larger nuclear units were cheaper to build per kilowatt of capacity. Moreover, utilities were growing rapidly and needed large increments of new power. The situation is very different now. Many seem to feel that a carefully designed small reactor might be easier to understand, more manageable to construct, and safer to operate. Although many of these claims seem intuitively convincing, they are difficult, if not impossible, to substantiate. OTA sponsored a search for evidence that small plants have any advantages over large plants in terms of safety, cost, or operability (20). This search revealed no firm statistical data in support of the small reactor, although it summarized some of the arguments that make it an attractive concept (see vol. II).

Utilities may find small plants especially appealing today because they allow more flexibility in planning for the total load of the utility. In addition, the consequences of an outage would have a smaller impact on the overall grid. Furthermore, reducing the size of plants would limit the financial exposure of the utility to loss and increase overall system reliability. Initially, small plants appear to suffer a disadvantage in unit construction costs since they cannot realize the full benefits of economies-of-scale. However, more of the plant could be fabricated in the factory rather than constructed in the field, and this could result in large cost savings if the market is large enough to justify investment in new production facilities. Moreover, the construction times for small plants would probably be much less than for their larger counterparts. Overall, it is not clear that small plants would necessarily be more expensive than today's large ones.

The operability of different sized plants may be compared on the basis of availability. As shown in figure 27 A & B, the availability of small plants generally exceeds that of currently operating larger plants, although only by about 5 percent. This trend surfaces both when availabilities are plotted as a function of the number of years after start of operation (which compares plants of the same age) and as a function of calendar year (which compares plants operating in the same environment). These differences could be due to either the number or duration of outages at smaller plants, which indicates that small plants

events is independent of reactor size. However, the initiating events that do occur at small plants may be easier to manage than similar events at large plants. This result is based on too small a sample to be conclusive, but it may warrant further study. Another safety comparison can be made on the basis of the consequences of an accident. The worst-case accident in a small plant would be less damaging because the fission product inventory is much less in a smaller plant. This effect, however, might be offset by the larger

number of small units needed to comprise the same generating capacity.

Small reactors are unlikely to be able to compete commercially with their larger counterparts unless R&D that specifically exploits the potential for modular, shop fabrication of components is sponsored. This would allow small plants to take full advantage of the increased productivity and quality of work in a factory setting.

The concept of a standardized reactor has been widely discussed for years in the industry but has yet to become a reality (18). The advantages are many: more mutual learning from experience among reactor operators, greater opportunity for indepth understanding of one reactor type, and sharing of resources for training operators or developing procedures. Since much of the concern over current reactors centers on their management rather than on their design, the opportunity to concentrate on learning the correct application of one well-understood design is appealing.

Utilities and vendors would be especially enthusiastic about standardized designs if that concept were coupled with one-step or streamlined licensing. The simplification of the licensing process might bring concomitant benefits in reduced capital costs. If the plants were smaller than those of the current generation, larger numbers of small plants would be built to meet a given demand, and this would facilitate standardization.

A major barrier to designing a standard plant is the difficulty in marketing identical reactors, given the current industry structure and regulatory climate in the United States. There are many opportunities for changes in today's plants, such as to match a particular site, to meet the needs of a specific utility, and to accommodate NRC regulations. In addition, the existing institutional structure does not lend itself easily to industrywide standardization. There are currently five reactor suppliers and more than a dozen AE firms. While each reactor vendor is moving toward a single standardized design, balance of plant designs by the AEs continue to vary. It is unlikely that a single dominant plant design will arise out of all combinations of vendors and AEs, which implies that there may not be industrywide standardization. However, it is possible that a few prominent combinations of the more successful reactor suppliers and AEs will join forces to produce a more manageable number of standardized designs.

No single reactor concept emerges as clearly superior to the others since the preferred design varies with the selection criterion. If safety is of paramount concern, the reactors that incorporate many inherent safety features, such as PIUS or the modular HTGR, are very attractive. In such reactors, the critical safety functions of reactor

shutdown, decay heat removal, and fission product containment are provided by simple, passive systems which do not depend on operator action or control by mechanical or electrical means. The full-scale HTGR is also attractive in terms of safety since it provides more time than any of the water-cooled concepts for the operator to res

pond before the core overheats. The remaining reactors appear to be roughly comparable regarding safety features. The HWR has the lowest inventory of radioactive materials, and the independent moderator loop serves as a passive, alternative decay heat removal system. In addition, the HWR has compiled a superb record in Canada. Advanced LWRs incorporate the benefits accrued from many years of extensive operational experience. Finally, small and/or standardized reactors may have operational advantages resulting from a better understanding of and control over their designs.

