Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

ICF energy extraction

  1. Dec 26, 2011 #1
    The purpose of this thread is to spark discussion of how to generate energy from Inertial Confinement Fusion. I have a few questions about the process, so hopefully someone who's more in-the-know will have the time and the inclination to answer them. (I would appreciate it greatly)

    With respect to ICF (Inertial Confinement Fusion):

    If I understand correctly, you use mixture of Deuterium/Tritium pellets; you fire a laser (possibly even 192) at the pellet, causing fusion within the pellet. Once Fusion occurs in the pellet, you get much more energy in heat than you put in. That's the ideal scenario.

    Then there's the question; how are you going to make sure that enough heat goes to where you want it to go (i.e. a big "bucket" of water attached to a steam turbine)? It would be really interesting to play with LASNEX and/or HYDRA, and play with several design ideas; unfortunately, I have neither one on my calculator. ;) ;)

    1) Shoot pellet with lasers.
    2) Make pellet go boom.
    3) Move heat from boom to vat of water to generate steam, which will in turn generate electricity. (Note: none of these energy conversions occur at 100%. So you might assume 10% for a minimum amount)

    Anways, there are several problems with this approach:
    1) You have an energy budget throughout this process; you want to generate (at the steam generator in electricity) more than you put in.
    2) Heat is fickle, it tends to go where IT wants to go; not where YOU want it to go.
    (Is there an engineering idea analogous to electrical circuitry for controlling the motion of heat flux instead of electrical current? [I've never seen anything more advanced than what you'd see in a regular class on Thermodynamics/Statistical Mechanics])
    3) It's difficult to produce the temperatures and pressures necessary to generate fusion.

    Questions about the process:
    1) Fusion is tough to obtain; assuming 100% of the pellet's were to fuse, how much energy would that generate? Realistically, how much energy are we getting out of it? (Have experimental trials even begun?)
    2) We're using lasers to power the process; what does that electric bill look like?
    3) How accurately do the computer models reflect the measurements from experimental trials?
    4) Are there other techniques than generation of electricity via steam that have been suggested? (I.E. generation of electricity via heat like a thermocouple)

    Slightly unrelated question:
    Might it be possible in Magnetic Confinement Fusion to produce a magnetic field that's reactive to the motions of the plasma? When the plasma pushes the magnetic field detects the change and pulls and when the plasma pulls the magnetic fields push. (Is this made significantly more difficult by the fact that there aren't many stability solutions to the Magneto-Hydrodynamic Equations?)

    I apologize if any of these ideas appear unclear, I'll do my best to clarify these idea's when confusions arise.

    To throw an article out for discussion:

    Analysis of the National Ignition Facility Ignition Hohlraum Energetics Experiments
    Published in: Physics of Plasmas, vol. 18, no. 5, April 7, 2011, p.056302
    Last edited: Dec 26, 2011
  2. jcsd
  3. Dec 26, 2011 #2
    1) First question answered (roughly):

    In the reaction:
    Deuterium+Tritium ---> Helium
    (2H)+(3H)--->(4He)+17.6 MeV

    And assuming that we have a 10 mg pellet, which corresponds to about 2.5e21 particles. If half of those fuse with themselves, (perhaps assuming about a 50% Deuterium, 50% Tritium Composition). Roughly 1.25e21 undergo the reaction... this means that on the order of 10^9 Joules should be produced for a 100% reaction of the pellet. (Unless I made a stupid mistake, please point it out if I did). This gives us our maximum budget. This is equivalent to the chemical energy in a barrel of oil.
    Last edited: Dec 26, 2011
  4. Dec 26, 2011 #3
    2) A possible answer to the second question:

    -- National Ignition Facility (NIF) at LLNL in California, US [29]
    -- Laser Mégajoule of the Commissariat à l'Énergie Atomique in Bordeaux, France (under construction) [30]
    -- OMEGA EL Laser at the Laboratory for Laser Energetics, Rochester, US
    -- Gekko XII at the Institute for Laser Engineering in Osaka, Japan
    -- ISKRA-4 and ISKRA-5 Lasers at the Russian Federal Nuclear Center VNIIEF [31]
    -- Pharos laser, 2 beam 1 kJ/pulse (IR) Nd:Glass laser at the Naval Research Laboratories
    -- Vulcan laser at the central Laser Facility, Rutherford Appleton Laboratory, 2.6 kJ/pulse (IR) Nd:glass laser
    -- Trident laser, at LANL; 3 beams total; 2 x 400 J beams, 100 ps – 1 us; 1 beam ~100 J, 600 fs – 2 ns.

