back to article Amazing new algorithm makes fusion power slightly less incredibly inefficient

Google and Tri Alpha Energy, a Californian energy company, say they have come up with an algorithm that appears to help scientists generate hotter plasma more efficiently for nuclear fusion experiments. Keyword: experiments. Billions of dollars have been poured into achieving clean, carbon-free energy harnessed from smashing …

Optometrist Algorithm

Is this better, or worse?

<makes adjustment>

Better, or worse?

Which is clearer, 1?

<flips lens>

Or 2?

< flips back >

1?

<flips again>

Or 2?

32
0
Silver badge
Terminator

Re: Optometrist Algorithm

Nah, this is computerized. So it's more like:

Which is better?

0?

<flip>

or 1?

<flip>

0?

<flip>

or 1?

23
0
Silver badge

Re: Optometrist Algorithm

I've never figured out whether I want to tell him "better" is more accurately formed letters or higher contrast letters. Sometimes the choice is between a perfectly formed 'E' that's got medium level contrast, and an 'E' that has a slight aberration that's got very high contrast. Which is better for actual vision?

19
0
Silver badge

Re: Optometrist Algorithm

" Which is better for actual vision?" Doug's

I've a feeling your optometrist is asking you which you prefer

11
0
Gold badge
Boffin

Re: Optometrist Algorithm

I've never figured out whether I want to tell him "better" is more accurately formed letters or higher contrast letters.

Tell them what you're seeing. So they can choose the next lens to try with more information.

Admittedly things are rather different for me. When you've got about 5% of normal vision the 5%-odd inaccuracy of the test is more-or-less equal to what you're trying to measure.

So the correct way to test vision this rubbish is apparently to work out the prescription for the lenses, and work backwards from that. The difference is between not being able to read the top letter on the chart, or maybe being able to read it with glasses on. So admittedly my answer above is given based on the fact that I'm dealing with people who're probably better trained and certainly have more experience (as they work in an eye hospital).

The other difference to having your eye test done in a hospital is they have these weird, massive, metal chairs. Bolted to the ground and with foot-rests. Which look like they've probably got mountings, so they can strap you in. And the room is filled with all sorts of gear that looks like torture equipment. Especially the thing with the chin-rest and massive lights they shine in your eyes and the probe that has to touch your eyeball in order to do the pressure test.

Is it safe. Is it safe...

3
0
Silver badge

Re: Optometrist Algorithm

My optometrist has all that stuff you describe and he's just in a normal office, not in a hospital. Except that several years ago he replaced the probe that touches your eyeball with something that does the pressure test by blowing a puff of air into your eye. Still a bit annoying, but at least it means you no longer need the numbing drops!

2
0
Anonymous Coward

Re: Optometrist Algorithm

Is this better, or worse?

I asked several friends about this, and apparently, I've been going to the wrong sort of optometrist.

Mine sticks sharp wooden needles in my eye and asks if if hurts more or less than the previous.one.

I think I'd best go to a different one.

2
0

Re: Optometrist Algorithm

You are describing the epitome of research. Shame on you. Sticking pins in stuff to see what happens is how scientists all over the globe function. Sadly, most of this is a complete failure, and the pins are never recycled. But, they do use acres of paper made from trees to publicise their efforts. My own research area is how to recycle old scientists. Despite their undoubted scientific skills, most couldn't turn a sausage on barbecue if they tried. I have this theory ( yet to be published) that scientists would serve us better by helping chefs to cook stuff normally, and to avoid using foreign words on menus. :-)

1
0
Gold badge
Happy

Re: Optometrist Algorithm

I thought your sentence was going to end "scientists would serve us better as casserole".

But it turns out you decided they should teach us to cook. I'm happy to have a replacement for kitchen-french, but only if they don't replace it with kitchen-maths.

I don't want to have boeuf bourguignon replaced with:

beef³+seared@523.15K+boiled@373.15K+wine

0
0
Bronze badge

Re: Optometrist Algorithm

How backwards is your optometrist? I've been wearing glasses for nearly 40 years and I can't recall any other method but the air puff test.

