Woah, that's not the end of the story! I thought you said you'd done physics?
It's worse than that -- I worked for Nuclear Electric (now called British Energy to get that scary N-word out of the way) for a few years, in the Barnwood office. I worked on computer modelling software that calculates from neutron flux and coolant temperature data to deduce conditions inside the fuel pins. I've also stood on top of a fuel pond and marvelled at the Cherenkoff radiation.
The vital issue which makes nuclear power and reactor design such a tricky business, is that after you've shut the thing off, you've got decay heat to deal with.
Yes, I accept your very good point that I've just made a handwavy stupid argument as to why Torchwood's handwavy stupid argument is stupid. Mae culpa!
Once the batteries go flat, the water in the core will quickly boil off.
Pedant: only one reactor in the UK (Sizewell B) uses water as the coolant, the others are all CO2 cooled. I wonder what boiling off what would like. Sizewell B's outlet channel temperature is about 450degC if memory serves, and the water is held liquid with some stupendous amount of pressure; the pipework around it is quite impressive to behold. As temperature rises, eventually the pipework will give way and I'm willing to bet that would be spectacularly
Then the fuel channels will melt.
Or, excitingly, catch fire in an old school Magnox reactor (with solid Uranium fuel). That's quite fun apparantly.
Then the molten fuel, metal and other stuff will pool in the bottom of the reactor vessel - that's a meltdown. Heard of Three Mile Island? If fission begins again in the pool of stuff, then you've potentially got a China syndrome.
Now, I'm fuzzy on this bit. The question is, once you've flooded the reactor with a neutron poison, how long does it take to dissapate heat; you'll certainly put a stop to the neutron chain reaction damn fast, as that isn't actually all that stable in a uranium based reactor anyway (plute based fast breeders are a different story). My instincts are that, unless some other factor intervenes, the core cools off quite quickly once the neutron flux is eliminated. I must confess though that I have no data to back this up.
The two famous exceptions have other factors: Three Mile Island had been run for hours with a coolant release valve left open, which meant there was far less coolant than their should have been, and Chernobyl was running the reactor much hotter than it should have been due to a surplus of Xe-135. Neither reactor was SCRAMmed until it was tool late, so I don't think either accident invalidates my reasoning.
The point is: you can't walk away from present-day nuclear reactors, because they are not inherently safe. Making them safe requires a lot of very complex engineering band-aids, and everything to work as it should.
This is very true. As with aircraft, only more so, there are systems, and redundant systems, and backups, and backups of the backups, and redundant backups of the backups, and so forth. But it's a hell of a lot safer than it is usually presented in popular fiction, which usually suggests that at any second they can just shoot off into meltdown mode at the drop of a hat.
Also, fascinating semi-related link