The most expensive single part of a nuclear power station is the containment building and support structures for the reactor vessel itself, something that also costs serious bucks. Those are the two parts of a power station that can't be economically replaced during the plant's lifetime. Everything else gets swapped out and replaced over the reactor's lifespan, pretty much. Pumps, steam separators, turbogenerator units, control systems, transformers, switching gear etc. last for a decade or two or three and then get replaced and/or refurbished. It's part of the cost of operating any power plant, the same thing happens with coal and gas plants depending on the hardware involved. A steam turbine turbogenerator set will last 25 or 30 years, for example. In the case of 1970s vintage reactors the old analogue control systems have mostly been upgraded to digital systems with finer control and better reporting, new sensors etc.

As for reusing a site this is being done sometimes but in other cases the grid demands have changed over the decades and a new greenfield site close to a new urban area that's doubled in size in the past fifty years might be better than one closer to, say, a place like Detroit.

The US military tried operating small nuclear power plants in remote land bases such as the South Pole site and even one in Greenland (the Iceworm project). They were finicky, difficult to keep working and generally an economic and logistical failure. It turned out to be simpler and cheaper to fly liquid fuel to such bases to power generators as well as supplying vehicles, aircraft etc.

The WHO numbers are estimates, there aren't real dead bodies like there were at Aberfan or the real body count at various coal mining operations in China and the third world generally. Even the West with higher safety standards has dozens and sometimes hundreds of deaths in coal mines each year -- the single-incident high body counts are widely reported (12 dead in the Sago mine WV in 2006, 29 dead in the Pike River mine in New Zealand in 2010 etc.). The death of one or two people at a time barely breaks the surface.

If the nuclear power industry was slaughtering workers at that rate there would be a world-wide outcry declaring it totally unsafe... wait, there is a world-wide outcry declaring nuclear power totally unsafe. Coal, not a whimper. Weird that...

You can point to, for example, the Aberfan disaster and say "Coal killed a hundred kids" or to the death toll from coal mining and transport year on year and say "Coal killed these workers" (China proudly announced the death toll from coal mining had fallen below 3000 per annum a couple of years back. It used to be a lot higher). That's on top of the mercury, cadmium, radon, sulphuric acid fumes, dioxins, beryllium, arsenic and the thousands of tonnes of other toxic wastes spread through the atmosphere and over agricultural lands and deposited in rivers and oceans every year which kills and maims people who don't work with coal directly. But nuclear power is worse somehow.

Both the Allies and the Axis forces sank millions of tonnes of loaded oil tankers during WWII, not to mention similar tonnages of warships each with many tonnes of fuel oil on board. One of the submerged museum ships at Pearl Harbor was still leaking fuel decades after it was sunk in 1941. As far as I know this extended and untreated oil spillage has had little long-term effect on sealife and the general health of the oceans worldwide.

You mean like "640kB of RAM should be enough for anyone?"

Intel are reducing power consumption and maintaining performance faster than ARM can improve processing power while keeping power consumption down. The current version of the iPad has a lot more processing power than the first one did but it has a battery three times bigger to give it the same endurance between charges, in large part because the newer ARM chips suck more power than their predecessors did.

Intel-based tablets like the Iconia W series (i3/i5) or Toshiba Encore (Atom quad-core) have the same endurance as ARM-based tablets with similar battery capacities while running a full-fat desktop OS rather than a phone OS with delusions of competency.

Light-water, heavy-water and carbon moderated power reactors only breed U238 up into Pu239 and Pu240 by "accident", so to speak. They get a few percent of the total energy they produce from fissioning these products in-situ. Breeders meant to produce surplus fuel or "burn" waste require much higher fluxes, usually achieved in a small physical volume hence the higher temperatures involved and the use of sodium, lead/bismuth, helium etc. to conduct away the heat. The LFTR concept requires this high flux density, whether of moderated thermal neutrons or a mixture of fast and thermal neutrons while at the same time having the radiological problem of the high-temperature fuel being less constrained in liquid form.

