Energy

Why a Nuclear Reactor Can't Just Be Switched Off

By Lucian — builder & engineer, LK Forge

Dropping the control rods stops the chain reaction in under a second, but the reactor keeps making real heat for weeks afterward. We benchmarked the Nuclear Energy lab's Way-Wigner decay-heat model for a 3000 MW reactor after a full year of operation, sampled from one second to one month after shutdown, to see exactly how much heat is left and how long it takes to fade.

 ·  5 min read  ·  measured from the Way-Wigner decay-heat model

6.3912%
of full thermal power (191.74 MW) still generated the instant after SCRAM
1.0743%
still there a full hour after shutdown
~1.35 h
for decay heat to fall to 1% of thermal power
~56.9 days
for decay heat to fall to 0.1% of thermal power

The Way-Wigner curve

Decay heat isn't one thing decaying — it's hundreds of different fission-product isotopes decaying at once, each on its own clock. Summed together, the classic Way-Wigner approximation gives decay power as a fraction of the reactor's steady-state thermal power, Pth, as a function of the time since shutdown (tafter) and how long the reactor ran beforehand (toperate):

decay fraction = 0.066 × ( tafter−0.2 − (tafter + toperate)−0.2 )

Both terms fall as a slow power law, t−0.2 — not an exponential, which is why the tail is so long. An exponential decay would drop by a fixed percentage every fixed time interval; this power law keeps dropping, but ever more slowly, which is exactly the shape in the chart below.

Fast drop, long tail

Decay heat (as a percentage of full thermal power) versus time after shutdown, for a 3000 MW reactor that ran a full year before SCRAM. The time axis is log-scaled because the sampled range spans six orders of magnitude, from one second to one month.

0% 1% 2% 3% 4% 5% 6% 7% 1 s1 min1 hour1 day1 month time after shutdown (log scale) decay heat

The exact numbers

Every row below is a direct sample from the lab's decay-heat model for a 3000 MW reactor after a full year of operation.

Time after shutdown Decay heat (% of Pth) Decay heat (MW)
1 s 6.3912% 191.74
10 s 3.9555% 118.66
1 min 2.7013% 81.04
10 min 1.6273% 48.82
1 hour 1.0743% 32.23
8 hours 0.6378% 19.13
1 day 0.4709% 14.13
1 week 0.2525% 7.57
1 month 0.1387% 4.16

Time to fall to 1% of Pth: 1.35 hours. Time to fall to 0.1% of Pth: 56.9 days. Reactor: 3000 MW thermal, 1-year operating history. Generated 2026-09-10.

Try it yourself

Open the lab, SCRAM the reactor, and watch decay heat and core temperature race against the cooling system in real time.

Open the Nuclear Energy lab →
Share this X Facebook Reddit

Related reading

Common questions

Why does a shut-down nuclear reactor still produce heat?

Splitting uranium produces hundreds of different fission-product isotopes, and most of them are radioactive. Shutting down the reactor (a SCRAM, dropping the control rods) stops the fission chain reaction within a fraction of a second, but it does nothing to those already-created fission products — they keep decaying and releasing energy on their own schedules, some over seconds, some over years. That radioactive decay is what's called decay heat, and it does not care whether the chain reaction is running. This benchmark measured it directly from the lab's model: 6.3912% of full thermal power (191.74 MW for a 3000 MW reactor) is still being generated the instant after SCRAM.

How long does decay heat take to become negligible?

Decay heat falls quickly at first and then slows down a lot, because it's really a sum of many isotopes each decaying at its own rate — the fast-decaying ones burn off first, leaving the slower ones to dominate later. In this benchmark (3000 MW reactor, a full year of operation before shutdown) decay heat drops to 1% of full power in about 1.35 hours, but takes roughly 56.9 days to reach 0.1%. That long, shallow tail is exactly why spent fuel needs active cooling for weeks to months after shutdown, and continues generating measurable heat for years after that.

Does it matter how long the reactor ran before it was shut down?

Yes, but less than intuition suggests. Longer operation lets more fission products build up, so a reactor that ran longer before shutdown produces somewhat more decay heat afterward. In this dataset, one hour after shutdown a reactor that had run for just one day was releasing 0.6091% of full power, versus 1.0743% for one that had run a full year — a real difference, but not an order of magnitude. Most of the fission-product inventory that matters for early decay heat builds up within the first days to weeks of operation; running for a year instead of a month adds comparatively little on top of that.