7 ms·
And the engineering here is very problematic. Indeed, the issue mentioned may be a showstopper for DT fusion. The problem is limits on power/area through the
by pfdietz 4y ago
And the engineering here is very problematic. Indeed, the issue mentioned may be a showstopper for DT fusion.
The problem is limits on power/area through the wall of the reactor. Because all the produced energy has to go through the wall, and because the area of the wall grows as r^2, the volumetric power density of a DT reactor (that doesn't exceed the power/area limit on the wall) must decline with increasing size.
This leads to DT fusion reactors having horrible volumetric power density. ITER, for example, has a volumetric gross fusion power density of 0.05 MW/m^3. The 2014 ARC design, 0.5 MW/m^3. A PWR's reactor vessel? 20 MW/m^3.
(If you look at the power density of the plasma alone, or the fission reactor core alone, you get similar wildly discrepant numbers: https://cpb-us-w2.wpmucdn.com/research.seas.ucla.edu/dist/d/39/files/2022/06/Final-FINAL-CIMTEC-2022-copy-Perugia-Italy-6-29-2022.pdf https://cpb-us-w2.wpmucdn.com/research.seas.ucla.edu/dist/d/... see slide 26)
If you want to make heat to make steam to drive a turbine, a fission reactor will be much smaller, and much cheaper, than a DT fusion reactor.
- IX-103 4y agoWhat's wrong with pulling the energy from the magnetic containment system? We already are using the magnetic containment system to redirect charged particles in the plasma back into the plasma. If we do so less forcefully than they were heading out, but still enough to stop them from leaving, we can convert particle velocity directly into electricity (AC with a relatively linearly decreasing frequency distribution, I think). This also has the effect of cooling the plasma (since temperature is related to average particle velocity). The challenges I see here are: 1) The faster the control system, the more efficient it is at extracting energy -- it can extract the energy of higher frequency components. 2) You may need a very large magnetic field (larger than needed for containment) to make this practical 3) Coil geometry and efficiency becomes critical.
- pfdietz 4y ago80% of the fusion energy from DT fusion comes out in fast neutrons, which ignore magnetic fields. The other 20% has to come out somewhere. If it could be converted to electrical energy maybe it could avoid surfaces (Helion plans to do this, with DD + D3He) but when it's just 20% of the output that doesn't buy you much.
- vintermann 4y agoI knew neutrons was a headache with nuclear energy in general, but I didn't know it was that bad. I would not be excited about attacking the problem of how to safely get energy out of a soup of hot, corrosive, exotic radionuclides.
- pfdietz 4y agoAnd DT fusion neutrons are even worse than fission neutrons, since they are much more energetic (and because only 3% of the energy in a fission reactor is released in neutrons.) The 14 MeV DT neutrons are above the threshold for (n,p) and (n,alpha) reactions, which cause gas to accumulate in reactor materials. The helium produced by (n,alpha) reactions accumulates in tiny bubbles that would reach extreme pressures and rip materials apart.
- jjk166 4y agoAnd neutrons pass straight through solid matter, dumping their energy volumetrically, completely solving the surface area scaling problem.
- pfdietz 4y agoWell, no, not completely solving it. There are limits to neutron fluence materials will withstand. Even neutron heating can't be too nonuniform in the blanket.
- trhway 4y agoYes, the Sun has volumetric intensity of a compost pile.
- adrianN 4y agoFor entirely different reasons though.
- trhway 4y agoI see there the same relationship. If you have a fixed surface temperature then you have to decrease volumetric intensity when increasing volume (the ITER and further case). If you have a fixed volumetric intensity then increasing volume results in higher energy flux per unit of surface and thus results in surface temperature increase to increase the intensity (T^4) of radiation away of that energy - that is the Sun case.
- ricksunny 4y agoI like this term ‘volumetric intensity’. I don’t know how common the term is, (some cursory googling just turned up one fusion paper from the 1990’s)) but it would make appreciating Gauss’s Law in E+M much easier from the outset than the typical surface integral calculus typically used to introduce the subject.
