What's the status of fusion energy? (not CF )

[Chi Meson]

What's the status of fusion energy? (not "CF")

So, I know that fusion as an energy source has hit a lull of sorts. Could somebody fill me in on a few points (and correct any misconceptions I reveal in my questions). I'm working on a motivational lecture for my high school students, and I'd like to get things right. Note: I do not want anyone to talk about "cold fusion" in this thread.

So the ongoing problem is that fusion creates temperatures at tens of millions of degrees, and therefore is very hard to contain. The particles are held in intense electromagnetic fields that require more energy to create than is released by the reaction. I learned this when I got my degree 20 years ago, and this is still the standard story now. So what's been going on for two decades?

An associate of mine mentioned that when supercomputers get a few more orders of magnitude faster, we might build more accurate computer simulations of possible reaction situations. Sounded OK to me; Any validity to this conjecture?

Is there a theoretical certainty that people are working towards, or are we still stabbing away at things? Are we waiting for a technological breakthrough ('unobtanium" or something).

I'd appreciate any info or links. Thanks n advance.

 

[Astronuc]

Waiting for ITER to get up and running.

It's still confinement time and reaction rate.

 

[Chi Meson]

Waiting for ITER to get up and running.

It's still confinement time and reaction rate.

Reaction rate: is the rate too slow, or too fast? I'm thinking it's too fast (becuase "too hot" and "too fast" tend to go together).

Any opinions on the "fast computer simulation" conjecture?

 

[Chi Meson]

I looked up ITER;

yikes. 10 years to go. Any of my students could be doing their post-docs before it starts up!

What's the confidence-level among people foremost in the industry? ITER is a scaled-down model of possible emergy production plants, so it is mostly a test of the best idea so far, am I right? The entire scientific world is behind ITER (apparently but that's according to their own webstie) but is there chatter of hopefulness or cynicism or "wait and see" out there?

Are all the eggs in this one basket now (of course not, but are they)?

 

[mheslep]

Waiting for ITER to get up and running.

It's still confinement time and reaction rate.

Add: handling the neutron flux to the containment walls in a cost effective way. Last look even ITER documentation estimated annual replacement - that is of the innermost guts of the reactor, at least.

 

[Moridin]

It is not just the engineering features that needs to be handled and built to function. There is also the aspect of politics. Some parties have withdrawn and then rejoined and there is nothing that says that this will not happen again. Telling politicians to spend a total of about 10 billion USD on something which they have minimal to no knowledge of is a tricky business. Remember that this project is estimated to be around for 20 years to come, so there is going to be costs in the future as well.

 

[berkeman]

You've probably been following the pertinant threads in the Nuclear forum, including this interesting recent one about fusion waste products:

https://www.physicsforums.com/showthread.php?t=121166

Also, LLNL is leading the investigations into inertial confinement fusion:

http://www.llnl.gov/nif/icf/icf.html

partly to see if ICF is a viable energy producer, but also as an alternative to some of the underground nuke testing that is no longer an option. The latest ICF machine is due to go online in 2010 -- it is one big bad mama. Don't know what its prospects are for energy breakeven, though.

 

[Chi Meson]

You've probably been following the pertinant threads in the Nuclear forum, including this interesting recent one about fusion waste products:

https://www.physicsforums.com/showthread.php?t=121166

Also, LLNL is leading the investigations into inertial confinement fusion:

http://www.llnl.gov/nif/icf/icf.html

partly to see if ICF is a viable energy producer, but also as an alternative to some of the underground nuke testing that is no longer an option. The latest ICF machine is due to go online in 2010 -- it is one big bad mama. Don't know what its prospects are for energy breakeven, though.

I just now noticed that this is not IN the nuclear forum! My aim was off a bit on the forum-jump menu. Anyone care to move this for me?

 

[berkeman]

Moved to Nuclear Engineering.

 

[theCandyman]

partly to see if ICF is a viable energy producer, but also as an alternative to some of the underground nuke testing that is no longer an option.

I thought they got all their funding for that research from the Department of Defense? I rarely see anything about ICF in the journals I read.

 

[curie]

It is not just the engineering features that needs to be handled and built to function. There is also the aspect of politics. ..QUOTE]

Politicking contributed strongly to the delay in deciding where to site ITER for far longer than seemed necessary. As well as the politicians having a minimal knowledge of the technology they have to consider the demands of their public, who will have even less knowledge & will want to know why so much is being spent on a "boffins' toy" rather than new cancer treatment centres etc

It's a running joke that viable fusion power is always 20 years away

 

[vanesch]

It's a running joke that viable fusion power is always 20 years away

That's correct! I followed some special-option courses on fusion plasma physics when I did my engineering degree (1985-1990) and there too, the enthousiast professor was telling us that this would be up and running commercially in 20 years

I grew a bit skeptical over the years. Not that I think that it is impossible (on the contrary), but that by the time that all the engineering problems have been solved, the thing will be such a monster, that it is not the clean, nice, attractive "seawater" energy source it was supposed to be, and maybe not even commercially viable because of the complexity.

