5 ms·

This was described in depth in a recent Nova program called "The Nuclear Option"[1][2]. In it they referenced multiple other more safe methods than the most co

• acidburnNSA 10y ago
  I watched it too and can elaborate a bit on this since I'm in the SFR business (anyone want one?). When an nucleus fissions, not all the heat comes out at once. Roughly 7% of the energy comes out later, distributed in time as a decaying exponential [1]. The really key safety challenge of nuclear is cooling that after the chain reaction is stopped (because 7% of 3 billion Watts is 210,000,000 Watts). So if the power goes out (like in Fukushima), traditional plants rely on active cooling systems like diesel generators, fuel cells, steam turbines, etc. to run pumps that cool the fuel, preventing the fission products from emerging. There are some nuclear reactors that can handle the decay heat without active cooling systems. They're mostly low-pressure/exotic coolant systems, including liquid sodium metal, molten salt (FLiBe, NaCl, etc.), molten lead, etc. These can just naturally circulate and dump heat thorough ambient heat exchangers outside. And some gas-cooled Pebble Bed reactors can do it too because their fuel can get hot enough to just conduct it out. Worldwide, we have by far the most experience with sodium-cooled fast reactors (400 reactor-years). As pointed out in NOVA, two weeks before Chernobyl, the small sodium-cooled EBR-II in Idaho demonstrated station blackout conditions without scram and it just shut itself down and cooled itself. But sodium metal has a problem. It is quite reactive with air and water, so dealing with it can be an operational challenge. We know how to deal with sodium leaks in sodium-water steam generators (arguably the most "exciting" component in a SFR) and sodium fires, but they can still be expensive. The French SuperPhenix SFR suffered a series of political and weather-related challenges that ended up giving it a terrible operational record. The Japanese SFR Monju has a bad history too with some 10 year outages and whatnot. But EBR-II and FFTF in the US operated fantastically until Bill Clinton finished shutting down their funding following the long slowdown of nuclear research that came after the failed and super-expensive Clinch River Breeder Reactor Project, and the Russians now have the best and only commercial SFRs. So there's still hope for SFRs as Gen IV nukes. The French are working on a huge program called ASTRID to make a better SFR. The Koreans have KALIMER/PGSFR that's very far along in design. The US has TerraPower's Traveling Wave SFR, the Indians are turning a big one on now, China is operating a sodium-cooled test reactor (CEFR). The Russians are building another awesome SFR test reactor (MBIR, to replace BOR-60) and continue to operate and sell their BN-600/800, etc. series power plants. [1] https://whatisnuclear.com/physics/decay_heat.html https://whatisnuclear.com/physics/decay_heat.html
  • tbihl 10y ago
    So, at the risk of sounding foolish: what's the advantage of a fast reactor? If I'm remembering correctly, U-235 has a much better thermal than fast fission cross section. As to the rest, I'm sure there are risks to pressurizers and natural circulation, but they seem a lot more comfortable than trying to avoid sodium leaks. Or is there so little corrosion in sodium systems that it works out pretty well?
    • xjwm 10y ago
      I'm not an expert, but IIRC a fast neutron reactor can change power levels very rapidly. The rate of change for a thermal reactor needs to be much lower.
    • acidburnNSA 10y ago
      Not foolish. It's really quite non-intuitive. I'll try to break it down. Physical realities of note: 1. The only fissile nuclide (ie one that readily splits) that existed on Earth in 1938 is the minority uranium isotope, U-235, at a concentration of 0.7% compared to U-238. More exist now (such as Pu239, U233), but we had to synthesize them. 2. U-235 is 1000s of times more likely to split if it is hit with a slow (aka "thermal") neutron than a fast one. 3. If the (very plentiful) U-238 nuclide absorbs a neutron, it converts a few neutrons into protons until it becomes Plutonium-239, now a fissile nuclide(!). Same can be said about Thorium-232 and fissile U-233. 4. Every fissile nuclide releases substantially more secondary neutrons per neutron absorbed if hit with a fast neutron instead of a slow one. So the implications of these is as follows. To get a critical chain reaction working in the first place, Facts 1 and 2 necessitated the slowing-down of neutrons in an arrangement of natural (unenriched) uranium. This was originally done with very pure graphite in Chicago in a squash court by Enrico Fermi and co. in December, 1942. This led to natural-uranium fueled, thermal neutron plutonium-production reactors in Washington state using Fact 3 for the Manhattan Project. These reactors essentially converted diluted fissile material (natural uranium) into concentrated fissile material (chemically-separable Plutonium) for use in nuclear bombs. With Plutonium and enriched uranium available in the 1950s, it was now possible to start a chain reaction without slowing the neutrons down (getting around Fact 2). It was thought that Uranium was exceedingly scarce worldwide (turned out to be not entirely true), so rapidly converting lots of U-238 into Pu-239 fuel was thought essential to scale nuclear as a power source. To do this (without burning quickly through all the world's U-235), you need lots and lots of excess neutrons. Enter Fact 4. Facts 3 and 4 led to the development of fast breeder reactors, which can produce world-scale clean energy for thousands upon thousands of years using known resources. So really the sustainability of breeding is the key capability of fast reactors, and the roughly 100x improvement in safety (measured by core damage frequency) is a bonus. You can also burn nuclear waste if you do a fully-closed fuel cycle (because fast neutrons can split even non-fissile nuclides like Pu240, Np237, Am241), emitting only fission products that decay to stability in hundreds of years instead of hundreds of thousands of years.