Helion Energy to demonstrate net electricity production by 2024

[RERT]

Ok, feels like closing in.

Polaris runs at 0.1 Hz+. If each pulse is like the power plant, it would be net 5MJ, up to 45MJ in and 50MJ out net of losses.

So 500kW is ball-park power if each pulse is as good as the final design. So expectation must be lower.

You mentioned a light-bulb would do, but is there any meaningful estimate around of Polaris power?

[Skipjack]

Ok, feels like closing in.

Polaris runs at 0.1 Hz+. If each pulse is like the power plant, it would be net 5MJ, up to 45MJ in and 50MJ out net of losses.

So 500kW is ball-park power if each pulse is as good as the final design. So expectation must be lower.

You mentioned a light-bulb would do, but is there any meaningful estimate around of Polaris power?

Helion is playing some details close to their chest.
To the best of my understanding, the magnets in Polaris are weaker than those of the final power plant: 15+ T vs 20+ T. Though, they might upgrade them over the lifetime of the machine.

Then there is the question of "With what fuel?"
Polaris will run on D-D, D-T and D-He3.

From what I understand, they will try out D-D first, then D-T before moving on to D-He3.
IMHO (and that is pure speculation on my side):

Net electricity from D-D is unlikely with Polaris, though future power plants could (probably) be self sustaining on D-D, maybe even slightly net positive.
Net electricity from D-T is very likely with Polaris, though how much I don't know. It might even reach ignition with D-T.
Net electricity from D-He3 is the ultimate goal for Polaris, but might take more optimization and maybe some upgrades (magnets).

Either way, from what I understand, Helion will consider Polaris a success if they can power a literal or figurative light bulb with it.

[RERT]

Ok.

Presumably yield is strongly related to B, so it seems quite unlikely that a generator targeting low-Q will produce a remotely similar yield per pulse with the mag field down by a third.

The penny finally drops as to why they recently added DT to their program.

[Skipjack]

Ok.

Presumably yield is strongly related to B, so it seems quite unlikely that a generator targeting low-Q will produce a remotely similar yield per pulse with the mag field down by a third.

The penny finally drops as to why they recently added DT to their program.

It is down by a quarter. At B^4 a future power plant would be about 3 times as powerful as Polaris. Though, as I said, they might update Polaris with stronger magnets later on. That said, I am pretty confident that they can power a lightbulb with D-He3 with Polaris.

[TallDave]

HERT:

So 1Hz, circa 50 MJ in, say 55 MJ out, is 5MW.

yes but unfortunately it's only 5MW for 1ms, which is not very useful, practically speaking, as it would only provide 5KW over the 1 second, or enough to power an average home (although again I have to emphasize this is still 50 years ahead of major programs like ITER and NIF)

Uhm no. 50 MJ is almost 14 kWh. That would be 14 kWh per second. Assuming that half of it is recirculated into the machine, we get 7 kWh per second.
That would be 25 MW, if released to the grid continuously. Though with losses and everything else, it would likely be significantly less. I would already be happy if they managed to power a light bulb.

hmm yes I did get the fusion power confused with the grid output there

but it looks like maybe you confused the total pulse power with the gain

so starting from the gain of 5MJ, if you released that energy over 1 second to the grid, you would send 5MW to the grid over that 1 second

assuming you could produce the 5MJ gain again each second, you could send 5MW to the grid continuously, which is actually pretty impressive!

of course that's the optimal case, getting literally any usable gain from any size pulse will shock the world

at the design spec of .1 Hz we'd be looking at .5MW at a Q of 1.1

but based on Kirtley's paper, Q might be as high as 5 (and he says that's conservative due to the ion heating)

so at an ultra-optimistic Q=5 at 1HZ we'd have a quite astounding 50MJ in, 250MJ out, 200MJ usable power = 200MW to the grid(!)

of course they can't possibly have enough capacitance for that, so let's try again with the 50MJ capacitance as the limiting factor on gain

so they start at 10MJ, full-up the capacitors to 50MJ on each pulse, and send 40MW to the grid in the ultra-optimistic capacitance-constrained scenario

at any rate, if I worked there, I would ask to plug my phone into the capacitors so I could be among the first humans ever to use fusion power

[TallDave]

Ok. Presumably yield is strongly related to B, so it seems quite unlikely that a generator targeting low-Q will produce a remotely similar yield per pulse with the mag field down by a third. The penny finally drops as to why they recently added DT to their program.