    Using 1 kJ/Pulse, assuming perhaps 100 lasers firing on the same point. This would be 10^5 J per shot. Considering our total budget of: 10^9 J, this doesn't seem to be so bad... of course, I have no real idea if the above estimate for the laser pulses is really in the ballpark; when laser energies are reported, that could either be the energy contained in the beam itself (meaning that there's more energy that is required to generate the pulse). Or the laser energies that are reported could be the pulse generation energy. Or it could be something else. (If you know, please tell me)
  5. Dec 26, 2011 #4
    The third question is answered in the article I cited after reading it more carefully...
  6. Dec 27, 2011 #5
    AFAIK ICF is in practice just nuclear weapon research with a jacket of civil use. They are not working at all on methods on how to use the fusion energy gradually but to better understand how the Teller mechanism of radiation-compressed fusion material works.
  7. Dec 27, 2011 #6
    Well, the energy limits cut kind of close by a back of the envelope calculation... but that was the nature of fusion research in general from what I understand.
  8. Dec 27, 2011 #7
    Well, one of the ITER reactor's main design issues is to design a casing that can absorb the generated energy (in the form of neutrons) without becoming too quickly damaged by the intense neutron radiation and without poisoning the plasma with heavy ions (who loose energy too quickly by bremstrahlung because this radition loss goes with the 6th power of the ion charge). A completely new design is installed in the JET reactor now and being tested.
  9. Dec 27, 2011 #8
    Do you know how the ITER reactor's magnetic field is composed? Is it just a strong magnetic field that lenses at a specific volume? Or does the magnetic field actually react to the motions of the plasma? The ideal scenario that I'm imagining would be that the magnetic field could 'sense' if the plasma is too sparse or is likely to diffuse in a certain area and strengthen itself accordingly. I'm not sure if this is technologically possible from an engineering standpoint... I would imagine that programming the field to adjust itself in a way that would keep the main body of plasma stable would be pretty difficult.

    Are the design specifications for the ITER reactor public domain?
  10. Dec 27, 2011 #9
    The magnetic field of all tokamaks contains of a combination of a poloidal field, generated by the electrical current through the plasma, and a toroidal field, generated by external coils around the plasma. Together this results in a helical field, a field that is wound around the current. Electrically charged particles circle around these field lines, but also have a noticable influence on them. Instabilities can easily occur, lots research is done in this field.

    The design specs for ITER can probably be found, but I don't know exactly into which detail.
  11. Dec 27, 2011 #10
    I see. That's the impression that I got...


    The figure up above is supposed to represent a toroid. If the poloidal field from the plasma in section 3 is becoming too diffuse, then doesn't that mean that all you can do is push more current through the coils and hope that'll restabilize the plasma? Is there no way to identify the instability in section 3 and strengthen just that section? I'm pushing at a concept of a dynamic equilibrium which is used in our aircraft design.

    Simply making the total field strong and hoping that'll keep the plasma tight enough for fusion might be analogous to having an 'ant' problem and trying to solve it by crushing the ant colony with your bare hands... you might kill the bulk of the population that you grab, but a lot will slip through your fingers in the process...
  12. Dec 28, 2011 #11
    Well, one problem is measuring the magnetic field in real-time. It's not like hanging a probe in a plasma of this temperature works. Diagnostics are developed, and theory is being worked on. In JET they can hold the plasma stable long enough (I believe 20-30 mins), that is: if it were a real fusion plasma, they would have to replace the entire plasma due to impurities like helium nuclei anyway after that time, so keeping it stable for a longer time is not necessary.
  13. Dec 28, 2011 #12
    So no filtration techniques exist to separate the heavy ions from the lighter ions? There is no such thing as a molecular-level centrifuge?

    It looks like something like that does exist...

    Joanna Karczmarek, James Wright, et al. "Optical Centrifuge for Molecules". Physical Review Letters. Volume 82, Number 17. 1999.

    J. Karczmarek mentions that this technique was used to separate Cl35 from Cl37. She also mentions that it should work for any anisotropic molecule. It involves placing an anisotropic molecule in a linearly polarized infrared laser field. The principle that's exploited is the idea that you can distinguish between molecules based on their moments of Inertia. Could you really adapt this for a Tokamak? I'm not an expert of this technique, but the author of this paper and others in her field might know. I'm just throwing a bunch of ideas out there and hoping that they stick. :) ;) I want to see commercial fusion for energy generation in my lifetime.

