News for hackers. Stuff that matters.
Re:A faster, cheaper path to fusion energy by noetic_techy on Monday November 05, 2018 @05:34PM (#18386608)
I've seen the MIT lecture that presented this. Here are the main hurdles:

1) Estimated $40 billion USD need to build a test reactor.

2) Not enough FLiBe fluid (Low-Z fluid) on the planet for the reaction chamber. Would need large scale manufacturing.

Source: https://youtu.be/KkpqA8yG9T4

Re:A faster, cheaper path to fusion energy by baking on Monday November 05, 2018 @11:10PM (#18388290)
The large scale project using low temperature superconductors is ITER, initially proposed at $5 billion and currently estimated at $20 billion for a scientific test reactor.

The MIT ARC design is half the size using high temperature superconductors and should be in the $1-2 billion range for a full-scale 500MW pilot production fusion reactor.

What is currently being proposed (the subject of the papers being delivered tomorrow) and funded with private money is a $200-250 million SPARC (Smallest Possible ARC) that is half the size again of the full-size ARC.

So yeah, you are off by a factor of 160.

Re:A faster, cheaper path to fusion energy by pontifier on Tuesday November 06, 2018 @04:30AM (#18389519)
I think I could build my prototype fusion device for under $1M

Maybe if I said it would cost more to build, people would think it more likely to work.

Re:A faster, cheaper path to fusion energy by FiatLuxDave on Tuesday November 06, 2018 @09:29AM (#18391252)
That may be possible. I built my fusion prototype for about $600k in 1990s dollars.

Being more expensive doesn't make people think it is more likely to work. The easiest way to make people think your design will work is to take an existing design with known parameters and make it 'better'. Unfortunately, 'better' often means bigger and more expensive, which is how we get to projects like ITER. SPARC is a version of a compact tokamak using advances in high temperature superconductors - thus, 'better'.

A different way is to propose something completely different than what has been done before. Then, if it is cheap enough, people may be willing to take a chance on it. This is what I did. However, it needs to be different enough to not be easily dismissable as having the same problems as an earlier design. For example, if you are doing a mirror design, you better have a good answer for fixing the ends.

One thing that often gets overlooked when talking about cheaper fusion designs is speed. Big projects like ITER take a long time to build, and then a lot of experiments get done on them, taking more years, because you need to get your money's worth from them. Smaller designs get the knowledge gains out of the design much faster. Robert Hirsch has been trying to bring this to our collective attention ever since he converted from a tokamak proponent to opponent decades ago.

Re:A faster, cheaper path to fusion energy by baking on Tuesday November 06, 2018 @11:12AM (#18392240)
I think this talk (https://youtu.be/L0KuAx1COEk) makes a pretty good case for why a project of the scale of SPARC makes sense from an engineering standpoint.
Re:A faster, cheaper path to fusion energy by pontifier on Wednesday November 07, 2018 @12:49AM (#18397335)
Very interesting video...

My answers to the questions in the video are:

Fuel: D-D

Temperature target: 15kev

Confinement: hopefully approaching ideal at small ion counts. (penning traps are very good at trapping particles for long periods of time)

Instabilities: unknown but probable at high ion counts. I know there will be challenges to solve, but hopefully they are not fatal problems.

I believe the first prototype could achieve Q>1

Re:A faster, cheaper path to fusion energy by baking on Tuesday November 06, 2018 @08:56AM (#18390954)
SPARC is trying for Q (power out/power in) in the 1-5 range. They see this as critical for getting future funding for ARC. If you think you can get net power, what are you waiting for?
Re:A faster, cheaper path to fusion energy by pontifier on Tuesday November 06, 2018 @02:08PM (#18393987)
The prototype would almost certainly be break even at very low powers and densities. There are very few places where my design would lose energy, and basically gives ions unlimited chances to fuse once accelerated.

I had many discussions with the head of the physics department at the local university who is a specialist in plasma physics. He was of the opinion that it would work at very low pressures/densities, but that the total output would not be high enough to be commercially viable. What is the actual benefit of a room sized $100k fusion reactor that can only put out 100 watts of energy?

I believe I've solved many of those problems, but building a prototype costs $$ that I don't have to spare...

I could bankrupt myself to build the prototype but I can't gamble like that with assets that currently feed my family.

