Nuclear fusion!

[PaulineLovesPhysics]

I KNOW the difference between fission and fusion. What I want to know is: is nuclear fusion better than fission? Is it cleaner? Why should we invest in fusion?
I WOULD LOVE SOME LINKS TO WEBSITES ON THIS!
Thanks

[gmalivuk]

It's expected to be cleaner, yes, but different types of fission reactors could in principle be very clean, as well.

The main advantages are greater mass efficiency (a smaller mass of fuel can give a larger energy output) and more abundant fuel (hydrogen instead of uranium).

[gorcee]

In addition, fusion reactors cannot be used to generate weapons-grade materials. There is no risk of a catastrophic event happening in a fusion reactor -- no chance of meltdown. There are very few radioactive byproducts; although the reactor housing becomes radioactive over time, it would, in principle, take a very long time for this to happen, whereas fission fuel is often very radioactive (different reactors use different fuels, so this is not a general statement, but rather reflects the current operational state -- new reactor designs can be cleaner, as gmalivuk said, but they are not in the majority of currently operational designs).

In addition to being abundant, fusion would be inexpensive to deploy. You could build a reactor in a third world country, operate it very cheaply, and provide power to the citizens without any fear that the local government would attempt to use it to build a bomb. Reactors would also be very small, rather than multi-thousand acre facilities requiring dams to create artificial reservoirs, facilities could be built on land the size of, say, a local high school. However, you still might require a nearby water source for efficient steam generation.

[thoughtfully]

In addition to being abundant, fusion would be inexpensive to deploy. You could build a reactor in a third world country, operate it very cheaply, and provide power to the citizens without any fear that the local government would attempt to use it to build a bomb. Reactors would also be very small, rather than multi-thousand acre facilities requiring dams to create artificial reservoirs, facilities could be built on land the size of, say, a local high school. However, you still might require a nearby water source for efficient steam generation.

Citation please? Every indication is that small tokamaks don't work. The ITER reactor is going to cost on the order of hundreds of billions of dollars and require highly trained staff. Inertial fusion isn't much better. Something like polywell might be cheap, but it's far from clear that it'll ever work. Let's also keep in mind that the cost of the power that comes out hasn't got much to do with the cost of the fuel. "Too cheap to meter" is going to apply to fusion about as well as it worked out for fission, unless something unexpected comes along. Even for Uranium, it's still cheaper to throw out 90% of it and dig more out of the ground than it is to reprocess the fuel.

Something like a molten salt Thorium reactor would work better in the near term, and not have any proliferation concerns. China and India are building them, why aren't we?

some links:

National Ignition Facility (inertial confinement)
http://www.strongforce.org/

search results for topics containing the word "fusion" in the Science and Serious Business forums (and any subfora)

[gorcee]

In addition to being abundant, fusion would be inexpensive to deploy. You could build a reactor in a third world country, operate it very cheaply, and provide power to the citizens without any fear that the local government would attempt to use it to build a bomb. Reactors would also be very small, rather than multi-thousand acre facilities requiring dams to create artificial reservoirs, facilities could be built on land the size of, say, a local high school. However, you still might require a nearby water source for efficient steam generation.

Citation please? Every indication is that small tokamaks don't work. The ITER reactor is going to cost on the order of hundreds of billions of dollars and require highly trained staff. Inertial fusion isn't much better. Something like polywell might be cheap, but it's far from clear that it'll ever work. Let's also keep in mind that the cost of the power that comes out hasn't got much to do with the cost of the fuel. "Too cheap to meter" is going to apply to fusion about as well as it worked out for fission, unless something unexpected comes along. Even for Uranium, it's still cheaper to throw out 90% of it and dig more out of the ground than it is to reprocess the fuel.

Something like a molten salt Thorium reactor would work better in the near term, and not have any proliferation concerns. China and India are building them, why aren't we?

So I'm talking about the concept of a hypothetical commercial reactor, and you're talking about present research projects.

The OP is asking about what the advantages of fusion might be. I answered his question.