    Regarding how to detect the poloidal magnetic field; introducing a toroidal magnetic field induces stresses in the frame of the tokamak, right? Would it be possible to embed stress sensors into the frame and use the detected stresses to extrapolate the magnetic field? The stresses come as a result of the toroidal magnetic field from the tokamak interacting with the magnetic field of the plasma. Since you're generating the magnetic field for the tokamak, that is a known quantity. Would this not allow you to measure the magnetic field in real time?
  14. Dec 28, 2011 #13
    Extracting the heavy ions alone would make the plasma electrically charged, and that causes complications you don't want. And anyway, in practice this is not a problem, maybe unless you want to extract helium, but because that is used to heat up the plasma (this is the helium that is produced in the DT fusion reaction) I don't think anyone suggested that. Most studies I read assume that periodically cleaning up the plasma is a good idea anyway.
  15. Dec 28, 2011 #14
    So that's a possible hit then. :)

    Helium is used to heat the plasma? How does that work? I thought that it was merely a byproduct of the fusion reactions, after those reactions occur it becomes an impurity in the plasma, right?
  16. Dec 28, 2011 #15
    Helium is a product of the fusion reaction. About 20% of the energy that is released in the reaction comes free as kinetic energy of the He nucleus (the other 80% is the kinetic energy of the neutrons, which are supposed to heat up the reactor mantle). The He nuclei interact with the plasma and lose some of their energy to it.
    Last edited: Dec 28, 2011
  17. Dec 28, 2011 #16
    Oi... that means that's 20% of the fusion energy that can't be claimed for electricity generation right there! (Unless someone has a clever way to extract it, but I have no clue how you'd do that... that's bound in the plasma) I'm not quite sure what you mean by the reactor mantle... is mantle another name for the plasma, or is it the fluid that's supposed to carry the heat for the electricity generation? Is the heating efficiency of the mantle around 25%, or lower?

    Energy flow chart:
    [All percentages displayed in fractions of the principle amount of energy from the rxn (our budget)]

    "Fusion Rxn" ---> "4He (20%)+n(80%)" ((Isn't there some X-ray radiation that's generated too?))

    "n(80%)" ---> "Heated_mantle(20%?)+n(??%)+radiation(??%)"

    I'm assuming a 25% efficiency for heat conversion.
    "Heated_mantle(20%)" ---> "Water_Heat(5%)+radiation_in_water(??%)"

    "Water_Heat(5%)" ---> "Turbine_Motion(??%)"

    "Turbine_Motion(??%)" ---> "Electricity(??%)"
    (I'm sure there's a lot here that I'm missing)

    In short what was the principle energy that was budgeted for the reaction? How much of it was lost? Where and by how much?
  18. Dec 28, 2011 #17
    I appreciate the time that you've spent humoring my questions, johanw. :)
  19. Dec 28, 2011 #18
    That's true, but since keeping the plasma temperature high also requires a lot of energy that doesn't really matter. It would otherwise have to be inserted in another way.

    No, it's the material of the reactor surrounding the plasma. It used to be made often of carbon because of its low Z (so low energy losses when carbon atoms get into the plasma), but the latest version installed in JET (which is based on the ITER design) is made of berylium, with some tungsten in high-temperature area's.

    According to the calculations I know, the overall efficiency will be that about 30-35% of the energy that is released in the fusion reaction can be converted into electricity. That's approximately the same efficiency as most other power generating processes currently in use.
  20. Dec 28, 2011 #19
    That's a lot better than I expected. So, I guess that the trick is to try and generate more fusion so that there's a higher principle energy to work with.

    It looked like you did the programming for your thesis in Java, which is similar to c++; I suspect that with all of your references to the JET reactor that it may be possible that you're working on it as a programmer. Is that correct? Or do you just have a strong interest in this subject?
    Which computer languages do you use?

    Perhaps the question is inappropriate, if so, I apologize... you're not at all obligated to respond to it; I wouldn't hold it against you. :)
    Last edited: Dec 28, 2011
  21. Dec 29, 2011 #20
    Yes, that's what ITER is supposed to do.

    If you are refering to http://johanw.home.xs4all.nl/afstud-l.zip : no, that's Pascal. When I graduated in 1993, Java didn't even exist yet.

    I work as a programmer but not on JET or any other fusion lab, I just prefer to know a little what's going on in that buisiness. I work mostly with C++ and Delphi, and sometimes with C#.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook