Safety, emergency planning zones and the new NRC rules

A reactor doesn't stop making heat when it stops the chain reaction. Fission products keep decaying, and right after shutdown that decay heat is roughly 6 to 7 percent of full power, falling to around 1 percent within hours and lower over days. At Fukushima Daiichi in 2011 the reactors shut down as designed. Then the tsunami knocked out the power that ran the cooling pumps, decay heat had nowhere to go, and the fuel melted. Almost every SMR safety pitch is an answer to that one failure: remove decay heat with no pumps, no power and no operator.

The safety and radiation lesson covers the accident record and dose numbers. This lesson is about how the small designs claim to be safer, what the skeptics answer, and how the NRC's rules are changing because of those claims.

Passive safety, family by family

Family	How decay heat is supposed to leave	Main safety argument	Main concern raised
Light-water SMR (NuScale, BWRX-300)	Natural circulation of water; NuScale modules sit in a large pool that absorbs heat	Smaller cores, fewer large pipes to break, passive systems with long coping times	Still high-pressure water and steam; standard LWR accident types remain possible
Sodium fast (Natrium, Aurora)	Natural circulation of sodium; air flow around the vessel	Low pressure; EBR-II shut itself down with pumps off in 1986 tests	Sodium burns in air and reacts with water; fast-reactor cores can respond differently to some upsets
Helium gas with TRISO (Xe-100, Kaleidos)	Conduction and radiation out through the core and vessel	TRISO particles hold fission products at very high temperatures; helium is inert	Air or water getting into a hot graphite core; quality control across billions of particles
Salt-cooled (Kairos)	Natural circulation of salt, which has a high boiling point	Low pressure; salt holds many fission products chemically	Corrosion and salt chemistry over decades; little operating history
Heat pipe (eVinci)	Sealed heat pipes, no pumps	Very few moving parts; small inventory	Failure of many pipes at once; unproven at power

Small size helps physically. A small core has more surface area for its volume, so heat leaves it more easily, and it holds less radioactive material to release in the first place. Radiant's test plan at INL's DOME includes at least 150 hours at full power with no operator intervention, which is the kind of demonstration that turns a claim into data.

The skeptics' case

The fullest skeptical case is the Union of Concerned Scientists' 2021 report "Advanced" Isn't Always Better, by physicist Edwin Lyman. It concluded that the non-light-water designs it reviewed are not likely to be significantly safer than today's plants, and that some pose greater safety, security, proliferation or environmental risks. Lyman's points include sodium's chemistry, the possibility of air entering hot graphite in gas reactors, and the lack of operating data for salt designs.

There are wider concerns too. Hundreds of small units spread across many sites mean more places to guard and inspect. Smaller staffs and remote monitoring, central to SMR economics, leave fewer people on site when something odd happens. And until a design has run for years, its safety case rests on analysis and tests.

Developers and DOE answer that the designs remove the station blackout failure that caused the Fukushima meltdowns, that the EBR-II tests and decades of TRISO testing are real data, and that light-water SMRs inherit a 60-year operating record. Where the evidence points: the physics of passive decay heat removal is well understood and has been demonstrated in test reactors; whether commercial units deliver it reliably for decades is not yet shown, because none has operated.

Emergency planning zones

Every large U.S. plant has a plume exposure planning zone of about 10 miles, where officials plan sheltering and evacuation, and an ingestion pathway zone of about 50 miles for food and water protection. Counties inside the 10-mile zone run sirens, drills and evacuation plans, with plant owners funding much of it.

In November 2023 the NRC published a rule titled "Emergency Preparedness for Small Modular Reactors and Other New Technologies" (88 FR 80050), effective December 18, 2023. It lets SMRs (defined as up to 1,000 MWt per module), non-light-water power reactors and some other facilities opt into a performance-based program. The plume exposure zone is set by a site-specific analysis: the area where a release could give the public more than 1 rem total effective dose over 96 hours, plus the area where prompt protective measures are needed. There is no fixed distance. The zone could be larger or smaller than 10 miles, depending on the design and the analysis, and the NRC reviews each case. Historically, a few small plants (Fort St. Vrain, Big Rock Point and La Crosse) used 5-mile zones.

Supporters say the dose-based approach is just honest physics: a small reactor with a small source term shouldn't carry a large reactor's zone. Skeptics answer that the analysis depends on accident scenarios chosen by the applicant, and that a zone at the fence line leaves local responders with little planning if the analysis turns out wrong.

Part 53 and the proposed Part 57

The NRC's Part 53 final rule was published in the Federal Register on March 30, 2026 (91 FR 15696). It is an optional licensing framework for any reactor type. Provisions listed on the NRC's page include flexible use of probabilistic risk assessment, generally licensed reactor operators for certain facilities, alternative siting criteria that allow higher-population sites (areas with more than about 25,000 residents), factory fuel loading, load-following provisions, no single failure criterion requirement and no traditional aircraft impact assessment requirement.

A separate proposed rule, Part 57, targets microreactors. ANS reported in April 2026 that it was set for Federal Register publication in May 2026; NRC Chairman Ho Nieh said it is meant for reactors of 100 MWe or smaller. It would allow approval of fleets of identical reactors, alternative design standards, streamlined environmental reviews for limited-impact projects, and some construction before NRC permitting. The NRC's advanced reactor page, updated September 21, 2026, still listed it as a proposed rule. New York, and a group of ten attorneys general including California's, filed comments in June 2026.

Both rules sit inside the larger fight over NRC speed and independence covered in who regulates nuclear power. Executive Order 14300 also told the NRC to reconsider the linear no-threshold model and the ALARA principle for radiation dose, which affects every safety analysis, including emergency zone sizing.