Generally we've discovered that very high neutron fluxes (thermal, fast or a mixture of the two) in restricted volumes required for high levels of breeding in reactors and the attendant high temperatures tend to break things, cause leaks and fires and expensive shutdowns. At the same time reactors that work on the basis of moving fuel around (mostly pebble-bed designs) have not had a happy time of it even with lower neutron fluxes and larger working volumes in the core compared to out-and-out breeder designs. LFTR combines both of these iffy concepts.

Steam pipes leak all the time in light-water reactors, usually in the steam generators in the case of PWRs. This isn't a radiological problem as the cooling/moderating water isn't radioactive as it never comes in direct contact with the fuel and its waste isotopes which are ceramic pellets housed in sealed tubes. The steam loop runs at about 400 deg C or thereabouts at high pressure. In the case of LFTR and other breeder designs the coolant loop is at up to 700 deg C at which point most steel alloys have lost half their tensile strength compared to room temperature. Breeders that have broken their cooling loops in the past released molten sodium or helium but this had never came in contact with the fuel or its waste products so it was not particularly a radiological hazard. This is not the case with LFTR, of course.

"You could get the experimental platform for a couple of orders of magnitude less money."

No you couldn't, demonstrably. If they could build an ITER-scale reactor for one-hundredth the price they would have. Large-scale sustainable high-Q fusion is difficult. It cost billions to build and operate JET and it was never meant to beat breakeven (Q > 1) but it's come the closest to that of any of the major tokamaks with a couple of seconds of fusion with a Q of about 0.6 back in the 1990s. Heck JET wasn't even built specifically to do fusion, it was mainly supposed to be for plasma research but it got repurposed as plasma control and generation techniques improved. ITER, if it works as planned and the physicists haven't dropped a decimal point here or there, is a fusion reactor which will eventually run with Q >= 10 for several thousand seconds at a time. Maybe.

The "E" in ITER stands for Experimental. It's a testbed platform for trying out stuff and seeing how it breaks, a rig to make mistakes on and gain knowledge. There are nebulous plans to build DEMO and the later PROTO which will be power generating fusion reactors but they'll still be less than fully-commercial designs, just another step closer to the rollout of workable and cost-effective fusion power generation. Nothing is guaranteed though.

Like I said, nobody's ever run a thorium-cycle liquid-salt reactor and there is no Santa Claus. As for a "thorium breeder blanket" add-on to the Oak Ridge reactor, huh? The LFTR concept mixes thorium into the molten-salt stream, breeds it up to U-233 and then fissions it within a moderator to slow down the neutron flux. There is no separate blanket, it's all in one stream, salt, kickstarter fuel (U-233 or U-235/Pu-239), thorium and waste products all at 700 deg C and more, mindbogglingly radioactive, radiochemically very complex and being continuously moved around lots of piping and heat exchangers and chemical processing plant and it has to generate electricity at about 5 cents per kWh to be competitive.

Any such reactor is going to require a neutron flux way higher than the ORNL reactor ever experienced, a mix of fast neutrons to do the breeding and slower neutrons to fission the resulting U-233. This isn't a problem for existing well-tested light-water and heavy-water reactors delivering about 15% of the world's electricity demand right now, of course. In their case the ceramic fuel sits in zirconium tubes and water circulates around them to transfer heat and in some cases moderate the neutron flux, no fast neutrons specifically required for breeding purposes (although some breeding does happen anyway). Much simpler and more reliable, no explosives required.

I agree that uranium will not be scarce for decades, at least one conventional and proven light-water/heavy-water reactor operation cycle of about 60 years. It's possible it would never be scarce at all if the process to extract from seawater can be operated commercially -- it's been tested, its cost is estimated at about three or four times the price of conventionally mined uranium today. Some countries don't have much uranium within their boundaries so ongoing supply is not guaranteed. India is one such country hence their interest in developing a fuel cycle involving thorium for their heavy-water reactors. They're still building and operating conventionally-fueled reactors too though.

The ITER is designed to do more than "break even", it's expected to return 10 times the energy input for heating and controlling the plasma -- a return of 500MW for an input of 50MW and to sustain this for periods of thousands of seconds. This is just heat, not electricity, there's no plans to try and extract energy from the system yet. It's an experimental platform, not a prototype power generating system.

