How would electricity be generated from a nuclear fusion reactor?

[ElliotSmith]

TL;DR Summary

How would we get energy from a fusion power plant?

How would electricity be generated from a nuclear fusion reactor?

How soon do you think that fusion power plants will become a reality?

 

[phinds]

Summary: How would we get energy from a fusion power plant?

How would electricity be generated from a nuclear fusion reactor?

How soon do you think that fusion power plants will become a reality?

What research have you done on this so far? What have you found out?

 

[russ_watters]

How would electricity be generated from a nuclear fusion reactor?

Same way most of our power is generated now: by boiling water and running it through a turbine.

How soon do you think that fusion power plants will become a reality?

The joke is that it is always 30 years away (and has been for 60 years). Scientists seem pretty confident that the next project will achieve break-even and the one after that will be a prototype power reactor. So 3 more generations of projects; still at least 30 years if the optimism is warranted.

And then we can answer the question of whether we want it or not.

 

[gleem]

Same way most of our power is generated now: by boiling water and running it through a turbine.

Good old 19 th century technology.

How soon do you think that fusion power plants will become a reality?

Researcher have been promising fusion for decades. MIT is promising a small reactor (100 MW) by 2025 don't know about the big industrial sizes that we need.

 

[anorlunda]

Does anyone remember Chaos Manner, the Jerry Pournoule column in Byte Magazine? His favorite phrase applies here. Fusion will arrive "real soon now."

 

[PeterDonis]

Fusion will arrive "real soon now."

I remember Chaos Manor very well. As I remember, he always capitalized it: Real Soon Now.

 

[berkeman]

How would electricity be generated from a nuclear fusion reactor?

I took a class on plasma physics and fusion reactors many years ago. I seem to remember that the instructor mentioned that one of the harder fusion reactions to get working would have the advantage of being able to use Magneto Hydrodynamic (MHD) power generation instead of a thermal cycle. But my Google-foo is failing me now -- Does anybody know which fusion Rx that might be, and why it lends itself to MHD power generation?

 

[Astronuc]

I took a class on plasma physics and fusion reactors many years ago. I seem to remember that the instructor mentioned that one of the harder fusion reactions to get working would have the advantage of being able to use Magneto Hydrodynamic (MHD) power generation instead of a thermal cycle. But my Google-foo is failing me now -- Does anybody know which fusion Rx that might be, and why it lends itself to MHD power generation?

Likely, it was an aneutronic reaction, such as DD or D³He. There were numerous concepts on 'direct energy conversion', in which one would want most, if not all, energy in the form of nuclei and electrons.

https://en.wikipedia.org/wiki/Direct_energy_conversionSee the section on Induction > conduction systems
https://en.wikipedia.org/wiki/Direct_energy_conversion#Conduction_systems

 

[essenmein]

Read a random thing the other day that said D-T fusion produces about 80% of its energy in the form of high energy neutrons (~14MeV), the hard part is slowing them down to get the heat out (thermalize them), also bit of a problem for any materials => resulting in potentially large amounts of radioactive waste due to the neutron bombardment of previously not radioactive materials.

Interestingly it was also mentioned that there is a significant proliferation risk since production of Pu239 is relatively easy by just putting some uranium (depleted or other wise) somewhere near those neutrons.

 

[etudiant]

Think that is perhaps why none of the fusion efforts are getting enthusiastic government support.
The idea of creating electricity from star fire by boiling water seems a little incongruous, sort of like the early requirement to have a man with a flag precede motor vehicles. Plus the current ITER concept is proving to be so expensive that it produce power economically, even in the revised improved follow on design. It will hopefully prove that fusion is practicable, then the engineers will work to make a viable system starting from there.

 

[anorlunda]

By the way, do they have a plan for the Tokamak type fusion reactors to convert the energy to steam?

 

[etudiant]

Afaik, the idea is that ITER provides proof of concept, based on which an actual prototype fusion power plant would be built around 2050. That was briefed as a steam turbine based concept when I visited ITER some years ago. I don't know whether anything has changed since.

 

[anorlunda]

What about power density and temperature gradients in the steam production?

One article (sorry, lost the link) said that we would need 10000 MW/m³ power density, and 1 million degrees K per m temperature gradient. It is not easy dealing with such extremes.