- api 4y agoThe sun's core where most fusion happens would be the appropriate comparison. The rest of it is a boiling blob heated by the core.
- diordiderot 4y agoCan you explain this like I'm a 5 year old?
- jjk166 4y agoThe overwhelming majority of the sun is made up of hydrogen, and specifically the isotope protium, which is just a proton with no neutrons. When two protons collide, the overwhelming majority of the time they just bounce off eachother. The diproton (a particle with two protons and no neutrons) isn't just unstable, it's basically forbidden. However there is an incredibly rare chance that when two protons come together instead of just splitting up, one of the protons will turn into a neutron, thus producing deuterium, the hydrogen isotope with one proton and one neutron. Deuterium is much more capable of fusing. Now when I say incredibly rare, I mean like "unlikely to happen to any given hydrogen atom over the lifetime of a star" rare. Luckily stars are big, and dense, and thus contain a lot of hydrogen atoms hitting each other very frequently. Thus you have a slow but consistent burn of hydrogen over billions of years. Once it's converted to deuterium, it fuses within about 1 second. This produces helium-3 (2 protons, 1 neutron), which is also capable of fusion, but not with the abundant protium[1], so it takes some time to find something to fuse with. A helium-3 will on average survive about 200 years before it fuses, which is still pretty short as far as the sun is concerned. This slow burn means that within any given chunk of the sun there aren't that many fusion events occurring. Even in the deepest core of the sun, power densities never exceed 275 Watts per cubic meter. By comparison, compost releases about 7000 Watts per cubic meter at its peak decay rate and even over longer time periods averages around 600 Watts per cubic meter. Human metabolism releases about 1600 Watts per cubic meter. But since the sun is so immensely large, it has a lot of cubic meters of fusing hydrogen. Overall, it's producing 3.8x10^26 Watts. [1] Actually helium-3 can (at least in theory) fuse with a proton, it's just extremely rare. In fact it is so rare we've never actually observed it.
- jiggawatts 4y agoAn idle thought I had related to this is that the real trick to scaling up fusion power may be to not try and extract the energy "through the walls", as you said. You know how they tell you to "visualise success" because you need to know what it looks like to aim for it? Along those lines I was visualising what it would look like if someone like Elon Musk was on YouTube showing off some sort of future fusion reactor that's "not just a piece of lab equipment", but something that would be dramatically better than any extant fission reactor or similar technology. I visualised it as an enormous stationary rocket engine, where cooling water was pumped through a relatively narrow (~50cm) "reactor tube". The water needs leave a hole in the middle for the fusion fuel gases. The hole is achieved by imparting a rotation to the water, so it "spins out" to the sides. Fusion would occur not throughout a large volume, but at a small number of "pinch points" along the axis set up using powerful magnets -- and not necessarily superconducting magnets! To get sufficient current through the coils, the charged exhaust is allowed to expand through a magnetic "rocket engine bell", producing moving (and accelerating) electric charges. This current is recirculated through the engine to provide the enormous magnetic field strengths required, via thick copper conductors aggressively cooled with water. Essentially it would be a magneto-hydrodynamic-fusion jet engine, using fields instead of impellers to achieve compression, expansion, and energy recovery to run the whole cycle. The bulk of the energy would be in the super-heated steam produced, which could be used as in traditional power plants, or used as a bone-fide rocket in space. Obviously there would be "challenges" to making this work, to put it mildly! But consider that with the kind of computer power we can throw at modelling physics these days, it may be possible to use "topology optimisation" style tricks to figure out the required geometry of the parts to achieve the conditions required for fusion while staying within material design constraints. If you ask me, a Manhattan-project style attempt at something like this would be a better way to "waste money" than yet another aircraft carrier, or whatever...
- randmeerkat 4y ago> “You know how they tell you to "visualise success" because you need to know what it looks like to aim for it? Along those lines I was visualising what it would look like if someone like Elon Musk was on YouTube showing off some sort of future fusion reactor that's "not just a piece of lab equipment"” The money went to his head, he’s not the innovator he was, now he’s just a billionaire that got bored with Mars and wants to own social media instead.