As for why we're not building molten salt reactors in the US, the answer is because there has been essentially a moratorium on new fission reactor construction, due to heavy political pressure from environmental lobbyists.

[firechicago]

It's worth noting that all of the above mentioned benefits and more were claimed for fission power. In the 50's and 60's the expectation was that electricity from fission plants would be literally "too cheap to meter." And no one started thinking seriously about the problems of nuclear waste until after the first plants were up and running.

But the main problem is that no one expected the things to be so damn expnsive to build. And given that the technical challenges of creating a sustained fusion reaction are, in many ways much greater than those involved in creating a sustained fission chain reaction, I find it highly unlikely that fusion reactors will be "inexpensive to deploy." Sure, they might be dirt cheap to run, but that doesn't help much if they cost you $20,000 per kWE to build.

[Zamfir]

It's worth noting that all of the above mentioned benefits and more were claimed for fission power. In the 50's and 60's the expectation was that electricity from fission plants would be literally "too cheap to meter." And no one started thinking seriously about the problems of nuclear waste until after the first plants were up and running.

Ironically, the "too cheap to meter" statement might actually have been referring to fusion energy. The guy making the statement (the head of the Atomic energy commission) was well aware of early 1950s US research into both hydrogen bombs and fusion energy, but both were kept rather classified to the general public at the time.

At that time the first viable fission plants were just around the corner, and their development had been expensive but basically unproblematic. So the expectation of some was that fusion plants would be the obvious next step, up and running within your lifetime. Overoptimistic, but perhaps not crazy at the time.

[gorcee]

It's worth noting that all of the above mentioned benefits and more were claimed for fission power. In the 50's and 60's the expectation was that electricity from fission plants would be literally "too cheap to meter." And no one started thinking seriously about the problems of nuclear waste until after the first plants were up and running.

But the main problem is that no one expected the things to be so damn expnsive to build. And given that the technical challenges of creating a sustained fusion reaction are, in many ways much greater than those involved in creating a sustained fission chain reaction, I find it highly unlikely that fusion reactors will be "inexpensive to deploy." Sure, they might be dirt cheap to run, but that doesn't help much if they cost you $20,000 per kWE to build.

But they're not that expensive to build. They're expensive because they're research projects, and rather than a bunch of certified engineers managing them, they have scores of highly-trained PhD-level research scientists working on them. These are also government-funded research projects, which means that there is no expectation for the government to recover research investment dollars once the technology becomes commercializable.

It's virtually certain that the cost of an effective fusion reactor is going to be much cheaper than an equivalent fission reactor. The lack of catastrophic risk means that far less contingency infrastructure must be put in place. No complex fuel rod storage depots with managed cooling systems. No multiply-redundant generators. No need for 0-fault earthquake protection.

All you need are magnets, a turbine, a building, and really good computers. And considering that computing power is dirt cheap, and turbines and buildings are not new tech, a commercial reactor would probably cost on the order of a mid-sized coal plant -- if not less.

[Zamfir]

All you need are magnets, a turbine, a building, and really good computers. And considering that computing power is dirt cheap, and turbines and buildings are not new tech, a commercial reactor would probably cost on the order of a mid-sized coal plant -- if not less.

By that logic, the LHC should be really cheap - not even a turbine! Or aircraft - just a load of aluminum, and some jet engines. Jet engines are hardly new tech. Oil refineries -> bunch of pipes and pumps, some controller circuits, usually left outside so no building necessary. The Three Gorges dam -> concrete and steel, plus some FEM calculations. Basically 19th century tech scaled up. And Chinese construction workers earn even less than PhDs research scientists (though the difference is shrinking). Rockets -> aluminum pipe filled with fuel, plus a 386 if you want to calculate a reentry trajectory. Ever seen a semiconductor factory? All you need is a building with machines inside.

Really, you start wondering where the money goes. Overpaid contractors, probably.

