Preprint flags hours 44-48 as flammable in a modeled NuScale hydrogen accident
A thermodynamic preprint applies a 1112.5 kelvin flame-temperature cutoff to rich hydrogen in small modular reactor containments and finds a four-hour flammable window in one NuScale severe-accident case. The result matters because compact containments can concentrate accident hydrogen, but it is a model checked against 21 historical points, not a plant test.
In an arXiv preprint, Sihyeong Yu proposes a thermodynamic test for whether hydrogen-rich gas in a small modular reactor containment could burn during a severe accident. The calculated adiabatic flame temperature is used to argue that lean-hydrogen rules miss a late window in NuScale case LCC-05T-03, when oxygen from water radiolysis builds beside hydrogen from hot zirconium cladding. The paper is a calculation, not a reactor experiment, and it has not been peer reviewed.
The criterion was checked at 21 points on the measured upper flammability limit in the U.S. Nuclear Regulatory Commission FITS hydrogen-air-steam database. From 5 to 50 percent steam, that limit temperature stayed within 17 percent of a steady value, and the minimum, 1112.5 kelvin, was adopted as a conservative threshold. In the NuScale scenario, the threshold marks hours 44 to 48 as flammable: about 8 to 10 percent oxygen, continued hydrogen generation, and low steam.
The pattern is familiar from Fukushima in 2011, when zirconium-steam reactions made hydrogen that burned once it met oxygen, damaging reactor buildings. SMR developers argue that a small, strong containment is safer, yet the same compactness can push a damaged core toward hydrogen-rich mixtures sooner. There oxygen is the scarce reactant, and the Lewis number, heat diffusion divided by mass diffusion, exceeds one, the hydrocarbon regime where this flame-temperature test was built. An oxygen-percentage screen alone cannot resolve that passing window, the preprint concludes.
Next, the 1112.5 kelvin cutoff needs a second rich-hydrogen dataset and a sensitivity study of steam, catalytic recombination, and venting before anyone treats hours 44 to 48 as a design fact. Methods note: the evidence is a thermodynamic calculation on 21 historical flammability points plus one modeled accident, so it cannot estimate ignition probability or containment failure. Until that check exists, the result is a sharper way to look for a transient risk, not proof that a next-generation reactor would explode.
Sources (1)
- [1]Applicability of a Thermodynamic Model to Rich-Hydrogen Flammability Assessment in SMR Containments(https://arxiv.org/abs/2611.12891)