It is down by a quarter. At B^4 a future power plant would be about 3 times as powerful as Polaris. Though, as I said, they might update Polaris with stronger magnets later on. That said, I am pretty confident that they can power a lightbulb with D-He3 with Polaris.

at some point though, power gains become much less desirable than frequency/utilization gains (partly due to the neutron/thermal load, but also because no matter how large Q is you still need enough capacitance to hold the gains)

e.g. seems likely a pulse producing the required 1TW gain for 1ms at 1Hz (.1% utilization) may create some intractable physical problems for a 1GW plant, even aside from the 1TJ capacitance requirement

presumably Helion would much rather increase the utilization to (say) 10% (say 5ms pulse every 50ms) and thus only require a 10GW pulse to produce the same 1GW grid output

seems to me the two major stumbling blocks that might kill this concept are 1) whether they can drive >20KeV plasma temps and 2) whether frequency can be high enough for a commercially feasible capacitance/load

[RERT]

Right now I’m willing to trust them that they can make a 50MW plant work. The internet seems to think that capacitors can have really enormous power densities, kind of supporting that.

My perspective is that the low core utilisation is a huge opportunity: if it really is busy for a few percent of the time (at 10Hz), then the core itself can produce far more power if the design can be improved.

Granted that might mean yet more capacitors.

Prosaically, improving the design might mean more cooling and shielding. More interesting is thinking about whether partial/intermittent evacuation and a ‘maturing’ fuel load could be made to work. Most radically, there might be an alternative to pumping to cleanse the chamber.

BUT: focus. The current design is targeting a Kitty-Hawk moment, and that will do - to say the least!

[charliem]

Mixing power and energy is obscuring things in regard to capacitor bank size. I don't see a problem.

Neither pulse frequency, nor peek power production determine capacitor size. Energy needed to get to reach fusion conditions, Q, and energy recovery efficiency do.

With an example:

• Lets say one of their production reactors needs 100 MJ to get to fusion conditions, and gets to a Q of 1.

• That'd mean a total of: 100 MJ injected + 100 MJ produced = 200 MJ.

• Now, let's say they can recover 75%, or 150 MJ.

• Set apart 100 MJ for the next cycle and we end with a net 50 MJ.

• Working at 1 Hz, this machine would give us 50 MWe (the size Helion says they are aiming for), with a capacitor bank of 150 MJ.

Given the 1:1000 time ratio, this machine peek power per cycle would be 100 GW.

Now, if they could make this same machine work at 10 Hz, then it would generate 500 MWe, with the same capacitor size and the same 100 GW peek power.

[jrvz]

I wonder if the chamber could be evacuated between pulses, at least in part, electromagnetically? After all, there have been proposals to use FRC fusion for spacecraft propulsion, which would require expelling the plasma at a high speed. For example:

Stephanie J. Thomas, Michael A. Paluszek, and Samuel A. Cohen. Fusion-enabled Pluto orbiter and lander. In AIAA SPACE Forum, page 5272, Orlando, FL, 2017.

Yosef S. Razin, Gary Pajer, Mary Breton, Eric Ham, Joseph Mueller, Michael Paluszek, Alan H. Glasser, and Samuel A. Cohen. A direct fusion drive for rocket propulsion. Acta Astronautica, 105(1):145 – 155, 2014.

[RERT]

Roughly my thought. And it doesn’t have to be fast. If they can exhaust the fusion plasma in a few milli-seconds over a distance of a few metres, then it’s perhaps 1/300 the velocity of the original FRCs. Sounds like not a prohibitively powerful pulse.

I have no real idea what happens when the exhaust gets back to the formation region(s?) though…

[usesbiggerwords]

I wonder if the chamber could be evacuated between pulses, at least in part, electromagnetically? After all, there have been proposals to use FRC fusion for spacecraft propulsion, which would require expelling the plasma at a high speed. For example:

Stephanie J. Thomas, Michael A. Paluszek, and Samuel A. Cohen. Fusion-enabled Pluto orbiter and lander. In AIAA SPACE Forum, page 5272, Orlando, FL, 2017.

Yosef S. Razin, Gary Pajer, Mary Breton, Eric Ham, Joseph Mueller, Michael Paluszek, Alan H. Glasser, and Samuel A. Cohen. A direct fusion drive for rocket propulsion. Acta Astronautica, 105(1):145 – 155, 2014.