Though I strongly believe it will work, I'm not going to go down the same hole that Tesla went down. I have no intention of dying penniless surrounded by pigeons. I am on track to build the prototype myself without going into debt in several years once a few of my mortgages are paid off.

Right now I own 100% of the IP, and have a submitted patent. I'm not actively looking for funding that would dilute me unless it's a very good deal. I'm kind of stuck bootstrapping unless I can get a grant. I have applied for several in the past, but to no effect.

Re:A faster, cheaper path to fusion energy by noetic_techy on Tuesday November 06, 2018 @05:11PM (#18395353)
I stand corrected. You're almost right, I should have said 4 billion, I remembered the number as 4-something. You can see at 1:16 in, the ARC pilot says 3,000? M$. Without going through the whole video again, I'm almost certain someone during the discussion mentioned 4 billion. You're right thought, a proof of concept can be done for an order of magnitude less.
Re:A faster, cheaper path to fusion energy by baking on Tuesday November 06, 2018 @09:35AM (#18391305)
I should clarify that by "half the size" I meant half the radius, which translates to about 1/8 the cost.
Re:A faster, cheaper path to fusion energy by pfdietz on Monday November 05, 2018 @05:44PM (#18386691)
World annual production of Be is just 220 tonnes.

A single ARC reactor would use 40% of this.

The total world Be resource is estimated at 100,000 tonnes, and if fully used in ARC reactors would supply just 1% of the world's primary energy demand.

Re:A faster, cheaper path to fusion energy by philipkglass on Monday November 05, 2018 @07:00PM (#18387111)
In mining parlance a resource is a function of economic demand as well as geology. Elements with limited demand may have "resource" levels far lower than those in regular industrial use, even if ore bodies of comparable grades actually exist for both.

Uranium was produced at a level of only hundreds of tonnes per year before the development of nuclear weapons and reactors. Now it is produced in quantities of tens of thousands of tonnes per year. Obviously the terrestrial geology of uranium did not change quickly; industrial demand is what changed. Uranium's crustal abundance is comparable to that of beryllium.

Despite cumulative production of more than 2.2 million tonnes uranium through 2003, additions to resource totals have kept pace with production so that overall resource levels have remained level or have increased over time. The ratio between Known Conventional Resources and reactor-related uranium requirements in 2003 was 52 compared to an average of 47 since 1985.

...

Uranium production in 1945 is estimated to have totaled 507 tonnes uranium. By 1965, when the first Red Book was published, production totaled 31,564 tonnes. Production peaked in 1980 at 69,692 tonnes from 22 countries. In 2003, uranium production was reported by 19 countries with output totaling 35,492 tonnes. Cumulative worldwide uranium production between 1945 and 2003 totaled 2,204,732 tonnes...

"Forty Years of Uranium Resources, Production and Demand in Perspective: The Red Book Retrospective"

https://www.oecd-nea.org/ndd/pubs/2006/6096-40-years-uranium...

It's a fair point that beryllium exploration and extraction would have to increase tremendously for routine construction of these reactors. But putting 100,000 tonnes of beryllium in ARC reactors would not mean that the Earth has then run out of beryllium.

Re:A faster, cheaper path to fusion energy by pfdietz on Monday November 05, 2018 @10:13PM (#18388031)
However, expanding the total resource of Be by a factor of 100 would likely mean a large increase in the price of the element.
Re:A faster, cheaper path to fusion energy by starbeast on Tuesday November 06, 2018 @12:41AM (#18388671)
Wouldn't economies of scale push in the opposite direction?
Re:A faster, cheaper path to fusion energy by AnimalMuppet on Tuesday November 06, 2018 @05:24AM (#18389769)
No. At price X, there are Y tons that are economically extractable; at price 2X there are more than Y tons that are economically extractable. If you want more than Y tons of it, the price is going to have to be more than X.
Re:A faster, cheaper path to fusion energy by VBprogrammer on Tuesday November 06, 2018 @05:53AM (#18389886)
I'm certainly no expert in the matter but couldn't this be constrained in other dimensions rather than availability of the source material. For example capital requirements of whatever extraction / purification steps are required vs the global demand. Therefore increasing demand could cause investment which increases supply?

That said, if Fusion power ever becomes a reality I hope it doesn't depend on some difficult to source material.