If the reactors were to be ranked on the basis of reliable operation and easy maintenance, a different order results. The advanced LWR is very attractive because these criteria have heavily influenced its design. Small reactors also appear high on the list because their size and shopfabricated components may facilitate operation, maintenance, and replacements. HWRs rate high because they have performed well to date, and they do not require an annual refueling shutdown. The few HTGRS that are in operation have had mixed performance records, but the newest design addresses some of the problems that contributed to poor reliability. One factor enhancing overall performance is the ease of maintenance in an HTGR resulting from inherently low radiation levels. There are many uncertainties associated with the PIUS concept. It is likely to pose maintenance problems. It is also possible that the behavior of the PIUS will be erratic in normal transients, thus increasing the difficulty of operation. In other ways, however, the PIUS could be simpler to operate.

Any attempt to rate these reactor concepts on the basis of economics is very difficult. Experience with LWRs indicates that the price of facilities of the same design can vary by more than a factor of 2, so estimates of costs of less developed reactors are highly suspect. Only a few speculative comments can be made. Small reactors suffer a capital-cost penalty due to lost economies-ofscale, but it is possible that this could be reduced by fabricating more components in factories and keeping construction times short. HWRs are expected to have comparable capital costs, but their lifetime costs may be lower than those of LWRS

since the HWRs have lower fuel costs. Standardization of any of the reactors discussed would reduce costs, if the reactors could be licensed and constructed more quickly. The HTGR appears to be comparable in cost to LWRs, but there are greater and different uncertainties associated with it. It is premature to estimate the cost of a PIUStype reactor for several reasons. First, it is still in the conceptual design phase, so types and amounts of materials cannot be determined precisely nor can construction practices and schedules be accurately anticipated. In addition, the PIUS designers are relying on low costs in the balance of plant to compensate for the higher costs of the nuclear island. It is not clear whether the balance of plant systems can be decoupled from their safety functions; the regulatory agencies obviously will have a major impact on this decision, and hence the cost of a PIUS-type plant.

A final criterion applied to these reactors might be the certainty of our knowledge of them. How predictable will their performance be? The ranking here is almost the reverse of that for safety. Advanced LWRs are clearly superior in terms of familiarity because they have evolved from plants that have operated in the U.S. for more than 20 years. HWRs have also compiled a long record, but design modifications might have to be made before the reactor could be licensed in the United States. There is much less experience with HTGRs in the United States, with only a single facility in operation. The PIUS concept lags far behind the other reactors in terms of certainty since it has never been tested on a large scale.

This survey has examined many reactor concepts and found that none were unambiguously superior in terms of greater safety, increased reliability, and acceptable cost. Most represent a compromise among these factors. A few could not be adequately compared because so many uncertainties surround the design at this stage. The present lull in nuclear orders provides an opportune time to reduce the uncertainties and expand our knowledge of the less well-tested concepts. A demonstration of advanced LWRs may soon occur in Japan, and the results should be valuable input to future decisions on the LWR concept. If continued, the Department of Energy's development program on HTGRs will con

tinue to provide information and experience that could make the HTGR a viable alternative to the LWR. It may also be valuable to examine the operation of Canada's HWRs to determine if any of their experience can be applied to U.S. reactors. If considerable sentiment continues to be expressed in favor of small reactors, some initial design work may be appropriate. Finally, a preliminary investigation of the PIUS reactor would teach us still more about a concept that is very promising. Work on this or another "fresh look" design would require government support since the existing reactor designers do not see

a big enough market to support new research programs.

Until the results of future investigations are in, nothing on the horizon appears dramatically better than the evolutionary designs of the LWR. There is a large inertia that resists any move away from the current reactor types, in which so much time has been spent and from which so much experience has been accrued. However, if today's light water reactors continue to be plagued by operational difficulties or incidents that raise safety concerns, more interest can be expected in alternative reactors.

1. Abel, P., Workshop Discussion on Technological Changes, Dec. 8, 1982, U.S. Congress, Office of Technology Assessment, Jan. 3, 1983.