Does anyone have a source on those issues?

 

[Xforce]

What about power density and temperature gradients in the steam production?

One article (sorry, lost the link) said that we would need 10000 MW/m³ power density, and 1 million degrees K per m temperature gradient. It is not easy dealing with such extremes.

Does anyone have a source on those issues?

Why do we have to use steam turbines? Again, that’s the technology of yesterday, too bulky and too many processes of energy conversion (radiation→Heat of reaction chamber→Heat of H₂O→Kinetic energy of turbines→electricity) something that often seen in a Rube Goldberg machine... If we have to use the heat generated by fusion (actually, I believe that heat should be maintaining the fusion reaction. So we don’t have to pump extra heat to the plasma) why don’t just convert the heat directly into heat via thermoelectric materials?

Or in the other case we convert the kinetic energy of charged particles to electricity (direct energy conversion)

 

[anorlunda]

See
https://en.wikipedia.org/wiki/Thermoelectric_generator

The typical efficiency of TEGs is around 5–8%.

 

[Xforce]

The typical efficiency of TEGs is around 5-8%https://en.wikipedia.org/wiki/Thermoelectric_generator

Well, there is no reason to think of this efficiency will not improve. TEG isn’t ready yet, but so is fusion. Consider the first generation of steam engines have a very low efficiency (below 5%) and now these gas turbines still can reach 50% or more... by the time fusion is ready, these TEGs should have enough efficiency though. And I believe TEGs are one of the key technologies that make miniature fusion generators possible.
Other than the thermoelectric cycle, I prefer more of the DEC cycle for aneutronic fusion (D-He3 is 16 times harder than D-T though) ... by directing the vector of speed of the charged particles, we focus them in a beam and harness their kinetic energy. This should be as efficient as a electric motor (since they works about the same) and up to 90% of efficiency (that’s way above the best turbines!) . Plus, Wikipedia says “Direct energy conversion (DEC) or simply direct conversion converts a charged particle's kinetic energy into a voltage. It is a scheme for power extraction from nuclear fusion.” This technology is optimized for fusion!
You got to see this! https://en.m.wikipedia.org/wiki/Direct_energy_conversion

 

[russ_watters]

Well, there is no reason to think of this efficiency will not improve.

Wishful thinking is not engineering.

Why do we have to use steam turbines? Again, that’s the technology of yesterday...

Not that it actually matters, but thermoelectric generators are about 2/3 as old as steam engines.

 

[anorlunda]

Even if we had direct conversion that is 90% (more likely 50%) efficient, there is still a huge amount of waste heat to get rid of. The problems with power density and thermal gradients remain.

At what temperature does a direct conversion device melt?

 

[gmax137]

It might be efficient, but boiling water and power a turbine might be common somehow, but it will be very bulky (not ideal if we want to fit it onto a spacecraft , or a ship (like carriers and nuclear subs).

Actually today's carriers and nuclear subs do use steam turbines to convert the heat produced in their (fission) reactors to propulsion. What seems Rube Goldberg to some is proven technology, reliable and well-understood.

Spacecraft is another realm completely, you're probably right about that.

 

[DEvens]

How would electricity be generated from a nuclear fusion reactor?

For several designs that's a difficult point.

Most of the variations on magnetic confinement are doing their absolute best to keep the heat in. So far, nobody has included any place to transport heat out to any kind of converter. So this would need to be at the next stage of development. You'd need to have some way to open some kind of passage for the heat. Maybe you have a "dump heat to the generator" mode in the magnetic field. So you might push the plasma to it's hottest, then dump the heat to a specially designed port or panel or something. Dump some heat then go back to normal operation mode. This is still to be worked out.

Possibly inertial confinement has an easier job on that point. The fusion lasts a very short time producing a pulse of energy that can be captured in any convenient format.

There is one outside-the-main-line design that has this already worked out. The General Fusion design uses molten lead as the wall of the fusion chamber. It circulates in, picks up the heat, and circulates out. Then the lead goes through a fairly familiar heat exchanger to pass the heat to turbines.

https://generalfusion.com/

How soon do you think that fusion power plants will become a reality?

