APPENDIX F

FACTORS BEARING ON THE CAPITAL COST DIFFERENCES BETWEEN LIGHTWATER COOLED AND LIQUID METAL COOLED BREEDER REACTORS

Many of the factors which bear on the relative capital costs of the LMFBR and of current U.S. reactors stem from the fact that the LMFBR uses a liquid metal as a coolant while most current U.S. reactors use ordinary light water.

The use of a liquid metal as a breeder reactor coolant stems from the stringent demands which breeding puts on the neutron economy of the reactor. Upon fissioning a chain reacting nucleus releases on the average between two and three neutrons. One of these neutrons on the average must cause the fission of another nucleus so as to continue the chain reaction. In order for breeding to occur, that is in order to get a net increase in the amount of chain reacting material in the reactor, it is necessary for at least one of the remaining neutrons to convert a fertile nucleus into a new chain reacting nucleus. Thus, on the average, at least two of the neutrons released in a fission process have to be utilized profitably in a breeder reactor and, since less than three are released in the first place very little wastage of neutrons can be allowed.

One of the ways in which the neutron economy of a breeder reactor is optimized is by adjusting the speed of the neutrons causing the fissions so that the maximum number of neutrons are released per neutron captured in the chain reacting fuel. For a breeder based on the fission of plutonium, as is the LMFBR, this ratio is maximized when the neutrons lose as little energy as possible between their emission and absorption. The reactor must be designed, therefore, so that: (1) a neutron in the chain reaction should bounce off as few atoms as possible between the fission event which produces it and that which it causes in turn, and (2) the neutron loses as little energy as possible in each collision that it does undergo.

Both of these conditions are met using a liquid metal coolant: (1) such a coolant is much more effective than water in removing heat from the surfaces of the fuel-consequently the fuel rods can be packed closer together in the coolant and a neutron going from one fuel rod to another has to penetrate fewer coolant atoms. (2) The light neutron will lose much less energy in a collision with a heavy metal atom because that atom-unlike the light hydrogen atom in water-will hardly recoil at all. Molten sodium has been selected as the coolant for the LMFBR.

Sodium has both economic advantages and disadvantages relative to water as a coolant. Its advantages stem primarily from its high boiling point, 1,620°F. One consequence is that it is possible to operate the reactor at low pressure-unlike water-cooled reactors where the water is kept at very high pressures (up to 2,500 pounds per square inch) so that it may be superheated to temperatures where conversion of heat into electrical energy is relatively efficient. With a low-pressure

system the tanks and pipes which constitute the reactor plumbing can be designed with thinner walls and quality standards are less critical to public safety.

Another advantage of the high boiling point of sodium is that it becomes possible to operate the reactor at higher temperatures than water-cooled reactors which operate at between 500 and 600°F. LMFBR's would probably heat their sodium coolant to temperatures of the order of 1,000°F which would give them a thermal conversion efficiency of about 40 percent-considerably higher than the 33 percent being achieved by water-cooled reactors. The higher thermal efficiency of the LMFBR would result in cost savings because a smaller turbine, less cooling water, smaller cooling towers et cetera are required per unit power output as the thermal efficiency increases.

The economic disadvantages of sodium as a coolant stem primarily from it having the unfortunate property of burning vigorously-even explosively-if it comes into contact with air or water. As a result, elaborate arrangements are required when loading or unloading the fuel in the reactor or performing other required maintenance operations to insure that air does not obtain access to the sodium. Since the LMFBR is designed to have a steam driven turbine-generator system for converting the heat in the sodium to electrical energy, great care is also required to prevent leakage between the sodium and water sides of the steam generators. In the current designs the heat is transferred first from the radioactive sodium which cools the reactor to nonradioactive sodium and then to the water. In this way, if a sodium-water fire should occur, at least it won't involve the highly radioactive primary coolant.