 

[Morbius]

I thought they got all their funding for that research from the Department of Defense? I rarely see anything about ICF in the journals I read.

Candyman,

Then you're not reading the right journals - because the research is publisched.

The Dept of Defense doesn't fund ICF research - the Dept of Energy does.

However, the Dept of Energy funds BOTH research for energy, and research on
nuclear weapons.

You don't find the funding for nuclear weapons in the Dept. of Defense budget; it's in the
Dept of Energy budget. The entire complex of design laboratories and production
facilities for nuclear weapons are part of a semi-autonomous agency within the
Dept. of Energy called the National Nuclear Security Agency or NNSA:

http://www.nnsa.doe.gov/

Not only does NNSA design and produce the USA nuclear stockpile; it does
work in non-proliferation, naval reactors, responds to nuclear emergencies...
To see the NNSA's mission statement:

http://www.nnsa.doe.gov/aboutnnsa.htm

ICF research is under the "Defense Programs" part of NNSA:

http://www.nnsa.doe.gov/defense.htm

Dr. Gregory Greenman
Physicist

 

[theCandyman]

ICF research is under the "Defense Programs" part of NNSA:

http://www.nnsa.doe.gov/defense.htm

That's what probably made me think it was funded by the Department of Defense.

I said I rarely see any articles about it, there might be 3 or 4 articles I see in the journals at my institution that are about ICF.

 

[Chi Meson]

Thanks for all the insight so far. It seems that there is a cynical attitude toward the success of fusion energy, is that fair to say? And the difficulty is almost entirely technological, as opposed to theoretical, yes?

Again, would computer simulates on super-superfast computers be a significant addition to achieving "break-even"? (I'm talking several orders of magnitude faster than today).

This was the notion of a friend of mine: the success of fusion could depend on subtle changes in the design of the reactor. Building and testing each design seems to take decades, whereas a significantly fast computer could simulate any design in ... I really don't know... My friend was saying "in seconds," but I know the nature of the reaction is so complicated, and the amount of matter involves absurdly large numbers of particles acting in quantum behavior... my mind was boggling just considering it.

But then I read that computers had already surpassed etaflops, and it's not egregious to think that well be a billion times faster in 10 to 20 years, plus with the headway into "quantum computers" I'm starting to think, "could be".

Keep in mind that I'm planning an end of year inspiration speech to my physics students. Fusion is not the only topic, but it's important, and I don't want to feed them BS.

If I were to try to persuade some of my students to go into research in fusion, would that be a mistake?

 

[Astronuc]

Thanks for all the insight so far. It seems that there is a cynical attitude toward the success of fusion energy, is that fair to say? And the difficulty is almost entirely technological, as opposed to theoretical, yes?

As for success, when I first considered fusion in the 1970's, it's development was 10 years from today (today then in the 1970's). In 1985, I was teaching a course in fusion and had taken a grad course in fusion as well - it was still 10 years away. Even in the 1990's, it was still 10 years, maybe 20 tops. Then in the late 90's (probably during Clinton's administration), I started hearing more like 50 years. I know a funding manager in Washington who thinks plasma physicists/engineers are liars - based on the contined requests for funding with little progress realized.

The challenges of magnetically confined fusion should not be underestimated or understated, and that perhaps has been the problem all along. It is not as simple as replicating the process in stars. Stars have much greater plasma densities and core temperatures, which produce pressures well beyond what we can achieve by man-made objects. One key constraint is the maximum strength of a magnetic field, which is constrained by the superconducting materials ability to handle the field without breaking down. Then there is what to do with that energy in an emergency dump. One certainly doesn't want a disruptive (exploding) magnetic/structure.

Again, would computer simulates on super-superfast computers be a significant addition to achieving "break-even"? (I'm talking several orders of magnitude faster than today).

That's already being done with supercomputers and massively parallel systems/clusters.

Experiment and numerical simulation go hand in hand. One does an experiment then builds a model (complex system of non-linear partial differential equations) based on the knowledge of the inputs/outputs and state variables of the experiment. Then the model gets tweaked to agree with the experiment. Then one does perturbations and/or extrapolations to new statepoints (different compositions, higher energy, higher temperature, higher density, and rates of change of these paramenters, and combinations of key variables/states and rates of change, and then different or 'new and/or improved' physical models). The progression of experiment and modeling takes time.

Honestly, we just don't know if and when controlled fusion will be successful. ITER may demonstrate that it is, or perhaps another concept will be revisited.