- cma 4y agoCan the wall be baffled like an air filter to increase area? I get that the flux through it in total at a given radius would be the same, but is the heat being pulled out through plumbing or something?
- dotancohen 4y agoThat only works to increase the surface area in materials with poor heat conduction, such as an aluminium-air interface on an air-cooled motor (air is a poor conductor of heat). Poor heat conduction is not the problem here.
- pfdietz 4y agoActually, heat conduction is something of a problem here in at least one way. Some of the energy going to the first wall impinges on the surface of the first wall and becomes heat there (photons from the plasma, direct impact of charged particles from the plasma on the surface). This must be conducted through a vacuum-tight barrier layer before reaching a coolant channel. The ability of this layer to dissipate heat (before stress from differential thermal expansion of the inner hot/outer cold parts cause it to fail) is proportional to its thermal conductivity. For ITER, the relevant layer of the first wall armor is a CuCrZr alloy. However, this alloy is unsuitable for use in a production DT reactor due to activation. For DEMO, this layer is going to have to be made of RAFM steel, which has an order of magnitude lower thermal conductivity. This problem MIGHT be solvable with a liquid first wall (flowing liquid lithium), but it's not clear all the penetrations of a fusion reactor can be shielded in that way -- and any that are not will face this issue. And the lithium doesn't solve the neutron wall limit problem unless its very thick.
- ncmncm 4y agoAs you have noted elsewhere, molten lithium would interfere with magnetic containment. Lithium hydride melts at a rather higher temperature. I don't know how reactive molten LiH would be with pipes. I know letting any air into contact with molten LiH would be, as we say, bad.
- melony 4y agoBack to stellarators? Mobius strips are very beautiful.
- qikInNdOutReply 4y agoStellarators do not solve the problem, they would heat evenly too. (TRIZ drums) What is needed is some sort of "abblative" migrating system, that leaves behind its heat to harvest, while the fusion process travels onwards, stressing other rejuvinated parts of the containement vessel. Basically, a series of reactors, pumping the plasma in circles once ignited. Or we dump the whole idea, call it a day and pump all the money into musks larange solar power plant. Which ironically is a fusion containement in the distance power scheme. Which then will have problem with solar winds. Funny how the small plant mirrors the problems of the larger soon to be plant.
- Mvandenbergh 4y agoOn the other hand, basically every complicated bit of engineering on the PWR is due to the fact that it does have such a high volumetric power density. A PWR is basically a fancy (very fancy) pressure vessel which is cheap (in relative terms) and then parallel redundant cooling systems and containment buildings because dealing with such high power density is hard. Right after shutdown, the power density from decay heat is still 1.3MW/m3 and dealing with that decay heat in all currently conceived circumstances is where all the money goes.
- concordDance 4y agoI'm confused about how 50kw/m^3 is "horrible". A 4m a side sphere would give 1.5mw, which is a decent sized plant. What's wrong with a 3-6m size reactor vessel?
- pfdietz 4y agoThe cost of the reactor vessel is directly related to how large it is (and how complex). The larger object uses more material, has more parts, has more joints between parts. Its reliability decreases along with its size, increasing cost even more (as all those parts and joints must now themselves have lower failure rates to maintain a given system failure rate.)
- throwaway0a5e 4y agoThis is an incredibly naive bordering on "probably just dishonesty masquerading as ignorance" thing to say. Of course cost goes up fairly linearly with size. More rebar, more pipes, more materials, more diesel in the bulldozer, more man hours, more welding rods, more welds to inspect, more money. But there are also a ton of one per instance overhead paper pushing costs, especially with something as onerously regulated as nuclear that are a large fraction of overall costs and will be roughly the same whether you're building a 1kw facility or a 1gw facility. All of that has to be amortized over your reactor. So if you build a small reactor each dollar of reactor might have $10 of overhead it needs to amortize. If you build a big one each dollar of reactor might only have another $1 it needs to amortize. Of course there's a sweet spot before the reactor is so massive the part count grows so high that the MBTF starts getting you. That sweet spot also has a low end below which the fixed regulatory and palm greasing costs are so high that your tiny reactor can never hope to pay them back in it's lifetime.
