Tuesday, September 18, 2007

I blog to think

About 30% of the entries I write for the blog never get posted, because I cannot get my reasoning straightened out. There are some entries that I post, that I shouldn't have for the same reason.

I write these blog postings because when I try to figure out something complicated, it helps to write it down. When I started a company by myself (10x), I had to make progress with no coworkers for three years. There was no-one with whom to talk over complex ideas. I ended up writing essays to myself, each designed to take someone from the state I had been in before writing the essay (confused) to the state I was in after writing the essay (enlightened).

I've just discovered Paul Graham's site. He does a better job of saying the same thing here.

Thursday, September 06, 2007

One of the problems with a Thorium-fuelled Molten Salt Reactor is starting it. Plutonium (and U-238) from reprocessed reactor waste is the most obvious start charge, but the problem is that it takes several tons of Plutonium (and thus a ton or more of U-238 that comes with it) to start the reactor, which will burn just one ton of actinides a year (which will soon be mostly the U-233 introduced from the blanket), and so it will be decades before the plutonium level is low enough not to be a problem in the waste stream.

The core geometry that David LeBlanc suggests is quite simple -- one big Hastelloy tube in a big unpressurized vat of blanket salt with no graphite. We could make it more complicated by having a three-fluid reactor, with two fuel salt tubes in a single big blanket. The first fuel salt would contain the start charge of Plutonium/U-238. The second fuel salt would be gradually charged with U-233 recovered from the blanket. Fission product seperation would run on the second fuel salt but not the first. The idea is that the fission product buildup in the first fuel salt wouldn't ever rise to a level that would kill the reactor completely, and by immersing the start charge in the neutron flux for decades, you could eventually burn all the transuranics.

One other advantage of this arrangement is that if you had a reactor with a breeding ratio of, say, 0.95, you could insert a small amount (70 kg/year) of reactor-grade Plutonium into the first fuel salt loop to make up for the insufficient breeding. The fission products from these later additions would never add up to the same level as from the initial Plutonium charge, and so they would not poison the reactor either.

There is one other point I'd like to make about reactors with less than unity breeding ratio: the reactor is quite insensitive to the actual fissile load it carries. It would be quite reasonable to have a big start charge and subsequent make-up charges of Plutonium breed an extra 50% or even 100% more fissile than needed, so that the reactor could go for one or two decades without any further make-up. During those decades, the actinides in the first fuel loop can burn down to nothing. After the third decade of operation, while you are replacing the radiation-damaged tubes in the core, you can seperate the Uranium and salt from the two fuel loops, dump the remainder as short-lived waste, and restart the reactor with the Uranium it stopped with, plus another, smaller start charge of Plutonium.

All this excess fissile material is a proliferation hazard in foreign countries. But the worst energy problem in the world is in the United States, where proliferation is not a problem -- we already have the Bomb. We do have to worry about diversion, but I frankly think that's a pretty small problem compared to the national security problem we face due to importation of oil.