[cphite]

All you need are magnets, a turbine, a building, and really good computers. And considering that computing power is dirt cheap, and turbines and buildings are not new tech, a commercial reactor would probably cost on the order of a mid-sized coal plant -- if not less.

Heck, I have magnets and computers at home... and my house is a building! All I need is a turbine and our energy crisis is solved!

No idea what those scientists are doing when it's all so simple...

[SU3SU2U1]

They're expensive because they're research projects, and rather than a bunch of certified engineers managing them, they have scores of highly-trained PhD-level research scientists working on them

Fission plants aren't all research projects and they aren't cheap to build, even the ones managed by certified engineers. Also, per man-hour you almost certainly pay engineers more than phd research scientists.

Also, to the best of my knowledge, no one has figured out a good first-wall material for a fusion reactor, so you should add that to list of things needed- magnets, turbine, building, materials-that-don't-exist-yet...

[gorcee]

All you need are magnets, a turbine, a building, and really good computers. And considering that computing power is dirt cheap, and turbines and buildings are not new tech, a commercial reactor would probably cost on the order of a mid-sized coal plant -- if not less.

By that logic, the LHC should be really cheap - not even a turbine! Or aircraft - just a load of aluminum, and some jet engines. Jet engines are hardly new tech. Oil refineries -> bunch of pipes and pumps, some controller circuits, usually left outside so no building necessary. The Three Gorges dam -> concrete and steel, plus some FEM calculations. Basically 19th century tech scaled up. And Chinese construction workers earn even less than PhDs research scientists (though the difference is shrinking). Rockets -> aluminum pipe filled with fuel, plus a 386 if you want to calculate a reentry trajectory. Ever seen a semiconductor factory? All you need is a building with machines inside.

Really, you start wondering where the money goes. Overpaid contractors, probably.

Are you all daft, or have you missed the part where I'm talking about hypothetical commercial reactors. I'm not talking about the cost of research and development. Because those costs are not paid in arrears, and they're being covered by federal funding, with no expectation of a return on investment.

So, my Nissan Juke cost $28,000. The prototype Juke probably cost n*$1,000,000 to make, where n > 1. Why is my car cheaper? Because when I bought the car I didn't also buy the engineer who designed the suspension.

[gorcee]

They're expensive because they're research projects, and rather than a bunch of certified engineers managing them, they have scores of highly-trained PhD-level research scientists working on them

Fission plants aren't all research projects and they aren't cheap to build, even the ones managed by certified engineers. Also, per man-hour you almost certainly pay engineers more than phd research scientists.

Also, to the best of my knowledge, no one has figured out a good first-wall material for a fusion reactor, so you should add that to list of things needed- magnets, turbine, building, materials-that-don't-exist-yet...

Right, because fission plants require lots of things. Like pools to store spent fuel rods, cooling systems for those pools, control systems for those cooling systems, redundant power supplies, lots of land near water (sometimes even dams built to supply that water), large concrete structures to shield the reactor core, etc.

The actual reactor unit itself is probably not that expensive. It's the infrastructure to support it.

Fusion plants would be similar, except the actual reactor unit would be a larger percent of the cost of the facility. However, once you had a commercial design, you have already done all the R&D regarding that design, so you just need to build it and ship it.

Once the first 747 was certified to fly, all they needed to do was build them and ship them. Yeah, they're expensive. But the 747s that, say, ANA buys are cheaper than the prototypes. And, the cost of those jets DOES include recovery of R&D costs to develop the plane. So, it's an unfair comparison to say, "ITER, a multinational research project costs X. Therefore, future fusion reactors will be expensive." No, ITER is a multinational research project intended to unlock science to enable commercial designs. It's not intended to create a design worth duplicating and selling. And part the cost of the project is to hire a bunch of people to do math that hasn't yet been done. A commercial design would take this science, develop a reactor somewhat based on it, use the math that's been done, and then create a product that can be sold. Even with some extra R&D costs needed, that R&D will apply to tens, dozens, or maybe hundreds of reactors, not just one. So those R&D costs will be distributed over those N reactors, and so the costs of building a hypothetical commercial reactor are *far less* than the costs of building a research design.