There is some question as to whether the economic advantages would outweigh the disadvantages of sodium as a coolant over the long term. The balance will be determined in part by whether it is decided in the future that current designs are overly conservative in, for example, the degree of separation between the steam generator and the primary sodium. In the short term, however, it appears clear that the capital costs for the LMFBR would be higher than those for water-cooled

APPENDIX G

SOME MANAGEMENT PROBLEMS IN THE LMFBR DEVELOPMENT PROGRAM AS DOCUMENTED IN GAO REPORTS

Fast flux test facility.-The FFTF is a nuclear reactor being built at ERDA's Hanford Engineering Development Laboratory in Washington State. In many ways it is the precurser of the demonstration breeder reactor which ERDA plans to build on the Clinch River in Tennessee-whose design is in fact in good part based on that of the FFTF. The FFTF will be a reactor with about 40 percent of the thermal power of the Clinch River reactor designed to test the properties of breeder fuel and materials under LMFBR operating conditions.1 According to a 1975 GAO review of the FFTF program:

AEC's initial cost estimate ($87.5 million) at project authorization was based upon several contractor-prepared conceptual design cost studies. In December 1968, AEC approved a changed core concept . . . The initial estimate was dependent upon the use of several components already proven in a sodium reactor environment. Because off-the-shelf items were not available, however, AEC subsequently was required to establish or reestablish an industrial capacity for manufacture of components of (sic) high temperature sodium service and to develop new nuclear industry standards for these new higher temperatures.

Several major components and facilities included in the conceptual design studies were deferred or deleted from the project and numerous consolidations and simplifications were made . . .

In July 1970 AEC presented to the Joint Committee a start of construction capital cost estimate of $102.8 million

On January 29, 1973, AEC advised the Joint Committee it was increasing the construction cost estimate from $102.8 to $187.8 million

.

In a letter of April 4, 1973 to AEC's general manager, the Joint Committee's Executive Director stated that the total costs associated with construction of the FFTF appear significantly greater than those which were included in the budget data on the construction project. He was also of the opinion that the Commission had not fully and promptly advised the committee of the changing cost estimates, schedule delays and other factors.

AEC was then requested by the Joint Committee to provide a current estimate of all costs associated with the FFTF, including those in the operating budget, as well as any plant and equipment obligations

On May 17, 1973, for the first time AEC provided the Joint Committee with a cost estimate in one place for the entire FFTF program-$509 million . . .' On March 11, 1975, Thomas A. Nemzek, ERDA's Director, Division of Reactor Research and Development, told the Joint Committee:

A change is not being proposed in the official ERDA estimate of FFTF project cost-$530 million at this time. However the project is experiencing substantial inflationary growth ... On the basis of these current trends, the project is forecasting a project cost of $622 million...3

The GAO review commented on the lateness of previous cost estimate increases as follows:

1 Thomas A. Nemzek, Director, Division of Reactor Research and Development, ERDA, information supplementary to testimony before the Joint Committee on Atomic Energy, March 11. 1975. pp. 54, 61.

2 GAO Staff Study, "Fast Flux Facility Program," 1975, pp. 11-14.

Ref. 1, p. 57.

From June 1970 until January 1973, AEC's plant and capital equipment estimate held at $102.8 million. On January 29, 1973, at which time costs totaling about 83 percent of the $102.8 million estimate were incurred or committed, AEC told the Joint Committee that it was increasing the FFTF estimate to $187.9 million.

In November 1973, at the request of the AEC Chairman, AEC and FFTF contractor officials developed a revised plant and capital equipment cost estimate for the project which amounted to $420 million . . . As in the case of the previous increase, funds equivalent to a major portion of the existing estimate (76 percent) had been incurred or committed.*

The GAO review also noted that:

The FFTF has experienced a substantial schedule slippage. In March 1967, shortly before authorization of the FFTF projects, AEC informed the Joint Committee that FFTF construction was expected to start by June 1968, and that full power operation would begin early in 1974. Because of considerable delays in the conceptual and preliminary design effort, however, FFTF construction did not actually start until July 1970-a slippage of about 2 years, AEC headquarters officials informed us that achievement of the full power operation milestone is not now expected until May 1979.

At start of FFTF construction, only limited detailed design effort had been accomplished and, since that time, design and construction have been accomplished concurrently."

Despite the increase in estimated costs by a factor of 7 and slippage of the full power operation date by 5 years, the current design is less flexible than that originally conceived and the GAO has expressed concern that "these changes may limit the number and type of experiments that can be performed" at the FFTF.

Sodium pump test facility.-A precedent exists in another LMFBR development program project for GAO's concerns about the ability of the redesigned FFTF to accomplish its mission. According to another GAO report:

The construction of the sodium pump test facility was authorized in the fiscal year 1966 budget. The estimate presented to Congress for approval at that time was $6.8 million. In 1969, a review of the project by a private architect-engineering firm revealed that the project, with its then current scope, would cost $25.2 million.