1
0
Mushroom

Rule 0

Do not make a fusion explosion.

11
3
Silver badge

Re: Rule 0

That's the easy one to keep! Though you'd be amazed at how irrational some people can be. I remember there being anti-nuclear protests outside the JET project at Culham, despite the physical impossibility of the machine doing anything dangerous (in the nuclear sense).

20
0
Anonymous Coward

Re: Rule 0

> I remember there being anti-nuclear protests outside the JET project at Culham, despite the physical impossibility of the machine doing anything dangerous (in the nuclear sense)

Mind you, any research which increases our knowledge of nuclear fusion processes is going to be gladly accepted by the people who make bombs.

0
2
Silver badge

Re: Rule 0

I think we know pretty well by now how to blow all the energy in the first few milliseconds...

it's the not blowing it in a few milliseconds that we have no idea how to do so far.

4
0
Silver badge

Re: Rule 0

Ah, but Bazza, that was in the previous universe. This one was created from the fireball generated by the "safe" experiment, and all the dimensions in which fusion works easily were rolled-up, leaving us with ony the previously-called "useless four".

1
0
Silver badge

Re: Rule 0 4 Simon Harris

Given the estimated fusion yields from warhead tests, I'd suggest we actually know very little about getting "all" the energy to blow in the first few milliseconds. That's why the military wants all those hypercomputer cycles for sims.

0
0
vir
Bronze badge

A Few Observations

This sounds like a species of genetic algorithm with operator input.

The company's name is Triple Alpha, but they're developing proton-boron fusion instead?

Could this be a viable source of helium? Proton-boron fusion creates 3 atoms of helium and 8.7 MeV of assorted energy. Assuming a 1 GW total energetic output from fusion, a plant would create 3592 mol/s, which is a hair under 900 grams of helium per second. A few plants could conceivably meet the total helium needs of the entire world. This seems like a lot of helium output, but the reaction is about 1/20 as energetic as fission, so maybe I didn't screw up the math. Anyone care to check?

4
1
Silver badge
Stop

Re: A Few Observations

There's no helium shortage*. What's more, your suggested manufacturing method is a very expensive way to get hold of some. Even this might be better...

http://thunderbirds.wikia.com/wiki/Sun_Probe/Episode_Guide

*https://www.wired.com/2016/06/dire-helium-shortage-vastly-inflated/

Back from when The Register was good:-

"stop us all worrying about running out of resources"

https://www.theregister.co.uk/2015/06/22/time_worstall_register_summer_lecture/

Bring back 'Time Worstal'!

6
1

Re: A Few Observations

According to Wikiwhatnot this company are pursuing the proton-boron fusion reaction precisely because this releases all its energy into the three alpha particles produced. No tricky neutron energy to absorb/convert. Trouble is it needs approx 10x the plasma pressure of D-T fusion.

Nice to see this example of machine learning even if it is human-in-the-loop.

3
0

Re: A Few Observations

You know darn well that Al Gore will demand Helium sequestration because we all KNOW that helium is the prime cause of climate disruption, which messes with old Al having issues keeping his swimming pools nice and warm in January.

0
1
Silver badge

Back in the day

When I was a lad at a school in Oxfordshire (*) in the eighties we went on a field trip to the local fusion reactor experiment. I seem to recall that they talked about orders of magnitude out from sustainable fusion rather than a factor of two. I remember seeing a graph of progress with a logarithmic Y axis and a fair way to go on the X axis.

They talked about "50 years away" (or was it 25 - that X axis was hard to interpret!) and I note that reasonably current articles in New Scientist still mention similar timescales. It will happen one day but it is rather expensive and rather hard to get politicians to sign off the vast sums needed for a very, very long slog.

(*) WTF Google - that's how its spelt - what's with the squiggly underscore?

7
0
Silver badge
Boffin

Re: Back in the day

The article talked about an improvement in the temperature and efficiency over current methods by a factor of 2.