[starslayer]

And considering that computing power is dirt cheap, and turbines and buildings are not new tech, a commercial reactor would probably cost on the order of a mid-sized coal plant -- if not less.

It's this sentence that makes you seem to be vastly underestimating the difficulties and costs in building a fusion reactor. Fission reactors really aren't new tech either, when you think about it; in fact, they require almost exactly the same things a fusion reactor does, since all they really are is a (radioactive) heat source surrounded by cooling equipment that runs a turbine. Bam! Simple! Oh wait, they cost a couple billion dollars or so (IIRC); whoops. Guess that cooling stuff is all pretty expensive to maintain and operate after all. A large part of that, of course, is that the heat doesn't shut off right when you tell it to, but another part is that this is an extremely concentrated, high-intensity, high-temperature heat source. Fusion has this problem, but even more so, since instead of ~1000-1500 K, you're dealing with tens of millions of K or more. Not to mention the pressure needed to get the stuff to fuse.

Also, fusion reactors generate radioactive waste too, since the fusion reactions within our reach all emit neutrons. That has to be disposed of somehow, likely using some of the same techniques as for fission reactors.

[firechicago]

The actual reactor unit itself is probably not that expensive. It's the infrastructure to support it.

That's ridiculous. Reactors are hugely expensive pieces of machinery that need to be minutely controlled remotely at extreme temperatures and pressures while encased in massive steel and concrete containment vessels. Keeping a bunch of waste in a cooling pool and circulating the water is a trivial problem compared to keeping a reactor running. I'm not finding any good sources online, and I can't raise any of the experts I know at the moment, but I would be shocked to learn that the reactor and its associated cooling, control and containment systems were less that 75% of the cost of a nuclear power plant.

More generally we get that you're talking about a hypothetical commercialized reactor which doesn't have any of the problems that current research reactors have run into. But assuming that all those problems can be solved in ways that are cheap to implement and easily scalable is far from a safe assumption. Maybe they will find such solutions, and maybe the Higgs mechanism will create free brownies, but I wouldn't put money on either proposition.

[gorcee]

And considering that computing power is dirt cheap, and turbines and buildings are not new tech, a commercial reactor would probably cost on the order of a mid-sized coal plant -- if not less.

It's this sentence that makes you seem to be vastly underestimating the difficulties and costs in building a fusion reactor. Fission reactors really aren't new tech either, when you think about it; in fact, they require almost exactly the same things a fusion reactor does, since all they really are is a (radioactive) heat source surrounded by cooling equipment that runs a turbine. Bam! Simple! Oh wait, they cost a couple billion dollars or so (IIRC); whoops. Guess that cooling stuff is all pretty expensive to maintain and operate after all. A large part of that, of course, is that the heat doesn't shut off right when you tell it to, but another part is that this is an extremely concentrated, high-intensity, high-temperature heat source. Fusion has this problem, but even more so, since instead of ~1000-1500 K, you're dealing with tens of millions of K or more. Not to mention the pressure needed to get the stuff to fuse.

Also, fusion reactors generate radioactive waste too, since the fusion reactions within our reach all emit neutrons. That has to be disposed of somehow, likely using some of the same techniques as for fission reactors.

Considering that I actually worked on the NSTX, I'm fairly certain that I understand quite a bit more about fusion reactors than you think.

[starslayer]

Okay then, what about your experience on that project made you think that fusion reactors will be rather cheap as power plants go? What gives you the optimism that cheap (or even cheap-ish) solutions will eventually be found for the problems the research reactors currently face?

[gorcee]

Okay then, what about your experience on that project made you think that fusion reactors will be rather cheap as power plants go? What gives you the optimism that cheap (or even cheap-ish) solutions will eventually be found for the problems the research reactors currently face?