To reduce estimated costs, the project scope was then revised to test sodium pumps having a capacity of about one-third the size of those initially anticipated to be tested. The reduced project scope resulted in a cost estimate of $12.5 million for the facility. This estimate was presented to and approved by the Congress as part of AEC's fiscal year 1972 budget request. In fiscal year 1974, this $12.5 million estimate was again revised up to $17.5 million. At that time, AEC stated that the reduced capability of the facility would not adversely affect the capability to test pumps up to the sizes needed for use in the foreseeable future of the LMFBR program.

ERDA is presently planning modifications to this facility so it can test CRBR (Clinch River breeder reactor)-size pumps, which are larger than the pumps for which the facility is presently designed. These modifications are presently estimated to cost $40 million, increasing the project's total cost to $57.5 million.'

To test full size plant components with sodium, ERDA has recently added to the LMFBR program a plant component test facility which is currently estimated to cost about $200 million and is planned for operation in the early 1980's.

Ref. 2, pp. 17, 18.

5 Ref. 2, p. 19.

e Ref. 2, p. 22.

GAO Report to the Congress, "The Liquid Metal Fast Breeder Program-Past, Present, and Future," pp. 25-26.

Clinch River breeder reactor.-The CRBR is supposed to be a demonstration commerical breeder reactor generating about one-quarter to one-third the power of the full sized LMFBR's the first of which ERDA expects to have operating in 1987.10

9

The CRBR is a joint government-industry effort. In August 1972: AEC estimated that $699 million would be required to design, construct, and operate the project, of which private project participants, primarily utilities were expected to provide from $274 to $294 million including $20 to $40 million from reactor manufacturers. AEC was authorized to contribute a total of about $422 million, $92 million of which was to be in direct financial assistance, $10 million in special nuclear materials, and $320 million in development work from AEC's ongoing LMFBR base program. Base program funds were limited to 50 percent of the then estimated capital cost of the plant. The direct assistance and base program funds were restricted as to what they could be used for. In general, they could not be used for end capital items for the plant.

ERDA's cost estimate for completing the CRBR project is now $1.736 billion— an increase of more than $1 billion. Because utility contributions were fixed, ERDA, by contract, accepted the open-end financial risks connected with the project and agreed to seek funds for any cost increase..."

The date for commercial operation of the CRBR has slipped by 3 years in 3 years to early 1983. According to a GAO report, additional delays may be expected:

Two important project milestones are (1) obtaining a limited work authorization by September 1, 1975, and (2) obtaining a construction permit by August 1, 1976.

Delays have already occurred in the licensing process. According to ERDA, neither the limited work authorization milestone nor the construction permit milestone will be met. A delay of 4 months could be expected in each category... The application for a limited work authorization was submitted to NRC (Nuclear Regulatory Commission) in October 1974. NRC, however, has not formally accepted the application for docketing because it feels additional information is necessary before a complete review of the application is possible..."

12

The GAO report goes on to describe other potential future causes of delay in the CRBR project: lack of timely and adequate funding, public hearings and outside legal interventions during the licensing process, delays in the delivery of long leadtime material and components, unavailability of craftsmen-particularly welders, and potential design changes-in particular those relating to a "core catcher" which is favored by the NRC but not by ERDA."

Due to the joint industry-ERDA funding of the CRBR project, a Project Management Corporation was established directed by a threeman steering committee representing ERDA, the Tennessee Valley Authority, and Commonwealth Edison. (The Tennessee Valley Authority is providing the site, will operate the plant, will purchase the power it produces, and will have the option to buy it after the project is over. Commonwealth Edison is providing engineering management and purchasing services for the project.) The GAO has described the organizational arrangement for the project as "complex and potentially cumbersome.” 14

Ref. 1. p. 61.

10 GAO Issue Paper to Congress: "The Liquid Metal Fast Breeder Reactor: Promises and Uncertainties" (1975), p. 101.

11 GAO, Report to the Joint Committee on Atomic Energy, "Comments on Energy Research and Development Administration's Proposed Arrangement for the Clinch River Breeder Reactor Demonstration Project" (1975).

12 GAO. Report to the Congress, "Cost and Schedule Estimates for the Nation's First Liquid Metal Fast Breeder Reactor Demonstration Power Plant" 1975, p. 27.

13 Ref. 12, pp. 27–33.