If the students like a challenge, then they could tackle fusion, but at the same time, I would recommend getting a diverse background in physics or engineering, that leaves open as many other paths as possible.

 

[Morbius]

Again, would computer simulates on super-superfast computers be a significant addition to achieving "break-even"? (I'm talking several orders of magnitude faster than today).

Chi Meson,

As Astronuc points out; that is already being done.

About a decade ago; the Dept of Energy instituted the Accelerated Strategic Computing
Initiative or ASCI; in order to boost development of a new generation of very powerful
supercomputers and simulation software in support of a new paradigm for the management
of nuclear weapons without testing; the "Science-based Stockpile Stewardship Program";
[now called just "Stockpile Stewardship Program".]

ASCI has led to "quantum leap" in the power of supercomputers compared to what was
available a decade ago. Some of the most powerful computers are at the national labs
that do fusion research. [ ASCI is now a mature program, not an initiative, and has been
renamed ASC. ]

THE most powerful supercomputer, according to the Top 500 list that tracks this:

http://www.top500.org/

is BlueGene/L at Lawrence Livermore with 131,072 processors:

http://www.llnl.gov/asc/computing_resources/bluegenel/

Sharing the same machine room with BlueGene/L is "ASC Purple" which is #4 on the list:

http://www.llnl.gov/asc/computing_resources/purple/purple_index.html

The #2 computer on the list is Sandia National Laboratory's "Red Storm":

http://www.sandia.gov/ASC/redstorm.html

Lawrence Livermore and Sandia are both home to programs involved in nuclear fusion
research; and those programs make estensive use of the computational facilities of
their parent institutions.

http://www.llnl.gov/nif/
http://www.llnl.gov/nif/icf/icf.html

http://www.sandia.gov/pulsedpower/prog_cap/index.html
http://zpinch.sandia.gov/Z/Images/z.jpg

Dr. Gregory Greenman
Physicist

 

[Aero Stud]

I think the same thing about computers. Maybe I don't know anough so I'm just being stupid here, but why on Earth not program and improve the computer simulations. Why spend 10 billion dollars to build this one single option, when you can spend that or a lot less, to keep improving and running computer simulations on that same design and countless others as well as fine-tuning each any way you like. How big is the maximum innacuracy with current computer power ? 10, 20, 30 percent ? Hard to believe it's that much. And if you got at least one best design giving you a stable gain with a positive uncertainty limmit in simulation - then spend the billions on construction. Right now they don't seem to have any idea if this design will do it, so you end up with the most expensive machine on Earth that does nothing except making you pay a large electrical bill. Even scientific projects need to think about time and money occasionally.

 

[Morbius]

I think the same thing about computers. Maybe I don't know anough so I'm just being stupid here, but why on Earth not program and improve the computer simulations. Why spend 10 billion dollars to build this one single option, when you can spend that or a lot less, to keep improving and running computer simulations on that same design and countless others as well as fine-tuning each any way you like.

Aero Stud,

Computer simulations of this area IS my field; I'm a computational physicist!

You can't program and simulate physics that you don't understand.

We DO simulate as much of the physics as we understand; but there's LOTS of
physics that we don't understand. We do experiments like this so that we can
develop computer models. Computer models and experiments go "hand-in-hand".

For example, you want to model a nuclear reaction in your computer model. In order
to do that, you need the reaction cross-section. Where do you get that?

You get the reaction cross-section from experiments. We don't know all the nuclear
physics that goes on here, ab initio.

Dr. Gregory Greenman
Physicist

 

[Aero Stud]

I see. I did not know that after reaching much higher energies in the various accelerators and building smaller tokamaks and astronimical observations, there was still such a big uncertainty about the nuclear reactions expected inside this thing. Well, hope it works then or at least helps.

 

[Morbius]

I see. I did not know that after reaching much higher energies in the various accelerators and building smaller tokamaks and astronimical observations, there was still such a big uncertainty about the nuclear reactions expected inside this thing.

Aero Stud,

Reaching high temperatures doesn't solve all your problems.

Let's take neutron cross-sections as an example.

For example, consider EXTREMELY high temperatures up in the MeV region for
some high Z materials. You'd like to know the reaction cross-sections for neutrons
on this material.

There's no problem getting high energy neutrons to toss at the material. However,
the nuclear cross-section has what are called resonances; big spikes in the quantity
due to quantum mechanical effects.

At very high energies; those resonances are so close together that you can't resolve
them. It's called the "unresolved resonance regime".

Additionally, there are collective hydrodynamic effects. Practically every new fusion
device finds a new plasma instability that we didn't know about.

So computer simulation without experiment to validate the computer models means
that you are just "spinning your wheels".

Dr. Gregory Greenman
Physicist