1.) NSTX is actually designed for > breakeven. One of major challenges there was beta loss (magnetic field pressure to plasma pressure) due to tearing mode instabilities. The design of the NSTX had much higher magnetic field pressure, so the computational demands for how to manage these problems were much greater. Improvements in computation capability have allowed for better feedback control of these instabilities. Conceptually, the problem is not dissimilar to controlling an aircraft unstable in some axis or axes. It's just harder to accomplish, but there is quite a lot of room for innovation in the computational abilities in this domain.

2.) Compact stellarator designs have, IIRC, never been observed to encounter a beta-decaying instability. Unfortunately, the NCSX project was cancelled. A substantial part of the cost to build NCSX was that it would not have been built in the same facility as NSTX, which was also the same facility as TFTR. So, a whole new infrastructure was requested so that NSTX and NCSX could have been run concurrently.

3.) Fusion reactors are not large and do not require large facilities to operate. PPPL is a small facility off the side of Rt. 1 in NJ. TFTR ran D-T plasmas, so PPPL was a tritium handling facility, which required it to be surrounded by a fair bit of land (but not that much). Here's PPPL: http://maps.google.com/?ll=40.34782,-74 ... 3&t=h&z=16. The facility is about the size of the nearby Bristol-Meyers Squibb facility. Note that PPPL also includes power-generation equipment on site to power the reactors. In a commercial facility, you'll also have to have turbines, etc. supporting the power generation process. By comparison, here is a coal-burning plant in Bridgeport, CT (which I have also worked on): http://maps.google.com/maps?q=bridgepor ... ticut&z=16.

NSTX, as I mentioned, is designed for greater than breakeven. Assume it could achieve breakeven to the point of power generation. Based on the fact that PPPL was designed to support 1300 researchers plus a reactor, the land utilization for a workable design similar to NSTX would be around the same as a coal-burning plant. The facilities requirements would likely be less: no need to have a pier and derricks to offload coal. No slurry pond. No pulverizers. Assuming that the steam turbines were essentially the same (they're more or less the same no matter what type of power plant you see, they just differ based on size), then you can assume they'll cost about the same. Thus, facilities for a fusion reactor would likely cost less, because there is no need for any of that support equipment.

4.) Computers are cheap. The main challenge to (magnetic) fusion power generation is the ability to control the plasma. Once you have developed the math to do it, all you need are computers to run the algorithms on. And, as I said, computers are cheap.

5.) "The total project cost was $23.6M; however the estimated value of the site credits is of order $77M." From: http://aries.ucsd.edu/LIB/MEETINGS/0103 ... -paper.pdf. Costs of advanced fission reactors are hard to find, but based on this press release, they seem to be on the order of $110M: http://www.bloomberg.com/news/2010-06-0 ... -2030.html. There is no word as to what role Hitachi will play in the construction of the facilities, but a nuclear power plant is much more expensive. For instance: "March 2008 -- For two new AP1000 reactors in Florida, Progress Energy announced that if built within 18 months of each other, the cost for the first would be $5144 per kilowatt and the second $3376/kW - total $9.4 billion. Including land, plant components, cooling towers, financing costs, license application, regulatory fees, initial fuel for two units, owner's costs, insurance, taxes, escalation, and contingencies, the total would be about $14 billion." (from http://en.wikipedia.org/wiki/Economics_ ... wer_plants).

Thus, even converting the $23.6M number to today's dollars, it's still below $110M. Of course, NSTX doesn't have some of the required equipment for neutron capture and heat conversion, along with tritium production. Nevertheless, the costs of adding these components is unlikely to quintuple the overall cost of the design.

[Zamfir]

Thise two fission reactors are intended to produce about 6000 mw of thermal power, nearly continuously for decades. Unsurprisingly, they are more expensive than a 10 mw machine dezigned to run for seconds at a time.

[Arisu]

The ITER reactor is going to cost on the order of hundreds of billions of dollars and require highly trained staff.

Just a quick question, but is it really that expensive? Wikipedia puts it at 15 billion Euro currently.
Or am I just missing / misreading something?

