The Amateur Nuclear Engineer
see two excellent comments to my pointless post below.
The second makes me wonder about how to make nuclear reactors safe.
One key aspect of extremely dangerous nuclear reactors is that the coolant is not the moderator.
Nuclear reactors require a moderator (sp?) to slow down neutrons so they are not absorbed by Uranium 238 but only by Uranium 235 which then fissions. Water is a fairly good moderator and is used in pressurized water reactors (standard in the USA). They have the advantage that loss of coolant is loss of moderator which slows down the reaction.
Another moderator is graphite. Graphite moderated reactors are terrible, because the graphite is solid and stays there causing fission no matter what. The Chernobyl reactor was graphite moderated. The Windscale reactor (which had to change its name to Sellafield to escape its aweful reputation) which had a non fatal fire is graphite moderated. The Yongbon reactor used by North Korea to make Plutonium for bombs is graphite moderated (graphite is also a poor moderator so neutrons are absorbed by Uranium 238 which turns into Plutonium 239).
Another very bad choice is liquid sodium (which is one of the most flamable substances known to man and hot). This was used in the Fermilab reactor star of the book "We Almost Lost Detroit".
Also the French use Sodium.
But wait, you ask, if there are all these adequate moderators, why did the Germans go to such length to obtain and purify heavy water during WWII ? The reason is that nuclear scientists at the kaiser Wilhelm institute in Berlin calculated that graphite wouldn't work. I prefer to assume that they were trying to sabotage the Nazi nuclear energy project and are world historic heroes.
They may have made an arithmetic mistake however.
heavy water is still the best moderator known. Also it is, you know, water. I mean liquid and chemically inert. Canada has relatively safe candu reactors which use unenriched Uranium and heavy water (worthwhile Canadian initiative strikes again -- what a bore).
I wonder if using heavy water as a moderator makes nuclear reactors safer as loss of coolant which is an excellent moderator would shut them down.
The general line in naturally safe reactors is based on two ideas. One is to make the fuel in "pebbles" of uranium coated with something tough and heat resistant. This keeps bits of Uranium apart. It also facilitates disposal. The other is "modular" construction with many small reactors (I think making steam for centralized generators).
I think that nuclear power generation got its awful reputation because of a few awful reactors, that people who actually work on this professionally can design an inherently safe reactor and that, so long as the question is "what is the worst possible accident" not "how likely is that accident" we can discuss the risks rationally. In the distant past, the question was the first, when people discovered that the answer was "catastrophic" they began guessing probabilities which were worthless.