You can go online and find YouTube vids on MIT research in to Fusion and their take on R&D outside of the large scale Tokamacs ?sp? where they want to use the plasma shape to help with its efficiency.

The issue is to raise the temps and the longevity of the 'burn'.

And yes... its still just 50 years away.

I blame Bush, Obama and Congress over the past 12+ years for killing high energy funding.

The interesting thing is the super conducting materials used to help create the magnetic fields has advanced. Kinda cool. So you should check out the vids.

6
0
Silver badge

Re: Back in the day

I was under the same impression - that they're several orders of magnitude away from the conditions required to "break even"; sure, any help is good, but this doesn't seem to even begin to make a dent in the problem...

1
0
Bronze badge

Re: Back in the day

"this doesn't seem to even begin to make a dent in the problem"

I quite agree - all this really consists of is a virtual board of knobs - the operator can then twiddle them one by one, (then all over again) until they get the best possible results from the current method - it doesn't directly alter the method at all, and unless they take some time to analyse the _reason_ why these settings are optimum, then they will have learned nothing.

But at least they are trying.

1
0
Silver badge
Boffin

Here are some free ideas

To the research scientists trying to understand fusion (instead of building reactors that make electricity).

Here's some free ideas. You're welcome.

1. start acting like this is a problem that needs to be SOLVED, instead of UNDERSTOOD

2. 1 word: RESONANCE [if you're a nuke scientist you know what this means]

3. Study how a 'travelling wave tube' works when it's creating or amplifying microwave frequencies, particularly with respect to "electron bunching" (and RESONANCE). Applications obvious.

4. Consider magnetic lenses and magnetic compression, not a toroid

5. There's an existing design that seeks to eliminate the effects of a torus causing the inside track to be slower than the outside track, something that was once pursued by the U.S. Navy, but somewhat recently abandoned [probably for something a whole lot better that's classified]

6. You're going to have to get energy out of the reaction at some point. Have you figured out how to do this? I suggest making the reaction happen inside a cavity within a large tank of water [aka 'boiler']. Steam systems would then attach to it. the rest is kind of obvious.

7. superconducting magnets lose their magnetism and/or superconducting properties under high doses of gamma radiation. don't use them. Pulsing electromagnets would have other benefits as well. Try those.

8. electrostatic focusing has been used for DECADES with various kinds of vacuum tubes/valves. Resurrect that tech, except using it for protons/deuterons/tritions etc.

9. deuterium and tritium are NOT the only fusion fuels. How about firing high energy protons at a solid lithium target? Or maybe lithium at lithium? Whatever, just try things. Inertial confinement seems to work kinda well with a solid fuel pellet, though it takes too much energy to burn the fuel. How about if that pellet crashes into "other fuel" ?

anyway, just thought I'd mention things like this in a different venue. who knows, it might work.

4
26
Silver badge

Re: Here are some free ideas

You need to check out some of the research by MIT.

There are designs that make using newer super conducting magnets viable. And newer materials.

Tokamack ?Sp? is based on toroid design but some of the incremental research is using linear systems.

More funding is needed and leaps in R&D and material science.

5
0
Bronze badge

Re: Here are some free ideas

just need to remove as much latency as you can from magnets to sensors to make fusion reactors alot more efficient, it needs to be in nano seconds, 2ms is a year to the life of plasma, so fibre and a quantum computer will save the day soon

uk's tokamack would have the upper hand by being new, with modern PCB design's etc

0
3
Silver badge

Re: Here are some free ideas

6. You're going to have to get energy out of the reaction at some point. Have you figured out how to do this? I suggest making the reaction happen inside a cavity within a large tank of water [aka 'boiler']. Steam systems would then attach to it. the rest is kind of obvious.

It still surprises me that nuclear power generation still relies on steam technology. Has there been any research on improved methods for converting energy into electricity? We got ride of steam trains for a reason.

