Although there was an early interest in severe accidents for LMRS, the analysis of those accidents did not progress very far before the start of the safety review of the Fast Flux Test Facility (FFTF) and the licensing review of the Clinch River Breeder Reactor (CRBR). In a letter of May 6, 1976, the NRC staff informed the CRBR Project Office of NRC's position on core disruption accidents (severe accidents). Basically, the letter said that core disruption accidents would not be considered as part of the design basis for CRBR, but they would be used as the basis for additional measures to be taken to limit consequences and reduce residual risk to ensure that the public health and safety was adequately protected. The research in LMR severe accidents during this period was oriented to better understanding the phenomena and devise ways to mitigate the effects of severe accidents. The NRC report, "An Analysis and Evaluation of the Clinch River Breeder Reactor Core Disruptive Accident Energetics," NUREG0122, published in March 1977, was typical of studies during that time. One of the severe accident issues discussed in that report that continues to this day for the LMR is the relatively large positive sodium void coefficient of reactivity. Although the CRBR mixed-oxide fuel design is quite different than the present metal-fuel LMR design and the LMR designers have reduced the probability of sodium voiding considerably, that troublesome characteristic still is intrinsic to the design.

With the exception of the positive sodium void coefficient issue, LMR designers have solved many of the safety problems associated with severe accidents. Proof testing under bounding accident conditions at the EBR-Il facility has contributed to the understanding of severe accidents. Added to this is PRA experience from the CRBR, from the Large-Scale Prototype Breeder project sponsored by DOE and EPRI, and from the more recent LMR design efforts--the Sodium Advanced Fast Reactor (SAFR) and the PRISM design by General Electric.

HWR Severe Accident History

DuPont has embarked on a major Probabilistic Risk Assessment program for the Savannah River heavy water reactor facility. This full-scope PRA is over 50% completed and, when finished in 1989, should provide a large body of information regarding the probability and consequences of severe accidents for that type of plant. This would provide a good starting point for extrapolation of the severe accident results to the HWR design candidate for the NPR.

The only other country to make significant use of heavy water technology is Canada in its unique pressure tube CANDU reactor. There has been no significant relationship between the U.S. and the Canadian HWR programs.

5. What are the possible solutions to any problems expected in providing adequate safety assurances for the NPR?

The process for assuring safety for the NPR will be complex and controversial. It is important for the DOE decision-makers to be aware of the difficulties that will arise in demonstrating adequate safety and to give consideration to ways to minimize them. This report has attempted to do that by describing the key elements that make up safety assurance today and by comparing the four candidate technologies against these key elements. From a regulatory, safety-assurance perspective, the candidate that looks most promising in minimizing these problems is the HTGR.

The key elements that make up safety assurance are:

The process of demonstrating adequate safety based on NRC policy
and practice;

The public and Congress as major contributors in establishing
safety assurance;

The role of severe accidents in determining adequate safety; and

The importance of the advanced reactor attributes--simplicity and
inherent safety--in providing adequate safety assurance.

These elements, taken together, constitute a significant new era in demonstrating safety for nuclear plants in the United States, especially for the NPR, which the DOE is determined to make at least as safe as modern commercial LWRS. Safety assurance today is substantially different from what it was a decade or even five years ago. Further, safety assurance has been markedly different in the commercial sector than in the defense sector. But DOE recently has set in motion an extensive program to upgrade safety assurance in the defense sector, bringing it more in line with that of the commercial sector.

When tested against these elements of the new era of safety assurance, the HWR has the most difficult problems ahead in assuring safety. Although significant problems exist for the LMR and for the LWR, they do not appear to be as problematic as the HWR. Finally, although problems also exist for the HTGR, it is the best choice from the safety assurance perspective provided in this report.

The Heavy Water Reactor

The HWR technology is only now, in 1988, beginning to employ modern regulatory and safety assurance practices. In virtually very aspect of the safety assurance process, the HWR has been shown to be seriously lacking. It has never participated in regulatory reviews, reviews that DOE will be insisting on for all of its hazardous facilities. To conform to a modern process will be a major undertaking. As revealed in the recent safety study of the Savannah River reactors by the National Academies, the HWR concept does not have the technological base to demonstrate safety. Codes, proof testing, and experimental data are lacking. And the design changes in going from the Savannah River design to the NPR design further complicate the situation. The public and the Congress will consider the HWR an unknown quantity. Their only acquaintance with the HWR is through the Savannah River Plant, which is a problem in its own right. The public and Congress have perceptions of safety that do not always coincide with the realities of safety. Perceived problems, which may not be real, are often correlated with unknowns or uncertainties in safety assessment, but it is the perceived problems that will drive the decision process.

