cost about $21 billion, but part of that cost is attributable to specifics of the plant's location that would not necessarily apply to a U.S. facility. Given the lack of numerous commercial reprocessing facilities to use as examples, it is difficult to know how much geographic location, economies of scale, and regulatory environment affect the cost of a reprocessing plant. All of the costs for reprocessing services included in this analysis assume that the plant would operate near capacity for its entire life. History, however, suggests that such an assumption might be optimistic and therefore that the unit cost of reprocessing could be higher than described here. Neither THORP nor La Hague has operated close to full capacity for a substantial period. THORP has been closed for more than two years after experiencing a radioactive leak. Before that, the plant operated at about 60 percent of capacity over its first 11 years. Although La Hague has not had the technical problems of the THORP facility, it too is operating well below full capacity: at approximately 65 percent, according to recent estimates. Operating at less than full capacity limits the amount of spent fuel that can be handled for a given cost. Another factor that could increase the estimated cost of reprocessing relative to direct disposal is the discount rate used in present-value calculations. The rate assumed in this analysis is similar to those used in the BCG and Kennedy studies and slightly above a risk-free government rate—but it is well below the rate that might be applied for this type of project. A higher discount rate would result in a larger cost difference between reprocessing and direct disposal. The relative cost of reprocessing is also affected by the market value of recycled fuel. As noted above, the fuel credits used in this analysis reflect front-end savings from using recycled fuel rather than newly mined uranium. If the costs of uranium mining and fuel preparation increased, and if recycled fuel proved to be a good substitute for newly mined uranium in nuclear reactors, higher fuel credits could offset the cost of reprocessing to a greater extent. Although uranium prices are currently high by historical standards, it is not certain whether high prices will continue in the future or whether current prices will encourage additional uranium development that could lower prices. Furthermore, modifying a nuclear reactor to use recycled fuel entails some costs, which would offset a portion of the potential fuel credits from reprocessing. Sensitivity to Varying Assumptions Although the size of the cost difference between reprocessing and direct disposal depends on inputs to specific models, the conclusion that reprocessing is more expensive than direct disposal generally applies under various assumptions. CBO tested the sensitivity of the results to changes in some of the key parameters of this analysis.
In conclusion, the cost of reprocessing may be comparable to that of direct disposal under limited circumstances, but under a wide variety of assumptions, reprocessing is more expensive (given current reactor technology). Policymakers weighing the merits of reprocessing and direct disposal may have other concerns besides cost—such as extending U.S. uranium resources, reducing the threat of nuclear proliferation by adopting advanced burner technologies, or lessening the demand for long-term storage space. Judging whether those goals justify the added costs of reprocessing is ultimately a decision for policymakers. Department of Energy, Energy Information Administration, Annual Energy Review 2006 (June 2007), Table 8.2a, available at www.eia.doe.gov/emeu/aer/pdf/pages/sec8_8.pdf. Statement of Edward F. Sproat, Director, Office of Civilian Radioactive Waste Management, at the 178th meeting of the Nuclear Regulatory Commission's Advisory Committee on Nuclear Waste, April 10, 2007, available at www.nrc.gov/reading-rm/doc-collections/acnw/tr/2007/nw041007.pdf As originally enacted, the Nuclear Waste Policy Act of 1982 called for studies of three potential sites for long-term geologic repositories. Sections 5011 and 5012 of the Nuclear Waste Policy Amendments Act of 1987 effectively cancelled any investigation into sites other than the one at Yucca Mountain. The "densification factor" describes that relationship; for example, a densification factor of 2 indicates that twice as much waste from reprocessing can be stored at the same total cost (in other words, that the unit cost of storage is half as much). Steve Fetter and Frank N. von Hippel, "Is U.S. Reprocessing Worth the Risk?" Arms Control Today (September 2005), available at www.armscontrol.org/act/2005_09/Fetter-VonHippel.asp. Anthony Andrews, Nuclear Fuel Reprocessing: U.S. Policy Development, Report for Congress RS22542 (Congressional Research Service, November 29, 2006). Uranium Information Centre, Fast Neutron Reactors, Briefing Paper No. 98 (Melbourne, Australia: Australia Uranium Association, June 2006), available at www.uic.com.au/nip98.htm. National Research Council, Nuclear Wastes: Technologies for Separations and Transmutations (Washington, D.C: National Academy Press, 1996); Massachusetts Institute of Technology, The Future of Nuclear Power (Cambridge, Mass.: MIT, 2003), available at http://web.mit.edu/nuclearpower; D.E. Shropshire, Advanced Fuel Cycle Cost Basis (Idaho Falls: Idaho National Laboratory, April 2007), available at www.inl.gov/technicalpublications/Documents/3667084.pdf; and Nuclear Energy Agency, The Economics of the Nuclear Fuel Cycle (Paris: Organization for Economic Co-operation and Development, 1994), available at www.nea.fr/html/ndd/reports/efc. Boston Consulting Group, Economic Assessment of Used Nuclear Fuel Management in the United States (study prepared for AREVA Inc., July 2006); and Matthew Bunn and others, The Economics of Reprocessing vs. Direct Disposal of Spent Nuclear Fuel (Cambridge, Mass.: Belfer Center for Science and International Affairs, John F. Kennedy School of Government, Harvard University, December 2003). Nuclear Energy Agency, Accelerator-Driven Systems and Fast Reactors in Advanced Nuclear Fuel Cycles: A Comparative Study (Paris: Organization for Economic Co-operation and Development, 2002), Chapter 6, p. 211, available at www.nea.fr/html/ndd/reports/2002/nea3109.html. Shropshire, Advanced Fuel Cycle Cost Basis, p. L-12. |