5
7
Silver badge

Re: Here are some free ideas

There are ideas for helium cooled fission reactors that use the hot helium directly in power turbines. This misses out the heat exchanger commonly found in fission reactor plants. The designs are quite clever - self regulating graphite pebble bed reactors could be entirely passively controlled. About the he only perceived difficulty is how to make the graphite pebbles not fall to pieces as a result of the neutron flux running through them. The design relies on cycling pebbles through the reactor and out the he other end when done with, but if one breaks then it's a big mess.

There's also ideas for direct working gas cooling of Americium reactors. You need far less material, and the Americium can be in direct physical contact with the working gas (it doesn't need to be a powdered oxide fuel like is needed for Uranium/Plutonium fuels).

6
0
Silver badge
Boffin

Re: Here are some free ideas

steam tech is pretty good for large amounts of power. all heat engines reject heat, and solar panels are no exception. I think their efficiency is in the 20-25% range nowadays, which is about like a 1000psi steam plant, i.e. your basic power plant. It's all 2nd law of thermodynamics and so on.

Some people like the idea of using air turbines instead, but it's harder to get the air flow to work that way. Steam naturally goes from liquid to gas and run down the pipes, then you condense it [in a vacuum of course] and pump back into the boiler. If you can find a liquid that works BETTER than water for that, well maybe you can improve the process a bit, but it's a tried+true tech and works exceedingly well.

The only other things I know of that convert heat or radiation directly into electricity are Peltier devices, and they'd probably reject even MORE heat than a steam plant [and be less efficient].

But yeah a linear reactor design - that's what I'd suggest, and I'm glad MIT is looking into it. It's something that's more likely to get you a sustained reaction and enough power to power itself.

and of course, some kind of "impulse drive". to which you'd have to add water. because momentum is mv, and you double mass flow rate to double thrust (which is twice the energy), or you quadruple energy to double velocity to double thrust. So you'd have to add mass for an impulse engine to work. So you'd still need fuel, but it could be something really heavy instead. [I figure best fusion rocket design would squirt water on the inside surface of the engine and into the center of it, to boil off and prevent engine melting, and provide the extra thrust from the mass - the fusion would provide the energy for velocity and therefore delta momentum = thrust]

but yeah, screwing around with Tokomak isn't getting anybody anyplace, but consumes a LOT of research money.

4
3
Silver badge
Devil

Re: Here are some free ideas

"This misses out the heat exchanger commonly found in fission reactor plants"

actually, the differential temperature across the tubes in a typical nuke plant boiler isn't all that much, maybe 40 deg F or so. The main limitation is the steam pressure. The Molier diagram for water more or less outlines the physical properties, and if you operate around 1000psi or so, you get maximum benefit. Most nuke plants will use oil-fired steam superheaters also, from what I understand, most likely to prevent condensation on the way to the turbines. Condensation is bad for turbines. It tends to damage the blades.

Anyway, all of this has been taken into account, more or less, which is why pressurized light water reactors are more common. Other designs exist, but these are generally the safest [unless something stupid happens, like no emergency cooling for several days (Fukushima), or the relief valve sticks open on the pressurizer and nobody notices because the indicator light is off and they're paying attention to pressurizer level, which is going up because of the pressure drop, so they DUMP COOLANT to lower it [this happened at 3 mile island].

Additionally the physics of water as a moderator is actually very good, much better than helium or heavy water would be. Still you can moderate with liquid sodium, and other materials, as long as the reactor design allows for it. Yet most reactors still use pressurized water. Must be something 'right' about it.

anyway... I used to operate a nuke plant for the U.S. Navy on a submarine a couple of decades ago. So I have some experience with it, though it's been "that long".

9
2
Silver badge

Re: Here are some free ideas

Ah yes, subs. I remember an RN sub steam kettler explaining the antiquity of the turbine machinery design, effectively harking back to the days before superheated steam in the early 20th century. Big slow turbines, to deal with the damp steam. Nothing like the small fast turbines found in surface ships circa 1940s.