Re:A faster, cheaper path to fusion energy by starbeast on Tuesday November 06, 2018 @05:48AM (#18389865)
If there is low demand for something, then there is low production and the production costs are high per unit. High demand will at first push the price up, but if there is ability to scale production of the resource, the price per unit should then come down, as the production costs per unit will drop due to economies of scale.
Re:A faster, cheaper path to fusion energy by AnimalMuppet on Tuesday November 06, 2018 @05:53AM (#18389889)
True, but in resource production, that rarely is what dominates. The problem is that, as you try to produce more, you're having to process ore that has lower and lower concentrations of whatever you're trying to extract. You can get a more efficient facility if you have the volume, but the lower concentration (usually) dominates, and so the cost goes up rather than down.
Re:A faster, cheaper path to fusion energy by starbeast on Tuesday November 06, 2018 @06:23AM (#18390018)
I don't think your analysis here is right.

If you look at the history of the price for aluminium, it has steadily dropped as production has risen, which is the opposite of what you are saying. And to pick another metal, the lowest copper price of the last 100 years was around the year 2000.

Re:A faster, cheaper path to fusion energy by philipkglass on Tuesday November 06, 2018 @12:57PM (#18393291)
I upvoted you, pfdietz, and AnimalMuppet because you are all correctly describing different phenomena that can increase and decrease the prices of a mineral commodity. Increasing demand pushes up the price, which spurs searches for additional resources and improved extraction techniques. Improved techniques and the discovery of more resources can push the price down again, sometimes even below the price before major industrial demand started. Uranium is cheaper in the 21st century than prior to its large scale industrial use. Platinum remained more expensive after demand shot up from catalysts, despite huge investments in improved extraction and the discovery of new resources.

I'm not trying to make predictions about the price movement of beryllium following a very successful ARC demonstration. I'm just pointing out that you can't look at current resource levels and divide by annual production rates to learn when resources will fall to 0. Making predictions about future prices is particularly hard when an element currently has only niche uses (e.g. beryllium, thorium, thallium, rubidium).

Re:A faster, cheaper path to fusion energy by jabl on Tuesday November 06, 2018 @02:01AM (#18388969)
Couldn't one use, say, FLiNaK instead of FLiBe?

Sodium and potassium are pretty abundant.

That of course leaves the problem of needing relatively large amounts of 6Li, but that's something one has to live with as long as one wants DT fusion, I guess. And it seems this is more with building up the enrichment capacity rather than some fundamental geological limits.

Re:A faster, cheaper path to fusion energy by pfdietz on Tuesday November 06, 2018 @05:19PM (#18395426)
Beryllium is needed for neutron multiplication. The 5th neutron in 9Be is the most weakly bound neutron in any stable isotope, with a binding energy of just 1.6 MeV.

It might be possible to use lead for neutron multiplication (it has a nice high cross section for (n,2n) reactions above 8 MeV), but it would not moderate the neutrons much, and 6Li breeds best off thermal neutrons. So the blanket and reactor would have to be bigger.

Re:A faster, cheaper path to fusion energy by JamesCoyne on Monday November 05, 2018 @06:20PM (#18386907)
Tossing in two cents to say the video you linked is a very good overview of the different experimental approaches to fusion. Presented by Dennis Whyte of MIT, it obviously extols the virtues of SPARC device.
Re:A faster, cheaper path to fusion energy by Demeisen on Monday November 05, 2018 @09:00PM (#18387745)
Another excellent talk, about a year later: https://youtu.be/L0KuAx1COEk
Re:A faster, cheaper path to fusion energy by gameswithgo on Monday November 05, 2018 @05:43PM (#18386685)
How hard is it to make FLiBe?
Re:A faster, cheaper path to fusion energy by pfdietz on Monday November 05, 2018 @05:46PM (#18386707)
The enriched 6Li in a single reactor would correspond to 1/4 of that produced for the entire US hydrogen bomb program. There is no facility in the world that could make the 6Li required, and the technology that was used in the US is now prohibited due to large leakag of elemental mercury into the environment (the plant used thousands of tons of liquid mercury metal.)
Re:A faster, cheaper path to fusion energy by MertsA on Monday November 05, 2018 @11:51PM (#18388462)
As a neat coincidence, molten salt fueled fission reactors also need enriched Lithium. However, those need Lithium 7 which would essentially be the depleted Lithium from enriching Lithium 6.
Re:A faster, cheaper path to fusion energy by pfdietz on Tuesday November 06, 2018 @12:28AM (#18388623)
No, MSRs do not necessarily need enriched lithium. For example, Moltex's fast MSR uses no lithium whatsoever. The (barren) coolant salt is zirconium sodium potassium fluoride, and the fuel salt (in tubes, where it does not mix with the coolant) is a mixture of sodium, zirconium, uranium, and plutonium chlorides, as well as fission products.