2. Agnew, H. M., "Gas-Cooled Nuclear Power Reactors," Scientific American, vol. 244, No. 6, June 1981.

3. Atomic Energy of Canada Limited, "CANDU-The Facts," November 1982.

4. Burnham, D., "Safety Goals Set by Nuclear Panel," New York Times, Jan. 11, 1983.

5. Civiak, R. L., "Potential for Reduction in the Predicted Release of Radioactive Material Following a Severe Nuclear Accident," Congressional Research Service, Mar. 21, 1983.

6. Fraas, A. P., "Survey and Assessment of the Technological Options Available to the Nuclear Industry in the 1980 to 2000 Period," Institute for Energy Analysis, January 1983.

7. Gas-Cooled Reactor Associates and General Atomic Company, "High Temperature GasCooled Reactor Steam Cycle/Cogeneration Design and Technology Development Plan for Nuclear Steam Supply System, Volume II Summary," GCRA82-011, July 1982.

8. George, R. A. and Paulson, C. K., "A Nuclear Plant Design for the 1990's-Meeting Tomorrow's Needs," presented to the American Power Conference, 1983.

9. Hannerz, K., "Towards Intrinsically Safe Light Water Reactors," Institute for Energy Analysis, June 14, 1982.

10. Kemeny, J. G., Chairman, "Report of the President's Commission on the Accident at Three Mile Island," Washington, D.C., October 1979.

11. Leggett, W. D., "Advances in Nuclear Power," presented at the Second Joint Nuclear Engineering Conference of the American Society of Mechanical Engineers and the American Nuclear Society, July 26, 1982.

12. Lewis, H., et al., "Risk Assessment Review Group Report to the U.S. Nuclear Regulatory Commission," NUREG/CR-0400, September 1978. 13. Martel, L., et al., "Summary of Discussions With Utilities and Resulting Conclusions," Electric Power Research Institute, June 1982.

14. McDonald, C. F. and Sonn, D. L., "A New Small HTGR Power Plant Concept With Inherently Safe Features-An Engineering and Economic Challenge," presented at the American Power Conference, Apr. 18-20, 1983.

15. MHB Technical Associates, "Issues Affecting the Viability and Acceptability of Nuclear Power Usage in the United States," Dec. 28, 1982. 16. "Mounting Reports of Failures in Electrical Breakers With Undervoltage Trip," Nucleonics Week, vol. 24, No. 14, Apr. 7, 1983.

17. National Electric Reliability Council Generating Availability Data System, "Ten Year Review 1971-1980 Report on Equipment Availability." 18. Office of Technology Assessment, U.S. Congress, Nuclear Powerplant Standardization: Light Water Reactors, OTA-E-134, April 1981.

19. Phung, D. L., "Light Water Reactor Safety Since the Three Mile Island Accident," Institute for Energy Analysis, July 1983.

20. Prelewicz, D. A., et al., "Nuclear Powerplant Size," ENSA, Inc., Nov. 19, 1982.

21. Robertson, J. A. L., "The CANDU Reactor System:

An Appropriate Technology," Science, vol. 199,
Feb. 10, 1978.

22. Smith, D., "A Reactor Designed for Sizewell," New Scientist, vol. 98, No. 1329, Oct. 28, 1982. 23. Tiren, I., "Safety Considerations for Light Water Reactor Nuclear Power Plants: A Swedish Perspective," Institute for Energy Analysis, ORAU/IEA-83-7, May 1983.

24. U.S. Department of Energy, "Heavy Water Reactors, Preliminary Safety and Environmental Information Document, Vol. II," Nonproliferation Alternative Systems Assessment Program, January 1980.

25. U.S. Department of Energy, "High Temperature Gas-Cooled Reactors, Preliminary Safety and Environmental Information Document, Vol. IV," Nonproliferation Alternative Systems Assessment Program, January 1980.

26. U.S. Nuclear Regulatory Commission, "Unresolved Safety Issues Summary," NUREG-0606, vol. 4, No. 3, Aug. 20, 1982.

27. Weitzberg, A., et al., "Reliability of Nuclear Power Plant Hardware-Past Performance and Future Trends," NUS Corporation for the Office of Technology Assessment, NUS-4315, Jan. 15, 1983.