My understanding is that the chamber is pumped only after the plasma has collapsed. Energetic changed particles don't play well with high speed rotating equipment for long.

[Skipjack]

The pulse only lasts 1 ms give or take a few microseconds. They can't keep it up much longer than that (for several reasons). So they have about 99 ms between pulses at 10Hz and 999 ms between pulses at 1Hz to evacuate the chamber using their custom turbo molecular pumps. Either one is a very long time for plasma. IF they can not make that work, then they would have to go with stronger pulses (more output energy in a single pulse) and a lower pulse rate. The latter would IMHO be less ideal since it puts more weight onto the power supply, capacitors, switches, which would drive up overnight cost. Though no matter what happens, they can always share the power supply between several machines, if they want to (or have to) build a bigger power plant.

[TallDave]

Mixing power and energy is obscuring things in regard to capacitor bank size. I don't see a problem.

Neither pulse frequency, nor peek power production determine capacitor size. Energy needed to get to reach fusion conditions, Q, and energy recovery efficiency do.

the capacitance really only represents the limit of how much power you can recover from a pulse

the energy required to reach fusion is always smaller than the recovered energy and thus irrelevant to the total required capacitance (except that you can't let the grid current drain it below that level or your reactor stops)

but the frequency also dictates the minimum capacitance required to supply a given current (e.g. to send 100MW to the grid over one second you need to send out 100MJ over that second, which you could get from either a single 100MJ pulse gain or (say) two 50MJ pulse gains over that same second)

With an example:
Lets say one of their production reactors needs 100 MJ to get to fusion conditions, and gets to a Q of 1.
That'd mean a total of: 100 MJ injected + 100 MJ produced = 200 MJ.
Now, let's say they can recover 75%, or 150 MJ.
Set apart 100 MJ for the next cycle and we end with a net 50 MJ.
Working at 1 Hz, this machine would give us 50 MWe (the size Helion says they are aiming for), with a capacitor bank of 150 MJ.
Given the 1:1000 time ratio, this machine peek power per cycle would be 100 GW.

yes, this is the same calculation I did but for a smaller plant, so that's reassuring

the problem is even the minimum 50GW (note this is an average, not a peak!) recapture of power per 1ms pulse required for a 50MW plant at 1 Hz (or .1% utilization) is probably going to melt something, as well as requiring 50MJ in capacitance

if they could run the same plant at 10Hz, they'd only have to deal with 5GW of recaptured power, and only need 5MJ of capacitance to hold it (of course that assumes you could still even produce fusion with a pulse using less than 5MJ... presumably there's a pretty hard floor somewhere in there)

a longer pulse could also reduce the required power of a given pulse gain (but not the required capacitance, since the total required pulse energy gain stays constant at a given frequency)

so they'll need to find a sweet spot between power, frequency, and sustainable pulse length, each of which offers its own unique challenges

now my local LWR runs at 2.5GW, but fortunately for Helion (as I'm sure they're well aware), the sweet spot for economic efficiency in a major grid is quite a bit smaller than that at around 100MW, or roughly in the range that their FRCs seem likely to find a workable solution

[charliem]

From what I remember, Helion think they can significantly avoid the high initial capital costs that hamper so many big projects, by going small.

The plan seems to consist in (at least initially) size their reactors so they can be mass produced in an assembly line, and transported by road or rail. Then, if/when GW escale production is needed, just add more reactors.

That is the root of the 50 MWe machine idea, if I'm not mistaken. I remember Dr. Kirley saying one of their goals is to be able to build them by the thousands in a not too distant future.

[TallDave]

the problem is even the minimum 50GW (note this is an average, not a peak!) recapture of power per 1ms pulse required for a 50MW plant at 1 Hz (or .1% utilization) is probably going to melt something,

to elucidate this a bit, it's not so much the 50GW power being safely recaptured and stored in the capacitors, it's the minimum 10GW in transport and brem flying heedlessly around the reaction chamber

along with the 75% of the 10GW D-D reaction that ends up in MeV neutrons, of course... might also take a stab at calculating peak neutronicity for the shielding requirements

I remember Dr. Kirley saying one of their goals is to be able to build them by the thousands in a not too distant future.

yes this makes them far more viable, particularly given the likely need to replace that central reaction chamber every year or so

lol just imagine doing that for the five-story DEMO 2GW tokamak chamber, going to take months if not years (and it only produces 750MW electricity)

meanwhile LWRs generally run at over 90% capacity