[Zamfir]

No , your pretty right. Construction is in the 10 to 15 billion euro range, operating costs another 5. Though hard numbers are a bit tricky, the partners are mostly required to deliver components, not cash money. The eu was going to deliver 45% of components (by esitmated cost ), and they currently project that share at 6 to 7 billion, somewhat more the double the original estimate. Presumably there will be some more overruns, but the whole machine should end up at less than 20 billion. Of course, that's not counting interest, as you would for a commercial facility.

[gorcee]

Thise two fission reactors are intended to produce about 6000 mw of thermal power, nearly continuously for decades. Unsurprisingly, they are more expensive than a 10 mw machine dezigned to run for seconds at a time.

Yes, and that cost includes 50 years of experience in commercial reactor design and implementation. If fusion reactors had 50 years of commercial success, as well, then I think we could scale these arguments a little better. Regardless, your comparison of seconds to decades is foolish. It's comparing apples to oranges. You're comparing the service life of a unit to amount of time it takes to generate sufficient energy to power itself. Why not compare service life to service life? How about you justify this apples to oranges comparison.

[eSOANEM]

4.) Computers are cheap. The main challenge to (magnetic) fusion power generation is the ability to control the plasma.

I visited JET (and MAST) about a year ago and, from what they were saying, this is no longer much of an issue for them. The issue they have is primarily that the magnets get too hot (which is why ITER will have superconducting magnets) and, given the data they have from JET, it seems highly likely that ITER should be able to produce >100% efficiency fusion power for sustained periods pretty much from the moment they'd properly calibrated everything so the plasma stayed in the right place.

[ekolis]

You could build a reactor in a third world country, operate it very cheaply, and provide power to the citizens without any fear that the local government would attempt to use it to build a bomb.

I'm actually curious - how is this true? Couldn't the local government pump ridiculously huge amounts of fuel into it, causing a chain reaction turning the reactor into a bomb? Or is there some self-limiting factor in the reaction that prevents this from happening? (After all, the sun hasn't blown up yet...)

[Robert'); DROP TABLE *;]

You could build a reactor in a third world country, operate it very cheaply, and provide power to the citizens without any fear that the local government would attempt to use it to build a bomb.

I'm actually curious - how is this true? Couldn't the local government pump ridiculously huge amounts of fuel into it, causing a chain reaction turning the reactor into a bomb? Or is there some self-limiting factor in the reaction that prevents this from happening? (After all, the sun hasn't blown up yet...)

A magnetic confinement fusion reactor is basically a stream of insanely high temperature/pressure plasma. To make it "explode" in any meaningful fashion means you have to make that stream even hotter and more pressurized than it already is; IOW, you need a more powerful and accurate machine. To hint at how easy that is: the major engineering problem with MCF ATM is that same confinement and pressurization. Effectively, to build a bomb out of a reactor, you need to make the reactor better at doing its job. (Rather than worse, as is the case with fission reactors.) If you have the engineering skill to do that, you might as well have just purpose-built a bomb originally.

(I believe similar reasoning applies to laser-ignited fusion, but I don't know enough to say for sure.)

[eSOANEM]

There's also the fact that the plasmas in tokamaks are not very dense at all so they are not particularly self-heating and, if the plasma does get unruly and start tearing the reactor apart, it will destroy the apparatus heating it and so will cool down very quickly indeed (because it is stupidly hot and very light) so, even if you could make it explode, it would be a heavily self-limiting explosion.

[gorcee]

You could build a reactor in a third world country, operate it very cheaply, and provide power to the citizens without any fear that the local government would attempt to use it to build a bomb.

I'm actually curious - how is this true? Couldn't the local government pump ridiculously huge amounts of fuel into it, causing a chain reaction turning the reactor into a bomb? Or is there some self-limiting factor in the reaction that prevents this from happening? (After all, the sun hasn't blown up yet...)

Even if they could do that, they'd have a stationary bomb in their own country. As far as I'm concerned, if the Iranians or North Koreans feel like blowing up their own country, they can go right ahead.