I'd say there are too many unknowns to hazard a guess. Primarily it's a question of exactly what is possible. We don't know yet which method or design will work. Or even if any of the current designs will work in an acceptable fashion.

After that there are tons of motivation questions. How much do people want it? What will be the price of fossil fuels? What will be the attitude towards fission reactors and what will be the price of Uranium or Thorium? What will be people's ideas about carbon into the atmosphere? How much electricity, and other forms of power, do people want? Will we continue to convert to electric cars and so need much more electricity generation? And bunches of other questions I've not thought of. These considerations will determine how much funding fusion gets, and so how hard people work at it.

In the meanwhile, we should not be ignoring things like improved fission reactors. And improved generation of all kinds. Fossil, wind, solar, tidal, geothermal, even things like burning garbage and waste products from saw mills should be investigated for improvement. Improved energy supply is important for nearly every activity of modern society. We should keep as many options going as we can afford.

 

[anorlunda]

So you might push the plasma to it's hottest, then dump the heat to a specially designed port or panel or something.

Am I correct that the chamber holding the plasma is maintained in vacuum?

If so, then the port would need to be transparent to particles and radiation, but also act as a pressure barrier to protect the vacuum. Are there any preliminary ideas for such a port?

 

[artis]

well maybe I'm mistaken and did not catch your idea here quite clearly but wasn't ITER and any major tokamak or spheromak design all about having the alpha particles help with plasma heating (after all they are charged and so can't escape the plasma easily) and the neutrons escaping the plasma doing the heating in the "blanket" from which heat is taken off by plain old circulation?
As far as I know there is no direct conversion for ITER and most other magnetic confinement designs , it's all about keeping the plasma "burn time" for as long as possible , taking off the heat produced by neutrons to convert to electricity and then recharging the torus plasma vessel by taking fused plasma leftovers out and fresh DT fuel in, I suppose in gas form which then gets ingnited once again and so the cycle repeats.

 

[mfb]

@artis: Right.

Getting the energy out is easy - it is unavoidable in DT fusion. The heat goes to the walls, the cooling fluid heats up, the rest is similar to a fission reactor.

 

[MrCosmic]

Likely, it was an aneutronic reaction, such as DD or D³He.

@Astronuc, most aneutronic designs use D-H3.
I've never seen anything on Helium3-Helium3. Are the fusion requirements to high compared to other aneutronics.
I wanted to write a story on realistic fusion space travel like The Expanse. I wanted to use H3-H3 fusion rockets. But can't find any numbers on that type of reaction.

 

[mfb]

H3-H3? Fusing tritium with tritium?

In increasing difficulty and decreasing neutron yield:
D+T (always produces a neutron)
D+D (50% of the reactions)
D+He3 (has some neutrons from D+D side reactions)
He3+He3 (a tiny bit of neutrons from He3+e -> T production)
p+B11 (neutron-free)

Note: We don't even know if D+D can burn stably on its own, i.e. can be used for net power output at all. We only know this about D+T.

 

[MrCosmic]

H3-H3? Fusing tritium with tritium?

In increasing difficulty and decreasing neutron yield:
D+T (always produces a neutron)
D+D (50% of the reactions)
D+He3 (has some neutrons from D+D side reactions)
He3+He3 (a tiny bit of neutrons from He3+e -> T production)
p+B11 (neutron-free)

Note: We don't even know if D+D can burn stably on its own, i.e. can be used for net power output at all. We only know this about D+T.

Yes, He3+He3(Helium3+Helium3)

I wanted to reduce radiators for the story, He3+He3 seemed like a good one.
Do you know the Lawson criteria for it?

 

[mfb]

Well, much larger than 1, the same for every reaction. The cross section is much lower and you need higher temperatures, that makes everything worse. https://www.researchgate.net/figure/Reaction-constants-for-different-fusion-reactions-7_fig1_283114377 - you need nearly two orders of magnitude higher temperatures (which means 6-8 orders of magnitude higher energy loss) and still get an order of magnitude lower fusion rates or so.

 

[cmb]

I've got a question which is more related to DEMO I guess than ITER, because the latter is not meant to be 'operational'. However, let's say ITER *is* a Q=10 device (that's the goal; https://www.iter.org/sci/Goals) so even if the breeder blanket technology is not present, should it not be a viable device for proving the thermal power can be extracted?

