>> you only have to ensure that the residual heat from normal decay doesn't melt the reactor.

Not hard if this is a passive-cooling design. My point is that at these power levels passive cooling isn't enough. Stop the pumps and the reactor may build up enough heat to start melting things very quickly, long before the pool boils dry.

Correct me if I'm wrong but isn't the idea of safe pool reactors that hydrogen-neutron heat transfer is more efficient than normal radioisotope-water heat transfer. If you shut down the pumps and heat increases, spectral hardening cools the reactor faster and disperses the energy into larger volume than convection using pumps.

Is that really true? My understanding is that ordinary boiling water reactors contain safety systems that work just by boiling. Reading the wikipedia article[1], the Isolation Condenser, the Automatic Depressurization System, and the Gravity-driven Cooling System all work just by water boiling and with natural convection in the water, without any active pumps, and they are supposedly enough to keep the core from melting.

[1] https://en.wikipedia.org/wiki/Boiling_water_reactor_safety_s...

Isn't the residual heat from normal decay something like 10% of the reactor output? When you split Uranium you get a lot of byproducts with short half-lives that continue decaying away even when Uranium chain reaction stops. So you've still got 40 MW to deal with, at least until those short-half-lived byproducts finish decaying.

According to one document[1]

> The decay heat produced by a reactor shutdown from full power is initially equivalent to about 5 to 6% of the thermal rating of the reactor. This decay heat generation rate diminishes to less than 1% approximately one hour after shutdown. However, even at these low levels, the amount of heat generated requires the continued removal of heat for an appreciable time after shutdown.

At 400 MW it takes about 4 hours to boil the water in an olympic swimming pool[1], so there should be some safety margin for the power to drop. Then once it is down to 1% decay heat it could boil for a couple of weeks before running dry.

[1] https://web.archive.org/web/20110316042315/http://www.hss.do...

[2] http://www.wolframalpha.com/input/?i=(specific+heat+of+vapor...

5% for the first few seconds: http://www.euronuclear.org/info/encyclopedia/r/residual-heat... but your point of course still stands. It's certainly not trivial to build passive-safe reactors, but it's also not impossible.