Re: heat exchangers - sure, they're pretty good, and of course they bring further benefits too in a nuclear plant; the turbine machinery shouldn't (all being well) get contaminated. That was the main attraction of helium as a coolant; it wouldn't be transmuted into anything else and become radiologically nasty. So provided the pebbles remain intact, the turbine machinery would remain clean despite being driven directly by the primary coolant.

I didn't know that civil power reactors often had oil fired superheaters. Seems like cheating somehow!

2
0
Silver badge

Re: Here are some free ideas

I think there is probably a place for a Raspberry Pi somewhere in a Fusion reactor.

Perhaps a little python script to control the magnets that do the suspension of the plasma and then trigger the bombardment of energy, all dependent on sensors feeding back through the IO pins.

Might have a go tonight when I get home. How hard can it be?

5
1
Silver badge
Trollface

Re: Here are some free ideas

"Might have a go tonight when I get home. How hard can it be?"

I was thinking that maybe they could set up some mirrors and direct all that light to some solar cells to reclaim some of the energy that way too :)

1
1
Bronze badge

Re: Here are some free ideas

use huge ceramic TEG peltiers and heat pipes so you can balance out the 1 million degree's through a big network

1
3
Anonymous Coward

Re: Another person puzzled by steam ...

Rather than heat, could we not look to harness electromagnetic directly ?

2
0
Silver badge

Re: Here are some free ideas

Steam loco's went because of fuel, oil products in combustion engines don't need the medium of steam to produce rotational energy. Reactors produce heat that needs to be converted into rotational energy in order to drive the generators, water has all of the properties required to achieve that and is easily managed, is cheap and even it's downsides like corrosion are well understood and can be dealt with relatively easily and economically

3
1
Silver badge

Re: Another person puzzled by steam ...

"Rather than heat, could we not look to harness electromagnetic directly ?"

Yes well see that's a bit of a problem when your main form of energy coming out of the reaction is neutrons moving really fast. Having no electric charge by definition, they're a bit hard to harness any other way than having them slam into stuff aka heating it up (and also making it radioactive, unfortunately). That said, depending on the specific fuel and fusion reaction you use there are options which aren't producing neutrons (surprisingly and inventively named "aneutronic fusion") which are actually being looked into by some researchers, but we're nowhere near reaching appropriate densities / temperatures to make it actually generate more that it takes to keep going. Fusion does happen mind you, they do make it "work", it just isn't outputting as much as it eats. Yet. However, if they can figure this one out, the resulting fast, charged particles could be used to extract electricity directly - although we'd need to also capture some other forms of energy the reaction generates (such as x-rays).

6
0
Silver badge

Re: Another person puzzled by steam ...

Rather than heat, could we not look to harness electromagnetic directly ?

Magnetohydrodynamic power conversion was a popular concept for fusion in the 1960s to 1980s, but MHD power plants (as demonstrated in some experimental coal and gas plants) can be very finicky and not competitive with conventional alternatives. They also complicate the whole process of containing the fusion plasma. You're already juggling thousands of variables to keep the plasma from touching anything but the diverter plates, so to toss in the electromagnetic variables in a MHD power system just complicates the whole mess.

It's not elegant, but it is easier for fusion plants to boil water or heat helium.

2
0
Silver badge
Childcatcher

Re: Here are some free ideas

Re: Water as a primary coolant. No. To drive the turbines, yes.

If you're designing a fission reactor to make energy, rather than one to make plutonium for the military (as most historical plants do) then you want a fast reactor, e.g. https://en.wikipedia.org/wiki/Integral_fast_reactor

These reactors are inherently safe, burn up almost all of the fuel and you can throw in any spare nuclear waste you want shot of. So, you wanna use sodium as the coolant.

https://en.wikipedia.org/wiki/Sodium-cooled_fast_reactor#Sodium_as_a_coolant

There's enough uranium in seawater to keep these plants going until the sun explodes.

http://tmtfree.hd.free.fr/albums/files/TMTisFree/Documents/Energy/Breeder_reactors_A_renewable_energy_source_pad11983cohen.pdf

2
1
Bronze badge

Re: Here are some free ideas

One of the main reasons for getting rid of steam locomotives was their need to carry large/heavy amounts of water around with them, which was then vented to the atmosphere. Stationary engines don't have supply issues, and don't need to vent steam either.