But no, this can't happen. Fusion plasmas are extremely un-dense (is there a word for the opposite of dense?) Like, around a billionth of the density of air, IIRC. So even if confinement is lost, nothing really happens. Don't forget, in current fusion tests, they heat the plasma for a 4-6 second run, and that's it. After that, confinement is lost. The reactors still work, however.

[Robert'); DROP TABLE *;]

(is there a word for the opposite of dense?)

AFIAK, "sparse" usually works.

[gorcee]

(is there a word for the opposite of dense?)

AFIAK, "sparse" usually works.

Durr, that, or "thin".

I was quite hungry when I made my last post, so I was acting quite un-sparse.

[idobox]

About the cost of reactors.

A significant part of a fission reactor cost is the pressure vessel: a big metal tank that must contain high temperature/high pressure water while bombarded by neutrons. Some generation 4, and fusion reactors won't need it, because they won't use water as a collant, and won't need to keep under very high pressure to prevent it from boiling.

On the other hand, commerical fusion reactors might require expensive stuff, like supraconductor coils, berilyum coating or whatever. Some parts might also need to be replaced often (or not, I don't know, really).

The last thing to consider is the size at which a fusion reactor becomes interresting. If fusion reactors become interresting at 1TW, it is going to cost a lot to build, and won't be fit for develloping countries and islands. while if it happens at 1MW, we could see them on civialian boats.

[Zamfir]

Yes, and that cost includes 50 years of experience in commercial reactor design and implementation. If fusion reactors had 50 years of commercial success, as well, then I think we could scale these arguments a little better. Regardless, your comparison of seconds to decades is foolish. It's comparing apples to oranges. You're comparing the service life of a unit to amount of time it takes to generate sufficient energy to power itself. Why not compare service life to service life? How about you justify this apples to oranges comparison.

My apologies for that formulation. I mainly wanted to compare "continuous operation" to "seconds", and the decades only distracted from that.

Do you much disagree with the point, though? That the NSTC is far, far smaller than the power plants you were comparing it too, and that you would require a significantly more expensive design to make the NSTC run on a power-plant style schedule? That is, at a steady state of full power, and reliably enough that maintenance outages are far apart and mostly scheduled.

I mean, yes, getting out of the experimental stage will make the machines cheaper. But factors like the above (and adding machinery to actually turn the generated energy into electric power) will also add to the cost. You seem remarkably certain that the outcome will be low costs, and I am honestly surprised by that certainty.

[PM 2Ring]

The last thing to consider is the size at which a fusion reactor becomes interesting. If fusion reactors become interesting at 1TW, it is going to cost a lot to build, and won't be fit for developing countries and islands. while if it happens at 1MW, we could see them on civilian boats.

Well, if the Focus Fusion approach favoured by the people at Lawrenceville Plasma Physics pans out, small fusion reactors may be feasible. While I find the concept of Focus Fusion interesting, I'm not totally confident in the theories of Eric Lerner of LPP, as he has some rather unconventional ideas regarding the Big Bang, but I'm trying to keep an open mind re Focus Fusion.

[DrPhil@JET]

Hi All, I work at JET fusion experiment, so I'll try and answer a few of the questions

Pauline - our website www.EFDA.org has info about experiments and fusion generally. Check outy the FAQs too. Also ITER's website ITER.org has a fair bit of material too. Please let me know if there is anything missing or unclear.

Budget - ITER is not hundreds of millions, it has just had the new figure approved of around 19 billion. Shared between the 7 countries, over 15 or so years, amounts to less than a dollar/euro/pound per person per year in Europe. Perhaps less in the other countries (e.g. US) because Europe pays more as host country.