HWR severe accident technology has not kept pace with the rising importance of severe accidents as an ingredient in safety assurance. Only recently have PRAs been undertaken for the HWRS at Savannah River. Severe accident research that addresses the core melt phenomena for the HWR is in its infancy.

While the HWR proponents claim that the HWR has certain safety attributes proposed by the NRC in the Advanced Reactor Policy Statement, the approach to inherent safety and simplicity is only piecemeal. Also, the passive cooling capability claimed for the new HWR designs has not been analyzed or tested.

The Liquid Metal Reactor

The LMR has an advantage over the HWR in the area of safety assurance process. Although the Clinch River Breeder Reactor (CRBR) was a different design than would be used for the NPR, there would be benefit from the CRBR licensing experience which included the completion of a Safety Evaluation Report and a construction permit hearing prior to cancellation of the project. More importantly, the commercial version of the LMR is pending before the NRC for a licensability review. However, there is no licensed LMR in the United States. Further, the CRBR review was not conducted against modern safety assurance standards, and operating experience that would be helpful for the LMR is limited.

As with the HWR, the LMR is not well known in Congress or among the American people. Thus, there is the risk that the LMR will be at the same disadvantage as the HWR; namely, in the public's eye, unknowns are equated with perceived risk. Further, the LMR has a relatively large positive sodium void coefficient, about the same size as the coolant void coefficient at Chernobyl. To dismiss the implications of this design characteristic on probabilistic grounds, even if sound technically, is difficult. Although LMR research in severe accidents is continuing, many of the severe accident issues have been resolved for the LMR.

The advanced LMR designs proposed for power plants have advantages over the HWR in inherent safety and in simplicity of operation. They also conform to the spirit of the NRC's advanced reactor policy statement.

The Light Water Reactor

The LWR has far more experience in the licensing and safety assurance process than the other three technologies. Also, the LWR operations experience is vast. But three important problems should be pointed out. First, both of the LWR candidates, the Westinghouse upgrade of its 4-loop design and the WNP-1 B&W unit, are old by advanced light water reactor standards. Second, the advanced light water safety standards are in a state of flux. Third, the NPR version of either the Westinghouse or the B&W-designed reactors will require redesign of the NSSS. Thorough evaluation of their safety would be warranted.

The LWR is not an unknown with the public or Congress. This is both an asset, in the sense that they have been educated regarding the LWR, and a liability, in the sense that they are aware of the accident at TMI, the B&W incidents at Davis Besse and Rancho Seco, and other LWR safety problems that have been visible over the past twenty years. The UCS and other intervenors have been vocal in their opposition to LWRS over the years and can be expected to raise many of the unresolved issues again.

The uncertainties related to severe accidents for the LWR are still large and continue to frustrate the resolution of major issues. For example "direct containment heating" is a major uncertainty which when resolved may mean backfits for some pressurized water reactors. The LWR is at a distinct disadvantage relative to the other three candidates since it is such a complicated machine with a potential for release of radionuclides not shared with the other three candidates.

The light water reactor requires active systems to assure safety. The motivation for the NRC's Advanced Reactor Safety Policy Statement was in large measure due to the problems with LWRs.

The High Temperature Gas-Cooled Reactor

The HTGR, compared to the other candidates, is in an advantageous position for demonstrating safety. It has up-to-date regulatory and operating experience at Ft. St. Vrain that cannot be matched by the HWR or the LMR. Further the commercial version of the modular HTGR is the lead advanced reactor under safety review consideration by the NRC. It leads the LMR in that regard, while the HWR and the LWR are not even competing.

The HTGR has a unique spot in relations with Congress and the general public. First it is a well known concept in Congress. In fact its year-by-year support comes from Congress. Further, the HTGR has been well received by groups that are traditionally hostile to nuclear power. This suggests that support outside of the technical establishment may be strong.

The HTGR risks from severe accidents have been known to be low for a long time. The fact that Ft. St. Vrain has no containment--the only licensed facility in the country without one-is evidence of this. Although there are some unresolved issues with the

NRC, the NRC staff believes the problems can be resolved; in July 1988, the NRC staff concluded that the commercial version of the HTGR will need no evacuation plan.

Simplicity and inherent safety are important attributes of the HTGR. No active systems are needed for shutdown and heat removal for weeks after an accident within the design basis.

Concluding Remarks

It is the conclusion of this report that, from a safety assurance perspective, the HTGR is the reactor of choice for the NPR. Stepping back from that perspective and considering some broader issues that affect safety of nuclear power in general, several further points are appropriate, all of which further support choosing the HTGR technology for the next NPR.

