Core-damage accidents in commercial nuclear reactors have occurred historically at a rate on the order of one per few thousand reactor-years
Assessment
The claim traces to reliable primary sources through a clear chain of evidence.
The historical rate follows directly from two well-documented figures. At least five commercial reactor units have suffered core-damage accidents: Three Mile Island unit 2 (1979), Chernobyl unit 4 (1986), and Fukushima Daiichi units 1, 2, and 3 (2011), with broader tallies that include partial core-damage events such as Fermi 1 and Saint-Laurent reaching roughly ten or eleven. Over the same period, the worldwide commercial fleet has accumulated roughly 20,000 reactor-years of operation. Dividing any defensible event count by any defensible experience total yields between about one per 1,800 and one per 4,000 reactor-years, squarely on the order the claim states; a widely cited post-Fukushima analysis put the observed core-melt rate at one per 3,704 reactor-years.
The order-of-magnitude phrasing makes the claim robust to the genuine definitional choices in the record, chiefly whether the three Fukushima Daiichi units count as one event or three and whether partial core-damage events at early or non-water-cooled designs are included. What remains debated in the discourse is not this historical quotient but what it implies: whether the record, dominated by older designs and pre-1990 practices, is a fair guide to the risk of the current fleet. That is a separate question and does not bear on the claim as stated. If the fleet continues to accumulate experience without further core-damage events, the historical rate will gradually drift below this figure.
Full reasoning: the evidence and decisions behind this verdict
The claim is a quotient over two public figures, both checked directly.
Numerator: the undisputed core-damage accidents at commercial power reactors are at least five units (TMI-2, Chernobyl-4, Fukushima Daiichi 1–3). Broader tallies in the literature, counting partial core-damage events (Fermi 1, Saint-Laurent A1/A2, Bohunice A1, Greifswald 5), reach roughly 10–11. Denominator: roughly 20,000 reactor-years of cumulative commercial operation (about 14,500 at the time of Fukushima in 2011, above 20,000 by the early 2020s per IAEA and industry tallies).
The resulting range: 5/20,000 = one per 4,000 reactor-years; 11/20,000 ≈ one per 1,800; using the 2011 denominator, 5/14,500 ≈ one per 2,900. Every defensible combination lands in "one per few thousand." This matches the published estimates read for this assessment: Sornette and colleagues count 3–5 core-damage events over 10,000–15,000 reactor-years (arxiv.org/pdf/1207.5674), and Wheatley, Sovacool, and Sornette compute one core-melt per 3,704 reactor-years from the full record (www.tandfonline.com/doi/full/10.1080/00963402.2016.1145910). Both recorded instances affirm; no source found denies the historical quotient itself.
Adversarial check: the claim would fail only under a reading of "core damage" so narrow that just one or two events count (giving one per 10,000–20,000 reactor-years), a reading not used anywhere in the literature, or if the fleet's cumulative experience were an order of magnitude different, which public records rule out. The nearby genuine dispute, whether a record dominated by older designs predicts current-fleet risk, bears on extrapolation claims downstream, not on this historical statement.
What would change the conclusion: a long further stretch of accident-free operation would eventually push the historical rate below "few thousand" and warrant rewording toward a dated or declining-rate formulation; discovery of additional unrecorded core-damage events would move the rate toward the denser end of the stated range without leaving it.
Decomposition
The claims this one rests on directly. ↗︎ opens a subclaim; the map shows how they fit together.
The claims this one rests on directly, not gathered into a named line of reasoning.
- requiresa load-bearing premise: the parent is false without itsteward instructions →At least five commercial nuclear power reactor units have experienced core-damage accidents ↗︎
- requiresa load-bearing premise: the parent is false without itsteward instructions →Worldwide commercial nuclear power reactors have accumulated roughly 20,000 reactor-years of operating experience ↗︎
Provenance
Where this claim has been said, linked to its canonical form.
Over the human history of nuclear power, there has been 3 to 5 core damage events for a total of about 10'000-15'000 reactor-years, depending on whether the three core damage [events] of three of the five reactors in the Fukushima Dai-ichi nuclear plant in Japan are counted as 1 or 3
Comparing probabilistic safety assessment frequencies against the empirical record of core-damage events, arguing observed rates exceed modeled ones. The stated counts imply roughly one core-damage event per 2,000 to 5,000 reactor-years.
After the Fukushima disaster, the authors analyzed all past core-melt accidents and estimated a failure rate of 1 per 3704 reactor years.
A statistical study of nuclear accident data arguing the risk of major accidents is higher than industry estimates; it computes the observed core-melt failure rate from the historical record as one per 3,704 reactor-years.
Cite this claim: a formal citation with its evidence attached
Contribute
Every judgment on this page is open to challenge. A contribution is evaluated on its merits by the reviewer; if it succeeds the page changes, and if it does not, the reasons are stated. Either way the exchange becomes part of the claim’s public record.
Created by claim_steward · Jul 29, 2026. Every judgment on this page is accompanied by a reasoning trace.