1
0

This post has been deleted by its author

Bronze badge

Re: Here are some free ideas

>We got rid of steam trains for a reason.

And that reason was coal.

2
0
Bronze badge

Re: Here are some free ideas

you can't use magnet's, magnetic friction is what creates plasma

0
2

Re: Here are some free ideas

One reason we got rid of steam trains is that it's hard to build an efficient condenser that uses air as the coolant; the heat transfer is too slow. Water, on the other hand, is quite good at heat transfer, so steam plants that use condensers cooled by water (as in lakes or oceans) are much more efficient. And as I'm sure you know, the efficiency of any heat engine is determined by the difference in temperature between the hot working fluid (steam, ICO a steam plant) and the cold working fluid (condensate).

Fossil-fueled steam-driven ships lasted a lot longer than steam trains, in part because the ocean makes a great coolant for the condenser. I was Main Propulsion Assistant on one of the last ones in the US Navy, decommissioned in 1993. Steam plants were replaced on non-nuclear naval vessels because the darned steam plants were getting too complicated and too dangerous: 1200psi steam at 975 degrees (F) of superheat, and we had around 5000 valves in our plant (four boilers, two main engines). I'm told that the gas turbines that replaced them are much simpler. (Possibly more efficient, I don't know.)

1
0
Silver badge

Re: Here are some free ideas

And that reason was coal.

Also ease of use and ease of maintenance. Diesel trains could be started and shut down in minutes. Steam engines could take hours to make a cold start and were, comparatively speaking, hangar queens. Dieselisation had enormous economic attraction to railways, except for all the steam engine workers that got fired by diesel engines.

0
0
Bronze badge
Boffin

Re: Here are some free ideas

The reaction Tri-Alpha and some other groups (Focus Fusion etc) are aiming for an Aneutronic design, these use pBoron-Hydrogen for the reaction rather than D-T.

The advantage of this is there are no neutrons released from the reaction and minimal neutrons released from subsequent interactions.

The advantage is the reaction produces helium ions, depending on the reactor type these can be focussed.

Therefore you can generate electricity directly from the reaction and the electricity generating equipment effectively becomes solid state.

The problem with this design is it requires much higher temperatures and pressures to 'ignite' the reaction.

0
0

Re: Here are some free ideas

You haven't heard of coal fired steam engines? You can burn coal all day long and not go anywhere, except release a bunch of ash and nasty things into the atmosphere, you have to convert the heat into rotational energy, a Stirling Engine, which are not know for their energy output.

0
1
Silver badge

"Oh look.. it's working well thanks to Google."

"Uh.. seems to be working too well."

"Oh sh......"

Moral: Let's be careful out there and not tweak things too fast or too much.

3
0
Bronze badge

Still not working...Sigh...

Must count.. Yes. When I was 14 I was working on fusion with my father at the Berkeley Rad Lab . We were using a neon filled axial fusion simulator with end mirrors. At the same time I was attending physics lectures and helping with the Standard Model by visually scanning the bubble chamber photos. It quickly became obvious that one cannot balance a repelling permanent magnet on top of another (not spinning).

Most of my work was taking pictures of the very high quality oscilloscopes and keeping accurate notes . Other than that it was very cool to see a 50 foot violet bar of ionized air from a beam dump on the Bevatron as well as watching the walking bubble chamber. Cooking a doughnut in the waveguide in 2 milliseconds was fun. Plenty of LN2 to play with.

I really wonder if this will ever work here on Earth? Gravity is the best container.

5
0

Page:

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Forums

Biting the hand that feeds IT © 1998–2017