The federal budget for nuclear technology, apart from the production
of nuclear defense materials, has been relatively constant for several
years, and all safety research has come out of that budget, whether it
is for commercial power or for DOE and military applications. If the
decision is made to go with the HWR, it will require large safety
expenditures for a technology unique to the NPR, thus diverting
needed funds from other important safety research programs. It could
detract from reactor safety research applicable to the commercial
sector.

Both the commercial and the military applications sectors could
benefit considerably from a shared safety technology. Safety
information developed for the commercial reactors could be of use by
the NPR and vice versa. The advantages of sharing safety
technology cannot be over estimated. This could not happen if the
HWR was chosen as the NPR.

It has been said that choosing a technology other than the HWR
would compromise the distinction between nuclear reactors for
weapons production and nuclear reactors for power production. This
argument may have some merit for one of the LWR candidates for the
NPR because that candidate is a pressurized water reactor that was
initially designed and built for commercial applications. However, it is
worth noting that the pressurized water reactor technology now used
around the world evolved from the U.S. Navy's nuclear propulsion
technology.

Finally, DOE, through its NPR program could make a significant
contribution to the future energy security of the United States by
choosing a reactor type that has potential to meet future energy needs
in a safe and environmentally acceptable way. In this way, the
Department also will serve to limit the number of nuclear reactor
designs in the United States. This is a problem of safety significance
that has plagued the country for a generation and that the Energy
Research Advisory Board has urged the Department to solve.

APPENDIX A

CREDENTIALS OF AUTHORS

Dr. James F. Meyer has a background in reactor safety and licensing that spans more than 20 years. He is an authority in the field of severe accident behavior (core melt accidents) of nuclear plants. After 5 years at Argonne National Laboratory where he performed reactor physics experiments, he joined the Nuclear Regulatory Commission where he spent more than 10 years dealing with severe accident analysis and other safety issues. Among the projects that he worked on were the Clinch River Breeder Reactor, the Fast Flux Test Reactor, the Zion and Indiana Point pressurized water reactors, and the Limerick power station. He also performed safety and "licensability" studies for candidate designs that were part of President Carter's non-proliferation program. He has policy level experience with severe accident and advanced reactor issues as the technical assistant to a Commissioner of the NRC from 1983 to 1987. Since leaving the NRC, Dr. Meyer has served as a consultant on severe accident safety issues.

Dr. Roger J. Mattson has more than 20 years of experience in nuclear reactor safety, most of which were spent at the Nuclear Regulatory Commission from 1967 to 1984. For seven of those years at NRC he directed most of the technical review of all licensed nuclear plants in the United States. He was a principal contributor to the NRC's Action Plan issued in 1980 after the accident at Three Mile Island, the NRC's Severe Accident Policy Statement issued in 1985, and the Safety Principles issued in 1988 by the International Nuclear Safety Advisory Group of the International Atomic Energy Agency. Among other assignments since leaving the NRC, Dr. Mattson has participated in design reviews of the production reactors at Hanford and Savannah River, served on the UNC/NI Corporate Nuclear Review Board for N Reactor, served on the Nuclear Review Board for Philadelphia Electric Company, and assisted the NRC in the interpretation of the accident sequence at the Chernobyl reactor in the Soviet Union.

[From The Energy Daily, Washington, DC, Monday, July 25, 1988]

HTGR BACKED FOR WEAPONS REACTOR

From a safety point of view, the Department of Energy should pick the High Temperature Gas Cooled Reactor when it selects the technology for its new weapons production reactor, according to a report by two veteran nuclear scientists scheduled for release this week. DOE is expected to make a multibillion dollar decision shortly on what kind of reactor to buy to produce tritium, a key ingredient in nuclear bombs. Four different reactor technologies are competing for the contract.

"When the alternative technologies are viewed from the perspective of providing the highest probability of demonstrating adequate safety in a timely and broadly acceptable fashion, the High Temperature Gas Cooled Reactor has significant advantages over the Light Water Reactor, the Liquid Metal Reactor and the Heavy Water Reactor," write Roger Mattson and James Meyer in the new study. "It has a better technical basis for responding to determined intervention, it has a proven track record in the NRC licensing process and it has lower severe accident risk. The HWR presents the highest risk because of the lack of regulatory background and the uncertainties surrounding severe accidents."

Mattson and Meyer are former NRC officials with extensive experience in nuclear reactor safety. Their study was commissioned by the National Defense Council Foundation, a 10-year-old Washington group that describes its purpose as being "in defense of free enterprise, country and Constitution."

If DOE decides to go with an HWR for its new tritium plant, the report says, large safety expenditures will be required for a technology unique to the new production reactor, thus diverting needed funds from other important safety research programs. It could detract from reactor safety research applicable to the commercial sector.