There appears to be a conflict in the goals where it is designed to operate for 400 to 600s, but it will take around 600s for the US proposed lithium lead breeder blanket to melt. Once the pulse stops, it'll freeze again and no thermal power can be circulated out of it?

So my questions are; if ITER is to be a Q>10 device and its design was taken forward as a power source, and a viable breeder blanket put around it, how would that breeder blanket work if it takes as long to melt the blanket coolant than the predicted pulse length? If the time to the next pulse was short, then obviously it could continue with then-melted coolant, so the next question is; how long would it take between pulses to reset a tokamak for the next pulse? I've not seen repetition rate mentioned as a goal for ITER.

Here's a calculation on how long the blanket takes to come up to operating temperature, please let me know if it is in error;
ITER's goals are 500MW for up to 600s

ITER will have a surface area of around 700m^2 (assume R=6m and r=3m)

Lithium has a specific heat capacity of 24.86J/mol/K, density of 0.534g/cm3, melting point of 180C and heat of fusion of 3kJ/mol. [Data from https://en.wikipedia.org/wiki/Lithium]

So, to take one cc of lithium from room temperature (say 20C) to a suitable temperature for a steam cycle, say 500C, would take; (480K * 24.86J/mol/K * 0.534g/(7g/mol)) + (0.534g/(7g/mol) * 3kJ/mol) = 910J + 228J = 1150J.

Lead has a specific heat capacity of 26.65J/mol/K, density of 11.34g/cm3, melting point of 600C and heat of fusion of 4.77kJ/mol. [Data from https://en.wikipedia.org/wiki/Lead]

So, to take one cc of lead from room temperature (say 20C) to a suitable temperature for a steam cycle, say 500C, would take; (480K * 26.65J/mol/K * 11.34g/(207g/mol)) + (11.34g/(207g/mol) * 4.77kJ/mol) = 700J + 261J = 961J.

Details of the US test blanket module for ITER are given here; http://www.telegrid.enea.it/Conferenze/SOFE05/DATA/05_11.PDF

It is a 41.3cm thick blanket of lithium lead eutectic at 17% Li and 83% lead.

This would therefore be approx 41.3cm * 700m^2 = 2.9E8 cm3, so 49E6cm3 of lithium and 240E6cm3 of lead.

Total energy to heat this blanket up, if it were to surround the ITER vessel, would be (49E6cm3 * 1150J/cm3) + (240E6cm3 * 961J/cm3) = 56E9J + 230E9J = 286E9J.

So, the total run time for ITER to melt its lithium lead blanket to molten and provide coolant and thermal output would be 286E9J/500MW = 572s.

 

[artis]

I would also love to hear the answers to your questions,
now as much as I know IIRC, is that previous test tokamaks did not have a blanket but then again their maximum power was much lower, in Iter case the power that they aim is high enough so a neutron barrier and heat removal is necessary , because it serves both as the coolant heater part as well as (probably most importantly) the buffer which would absorb most of the neutron flux which being very strong would otherwise probably damage the magnets and other systems rather fast.

 

[mfb]

You can water-cool a solid blanket.

I didn't see a repetition rate for ITER either but I could imagine that they start slowly. Make a single pulse, analyze it thoroughly. Make another pulse, study it. After a while: Make a short pulse, then do a quick analysis of the data until the following day, repeat, long-term analyses running in parallel. Once they are confident nothing breaks they might go to a few pulses per day and go to higher rates for endurance tests later (on a timescale of many years). Electricity production from such an infrequent operation is not useful, the installation of the infrastructure would cost more than selling the electricity would bring, and it would make the overall reactor more complicated.

The 500 W will only happen with deuterium-tritium operation after years of running without relevant fusion power, and even then they don't have to happen often.

 

[cmb]

You can water-cool a solid blanket.

Of course, but then you don't get tritium to keep the thing running.

There has to be at least one new recoverable tritium atom for every neutron released in a deuterium-tritium fusion reactor. This is a completely integral scientific and technical requirement to generating electricity from those neutrons. Each neutron has to both deliver its 14MeV of energy to the coolant and also generate the tritium. It's a big ask.

Tritium generation is part of the same question 'how would electricity be generated'.

