Planning ahead for advanced reactor safeguards and security

The National Academies of Sciences, Engineering, and Medicine was charged by Congress in the fiscal year 2020 Appropriations Act to form an expert committee to produce two reports, with consensus findings and recommendations for stakeholders, on advanced reactor designs that could be commercially deployed by 2050. The work of the committee on fuel cycles and technology options is being conducted in parallel with another study, Laying the Foundation for New and Advanced Nuclear Reactors in the United States.

Safeguards considerations: Jeremy Whitlock, section head for Concepts and Approaches in the Department of Safeguards at the International Atomic Energy Agency, acknowledged that while advanced reactor concepts may not be new—some were conceived decades ago—the application of IAEA safeguards to those technologies is new. Whitlock said that safeguards considerations for advanced reactors include new fuels and fuel cycles, new reactor designs, longer operation cycles, new supply arrangements, new spent fuel management, diverse operational roles, unattended monitoring systems, and remote distributed locations.

“For most of these advanced reactors,” Whitlock said, “we are well ahead in the development curve, and we have time to consider all of these issues. Technical solutions need time. . . . Everything is solvable, but the technology may be more complicated or may not even exist at this time and have to be developed.”

Incorporating safeguards into the reactor design process early on is voluntary, but can benefit vendors, Whitlock said. “We have a primordial soup of designs out there, and everybody wants to be the first advanced reactor to grow legs and walk out on dry land, and to say that you’ve discussed safeguards with the proper authorities . . . is actually a benefit to discussions with potential customers,” he said.

Proliferation risk: Matthew Bunn, co-principal investigator for the Managing the Atom Project at Harvard University’s Belfer Center, said most proliferation risk comes from enrichment and reprocessing, and not from the material characteristics of a particular design. A focus on reactor design characteristics misses most of the proliferation problem, according to Bunn, who said that expertise and infrastructure expansion, not technical processes, have historically posed the greatest proliferation risk.

Material accounting and machine learning: Ben Cipiti, of Sandia National Laboratories, described work on material protection accounting and control technologies, including a program to demonstrate safeguards and security by design for a notional electrochemical reprocessing facility.

Cipiti said that materials accountancy for aqueous reprocessing is well established but has measurement uncertainties that, at a large facility, can add up to a significant amount of a material of interest.

Cipiti said that one potential tool for process monitoring is machine learning, which could permit the use of unattended monitoring systems and reduce the frequency of sampling. “It’s similar to traditional safeguards approaches,” he said. “You’re limited ultimately by measurement uncertainty. With machine learning, maybe you can do a little bit better, but at the end of the day, it’s that fundamental limitation of how well you can measure the materials.”

Physical protection systems: Cipiti is the national technical director for the Advanced Reactor Safeguards and Security program within the DOE’s Office of Nuclear Energy, established in 2020 as part of appropriations for the Advanced Reactor Demonstration Program. The program focuses on reactor sites, and not the full scope of a reactor’s fuel cycle.

Security risks include theft—which could be abrupt or gradual—and sabotage, and physical protection systems are designed to meet those risks. Advanced reactors are typically designed with passive safety features. Determining whether those features can provide enough time to stop an attack and return a plant to a safe operational state “ultimately requires linking PRA and dynamic PRA tools with physical protection systems so that we can more fully run through these types of scenarios in the future,” Cipiti said.

The more robust a physical protection system is, the higher reactor costs will climb. “It will be interesting to see what will be acceptable from NRC’s standpoint, what that is actually going to cost, and then the fraction of that compared to the cost of the actual reactor itself,” Cipiti said.

HALEU: “Why does it matter if one is using 5 percent enriched or 20 percent enriched fuel, when they’re both LEU?” was a question posed and then answered by Warren Stern, deputy department chair of the Nonproliferation and National Security Department at Brookhaven National Laboratory. “The answer is,” he said, “when you’re talking about enriched material, the amount of effort that’s required to re-enrich it to something that might be of weapons use is less.”

While the NRC distinguishes between 5–10 percent and 10–20 percent HALEU, the IAEA does not currently have a separate safeguards category for HALEU, which is now in limited use worldwide but could be used broadly in advanced reactors. Widespread use of HALEU could have a significant impact on safeguards, Stern said. The frequency of inspections might have to be stepped up, for example, and the impact depends on the specific design of the reactor.

More available online: The preceding highlights reflect only the first day of the public meeting content. Several additional presenters, including experts from the National Nuclear Security Administration and the Nuclear Regulatory Commission, addressed the committee on May 18, on topics including the security of HALEU at fuel cycle facilities and in transit, proliferation risks of the laser enrichment of uranium, defense nuclear nonproliferation, and safeguards and material accounting topics.

A recording of the meeting and all presentation materials will be made available online, along with materials from other past public meetings of the committee.