The zirconium does not have to be "nuclear grade", since the design not only tolerates hafnium (a strong thermal neutron absorber) in the coolant salt, it depends on it to shield the reactor structure from thermal neutron degradation.

Lithium-7 will still lead to production of some tritium. Any MSR using lithium is going to need a tritium separation and capture system.

Re:A faster, cheaper path to fusion energy by baking on Tuesday November 06, 2018 @09:17AM (#18391139)
SPARC's models show a net positive for tritium breeding of 1.08 (tritium produced/tritium consumed) but I'm pretty sure this relies on enriched lithium, although I can't find where they specify the level of enrichment for that model. I'm sure it was a design constraint though.
Re:A faster, cheaper path to fusion energy by MertsA on Wednesday November 07, 2018 @05:48PM (#18403035)
MSRs that use lithium in the fuel salt do in fact need enriched lithium. The problem is that lithium 6 has a much larger neutron absorption cross section than lithium 7. It's on the order of 10,000x greater IIRC.
Re:A faster, cheaper path to fusion energy by pfdietz on Wednesday November 07, 2018 @07:00PM (#18403353)
Right. It's just that MSRs don't necessarily need lithium.
Re:A faster, cheaper path to fusion energy by MertsA on Thursday November 08, 2018 @03:33PM (#18409725)
Not all of them do, you're right about that.
Re:A faster, cheaper path to fusion energy by 08-15 on Tuesday November 06, 2018 @05:51AM (#18389877)
Why do you think it needs to be enriched?
Re:A faster, cheaper path to fusion energy by philipkglass on Tuesday November 06, 2018 @01:04PM (#18393351)
The ARC paper proposes that the lithium in the salt should be enriched to 90% lithium 6 to reach a tritium breeding ratio of 1.1. See section 5.3, "Tritium breeding."

https://arxiv.org/pdf/1409.3540.pdf

I presume that the breeding ratio would fall below 1.0 with natural lithium.

Re:A faster, cheaper path to fusion energy by 08-15 on Tuesday November 06, 2018 @04:51PM (#18395197)
For the first proof-of-concept reactor? That's borrowing trouble. If the things works at all and the only remaining problem is the tritium supply, funding won't be a problem anymore.
Re:A faster, cheaper path to fusion energy by baking on Wednesday November 07, 2018 @12:50AM (#18397344)
To clarify, the design that the paper is addressing is the ARC, a full-size pilot reactor intended to put 200MW out on the grid for 9 working years.

The demonstration reactor is the SPARC which is based on a scaled down half-sized (but 1/8 the mass and 1/8 the cost) version of ARC. While the main goal of SPARC is to achieve significant net power output, along with a proof-of-concept for many of the design features of ARC, some features may have to be left out for lack of space.

So I don't know if SPARC will be able to achieve net tritium breeding, but I presume they will be testing as many of the features as they can along with taking measurements and verifying the model. It is also noteworthy that the ARC/SPARC design allows for replacement of the vacuum vessel and cooling blanket without complete removal of the outside magnetic coils, so they foresee design iteration of those components unlike the ITER design which will be pretty much locked in.

Re:A faster, cheaper path to fusion energy by baking on Wednesday November 07, 2018 @02:02PM (#18401311)
For example, the blanket needs to be about a meter thick to capture the neutron so a smaller scale design will be inherently less efficient at breeding tritium.
Re:A faster, cheaper path to fusion energy by pfdietz on Tuesday November 06, 2018 @05:27PM (#18395481)
They have to breed tritium if they are to run the reactor for any significant length of time. Buying tritium externally would be very expensive -- it's $100M/kg and up, particularly if you exhaust sources like incidental production in commercial heavy water reactors.

DT fusion has the nasty circular dependency that breeding blankets are needed to make tritium, but they cannot be tested without working high intensity DT fusion neutron sources.

Re:A faster, cheaper path to fusion energy by agumonkey on Tuesday November 06, 2018 @07:14AM (#18390283)
So it's mainly theoretical ?