Whether ITER succeeds in this aim we won't know until it actually runs. One school of thought is that bigger tokamaks make it easier to control the plasma generated. Pessimists think more problems will crop up as the engineering scale increases. That's why they're building it, to find out.

No "they" didn't have a LFTR reactor working in the 70s. Nobody's EVER had an LFTR working. There is no liquid-fluorine thorium Santa Claus, just a lot of grad student Powerpoint presentations.

There was a molten-salt reactor, a laboratory-scale device fuelled with U-233 and later U-235 in intermittent operation at Oak Ridge National Laboratories for a few years in the 1960s. It never used thorium and wouldn't have been any good if it had because it couldn't breed thorium up into U-233 to fission for energy. It took a long time to decommission this small reactor in part as several bad things had happened to the piping inside it. Folks reckon the corrosion could have been fixed with a little tweak but you don't get to "tweak" sizeable reactors. Chernobyl 4 is a worked example of "tweaking" a large reactor.

China might sell you their CAP1400 light-water reactor design (an upgrade of the Westinghouse AP1000) or maybe their HTR-PM modular reactors; they're actually building one at the moment to test the concept and they have a small testbed gas-cooled pebble-bed reactor running at the moment. India is working on using thorium in regular heavy-water reactors as part of the fuel mix, not in molten-salt systems and nobody else is really interested in buying into what they're doing. Other folks are looking into pebble-bed reactors which can burn thorium as part of the fuel mix but the previous history of attempting this is not a success, mostly -- the Germans are still trying to figure out how to decommission their thorium-mix pebble-bed reactors. They've been filled with concrete for the moment to stop the leaks of radioactivity.

There are also experiments going on to see how thorium works in regular light-water reactors. The physics says it will work, it's not as energetic as regular uranium fuels though. Baby steps baby steps.

Ballistic coefficient varies with shape too but generally it's dependent on mass and size. A .223 bullet and a .50BMG bullet are similar in shape and muzzle velocity but the .50MBG goes a lot further because it loses speed less quickly than the smaller round. It still can't fly as far as a battleship gun round though even if it's a better shape.

As for drag, well we have worked examples of clean shapes flying at high speeds. The SR-71 flying at Mach 3 glowed a dull red from skin friction and that was at 80,000 feet where the air pressure is something like 0.5 lbs/square inch or 3% that at sea level. Drag and skin heating effect goes up as the square of the speed so the railgun dart at sea level and Mach 7 would experience something like 4 x 30 or more than a hundred times the amount of drag the SR-71 experienced, and the railgun dart doesn't have engines pushing it along and sustaining its velocity in flight.

BTW I was wrong about the late-model US battleship guns, the perils of working from memory rather than checking the numbers. Their muzzle velocity was similar to most rifles, about 2700 fps and they were of course 16" bore.

Modern large-calibre artillery shells usually have base-bleed which coverts them effectively into a full boat-tail configuration in flight by filling in the space at the base of the shell with hot gas. They're actually more efficient than the best rifle bullets in this regard.

I can't find this information on the web but has anyone actually fired a railgun projectile over the sorts of distances described in the goshwow articles and promotional bumpf designed to get more funding out of Congress? Has there actually been a 100-mile ballistic test of this system yet? 50-mile? 10?

The movies I've seen of railgun test firings have all been straight-line non-ballistic shots over a few dozen metres demonstrating the sort of armour penetration capabilities DU spears fired by 120mm smoothbores have been able to achieve for decades. I recall reading about folks experimenting with high-velocity wildcat rifles (.30 cal bullets in necked-down .50BMG cases and the like) who ran into problems with solid projectiles melting from air friction at muzzle velocities of only 4000fps (less than Mach 4) over a range of a few hundred metres. It's entirely possible a railgun round would vapourise if fired several kilometres through sea-level air at Mach 7. Not something, of course, a slower seaskimming Tomahawk missile has problems with even though it can fly seven times farther than this railgun can fire, can alter course, is terminally guided and can even carry EW jamming kit.