- pfdietz 4y agoThe argument being made doesn't require a linear relationship between size and cost. Remember, we're looking at two objects with the SAME power output, one much larger than the other. The MTBF/MTTR of fusion is actually looking like an extremely serious problem. Fission reactors can operate with 2% of their fuel rods leaking. A single leak through the vacuum boundary of a fusion reactor will contaminate the plasma and shut things down. And repairs on a fusion reactor are going to be much more difficult than simply replacing fuel rods. The innards of any fusion reactor will be so radioactive that hands-on maintenance will be impossible.
- go_elmo 4y agoVery wild idea: gaining electricity by induction instead heat since its charged matter.. but I know it doesnt make sense, as it only yields the kinetic energy which has too high entropy to be yieldable (no directed movement)..
- NegativeLatency 4y agoA lot of the particles emitted are neutrons (which have no charge)
- wiz21c 4y agoEuh ? In a few paragraph you basically say that even if fusion works, getting energy out of it will be almost impossible in a meaningful way (you're showing order of magnitudes of difference). But, don't the engineers at ITER and other facilities have already though about that ? It'd be very irresponsible to start such a project knowing in the beginning that, at the end, there's a high chance of failure...
- pfdietz 4y agoThat's a good question. Why do fusion people not pay attention to this issue? The answer is some do. And those people tend to get frustrated and quit the field (or, if they are lucky, notice the problem early and never go into fusion at all.) Lidsky is a famous example; Pfirsch and Schmitter in Europe are others. Others bottle up their objections and only let them out as they retire. The more common coping mechanism is just to assume someone else is going to solve it and not pay too much attention while working on other issues. It helps to focus on physics issues -- which are very interesting, after all -- and tell yourself that this problem is "just engineering" and can't be too hard to solve, in comparison, and that anyway telling yourself it's bad to think about it before the physics is nailed down (which is wrong, but lets you stop worrying that you've wasted your career.) As for ITER specifically: ITER is costing > $20B for something that could produce maybe 400 MW gross fusion power (and isn't engineered to produce tritium, electrical power, or to survive more that a few weeks at full power). Industrial levels of denial are needed to tell oneself something like that is on a trajectory to relevance. Perhaps an even better question is: why are people funding DT fusion? For governments, I think it's because in many cases the purpose of government research funding isn't to solve problems, it's to be seen working toward solutions in the distant future (and that perception can be achieved even if there is no actual chance of success, if the public is sufficiently unaware, which it is.) For venture capital I'm not clear what the explanation is. Spinoff technologies like high Tc superconducting magnets, maybe?
- Zoadian 4y agocompare it to the space program. that didn't really have any directly useable goal. but the technology produced in the process is used everywhere. DT fusion may or may not work. maybe the problems may get solved some day, maybe not. we'll never know if we don't try. and the technology we get in the process is valueable regardless.
- xani_ 4y agoCould the part of the plasma be diverted into heat exchanger and back into the stream somehow ? I guess problem there would be having one that doesn't just melt away. Maybe shoving water into it so it flash-evaporate and we're back to the ye olde steam?
- pfdietz 4y agoDoesn't help much in DT fusion, since 80% of the produced energy immediately leaves the plasma in neutrons. For non-DT fusion, there are alternatives, which is why I was explicitly saying "DT" in the critical comments. This is also why I consider Helion's the least dubious fusion effort out there today.
- dahfizz 4y agoTaking your number of 0.05 MW/m^3, Back of the napkin calculation is that the USA would need ~8,000 gigafacory-sized reactors to meet 100% of electricity demand.
- howenterprisey 4y agoDoes volumetric power density cause any issues besides making the devices relatively large (and thus more expensive)?