This is possibly a bigger scientific hurdle than creating a plasma that's hot enough. At least JET generated Q=0.6 for half a second. Has anyone ever generated recoverable tritium from a fusion reactor before?

It has to generate tritium. It can't buy the stuff for electricity generation. If it did, the costs work out like this;

Price of tritium from fission reactor production = $30,000/g

Total number of possible DT reactions per gramme of tritium at 100% fuel utilisation = (6E23/mol)/(3g/mol) = 2E23/g of reactions

Total neutron energy released = 2E23/g * 14MeV * 1.6E-19J/eV = 448GJ = 125MWh

Fuel cost per electricity unit (tritium cost only, ignoring over-night and operating costs) = $240/MWh

 

[mfb]

ITER will test some breeding concepts but it doesn't need to produce its own tritium. It is a scientific experiment, not a power plant. DEMO is expected to breed its own tritium after an initial supply.

 

[cmb]

ITER will test some breeding concepts but it doesn't need to produce its own tritium. It is a scientific experiment, not a power plant. DEMO is expected to breed its own tritium after an initial supply.

Of course this is the case.

But if I were to propose an experiment in which I aim to generate tomorrow's energy by burning gold, wouldn't the first (not second) question be where to get the gold from? If I said that'd be from transmuting base metals, wouldn't the scientific community ask for evidence that is possible, before putting much credence in gold as a source of future energy?

Tritium generation is a totally integral requirement to fusion power, and so the thread question, how to generate electricity from it, requires an answer to both absorbing neutron power and generating tritium from it. I don't think the OP's question was about ITER (it couldn't be, because ITER won't make electricity).

The additional dimension to the question is that if a future fusion power station were to operate in a pulsed mode, how long is the pulse/cycle time and what is the impact of that on generating power from it? If the future for a fusion power station is not a pulsed mode, then ITER can't really scientifically conclude much about continuous operation. So I don't think we know an answer to the OP's question, yet, and it's unclear if anyone really has a good idea about it.

 

[etudiant]

I do not have the actual data on the neutron economy, but afaik, the idea was to breed tritium from the lithium blanket around the tokamak. Afaik, the tritium supply has not been seen as an item of concern, as the expectation was that the operations would yield a surplus of tritium.

Re the power generation, I believe that the idea is to use the thermal inertia of the blankets to insure a continuous supply of steam even though the actual reactor operations are intermittent. Obviously one could rail at the idea of using fusion to boil water, but we are still a long ways from any direct fusion to electricity hardware, even if the concepts are pretty clear.

 

[PeterDonis]

Tritium generation is a totally integral requirement to fusion power, and so the thread question, how to generate electricity from it, requires an answer to both absorbing neutron power and generating tritium from it.

And you already have the answer to how tritium will be generated. The nuclear reactions involved in breeding tritium from lithium are well understood, so the expectations about being able to breed tritium from lithium in an actual power reactor are much better grounded than your hypothetical about making gold by transmutation of elements.

 

[mfb]

@cmb: Don't confuse "I don't know how it is done" with "it is unknown how it is done". Just because you are unaware of the research doesn't mean the research wouldn't exist. The tritium yield if you shoot neutrons into lithium is well-studied. How to extract the tritium efficiently is one of the things ITER will study. It doesn't need to produce all its tritium for that aspect, or even a significant fraction. If it can show that it could produce 8% of its tritium demand using 5% of the blanket area that is perfectly fine.

I don't think the OP's question was about ITER

It wasn't about tritium breeding either.

The additional dimension to the question is that if a future fusion power station were to operate in a pulsed mode, how long is the pulse/cycle time and what is the impact of that on generating power from it? If the future for a fusion power station is not a pulsed mode, then ITER can't really scientifically conclude much about continuous operation. So I don't think we know an answer to the OP's question, yet, and it's unclear if anyone really has a good idea about it.

ITER aims at 10 minutes, DEMO at 2 hours. Tokamaks have to be pulsed, but the time between pulses can be very short if the magnets can ramp fast enough. Short enough and you can produce electricity continuously. A bit longer and you want some short-term on-site energy storage. The impact on the cost per kWh would be negligible. And it doesn't impact the way electricity is generated from fusion.