SIZE - ITER's target is 500 MW output, but a future power plant would probably be bigger. The major issue is energy loss, through the surface of the plasma, so bigger is better - the usual surface area to volume ratio arguments... So, with current designs it doesn't look like it will be the kind of thing you could build in Rarotonga, first generation would need to be in fairly large developed countries. Who knows what developments may come in the future, such as the stellerator design, which may enable a smaller size (look out for the German Wendelstein 7X experiment starting up in 2015 or so)

POWER COST We have done some modelling of energy scenarios, and estimate that the cost will be roughly similar to fission. Fusion needs a fair bit of input power, heating systems, cooling systems and magnets - but gives you truckloads of energy out. If their is a price on carbon, we see it becoming a part of the world's energy mix around 2060, but if there is not a price on carbon, we won't be able to compete with cheap fossil fuels.

EXPLOSIONS . Fusion experiments run at very low pressure - so the lid will not pop off. If it did spring a leak, air would rush into the chamber. Regarding a runaway reaction, this would not happen as the fusion cross-section starts to decline much above our operating temperature of 10 - 15 keV (~100 - 150 million degrees) so actually it is self limiting. The amount of fuel is very finely balanced too - this is the pressure - too much or too little and the plasma dies. Injecting more fuel is actually one of our safety procedures if it looks like the plasma is getting too turbulent and we want to kill the reaction. ( if we do not, then the energy gets dumped into the wall and might only melt some of the tile surface; this is a bummer, but not dangerous like a melt down)

CHALLENGES yes there is still a bit of work to be done, mostly in working our what the materials will be for the final reactors. They need to resist heavy neutron bombardments with only minor development of radioactivty; cope with heavy heatloads; convert neutron (kinetic) energy into heat for turbines, and breed tritium from lithium in a viable way. There are research programs attacking all of these goals all round Europe, and probably the world.

oh, and "Hi Randall!" *excited*

Phil

[EdgePenguin]

The flavour of fusion that is currently under consideration produces neutrons (and lots of them, if I understand correctly) so I would assume that an unscrupulous individual could line the tokamak with U-238 and use it to make weapons grade Plutonium?

[DrPhil@JET]

yep, guess so.

[firechicago]

The flavour of fusion that is currently under consideration produces neutrons (and lots of them, if I understand correctly) so I would assume that an unscrupulous individual could line the tokamak with U-238 and use it to make weapons grade Plutonium?

Though if the resulting material is similar to the spent fuel coming out of a fission reactor, reprocessing it to extract the plutonium is a very difficult engineering problem of its own. Not quite as hard as enriching natural uranium to weapons grade, but very far from being something that you can do in your garage.

[Game_boy]

This is all on the assumption ITER works. Is there any chance we could get to the delivery date and it doesn't produce net power?

[EdgePenguin]

yep, guess so.

I really hope that fear does not stop this technology being deployed, if it lives up to its potential.

So, what is it like working at JET? When you leave the house in the morning, do you say to your other half "See you later, I'm off to save the world"?

[eSOANEM]

This is all on the assumption ITER works. Is there any chance we could get to the delivery date and it doesn't produce net power?

Of course there is. It's a new design of tokamak that's not a clone of a pre-existing one; the exact amount of energy it will produce can be predicted based on the energy output of other, smaller, differently designed tokamaks, but as it itself has never been tested, this is extrapolation and we know how that can go. Still, from what I gather, seeing as, discounting resistive losses in the magnets at JET (which would be eliminated at ITER), we've already got approximately break-even fusion power, it seems likely that, with a more optimised design than JET, ITER should produce a net output.

[Game_boy]

This is all on the assumption ITER works. Is there any chance we could get to the delivery date and it doesn't produce net power?

Of course there is. It's a new design of tokamak that's not a clone of a pre-existing one; the exact amount of energy it will produce can be predicted based on the energy output of other, smaller, differently designed tokamaks, but as it itself has never been tested, this is extrapolation and we know how that can go. Still, from what I gather, seeing as, discounting resistive losses in the magnets at JET (which would be eliminated at ITER), we've already got approximately break-even fusion power, it seems likely that, with a more optimised design than JET, ITER should produce a net output.

Thanks. I'm just seeing hype of the form of: ITER is a black box, you put $20b in and you get usable fusion power 10 years later. But good to hear it's not overpromising, based on existing results.