## Copper Electromagnets and Power and Heating limits

Discuss the technical details of an "open source" community-driven design of a polywell reactor.

Moderators: tonybarry, MSimon

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

### Copper Electromagnets and Power and Heating limits

It occurred to me that I had erroneously described the the heat limiting scaling of copper electromagnets in several different threads in the past. If the realitive deminsions of the magnets are maintained as the size scales up, I claimed that the internal volume would increase 4 fold for a doubling of magrid radius. The circumfrence would double, so I assumed that while the volume aviable for windings was 4X, the distance (circumfrence) doubled, so the net gain would be 2 x due to double the heating per tern would limit the current to 1/2 of the original. I was essentially thinking that double the current would be needed since the wire length per turn was doubled.
That was dumb. The current would be unchanged. The voltage might need to be doubled (to push the same current through twice the resistance) with the resultant doubling of the power. But, the heating in the wires comes from the amps, not the power (watts).
So an increase in the magrid to twice it's radius could allow for 4X as many turns and 4X increase in the B field.* The magnet heating rate would be the same. If WB8 is as much as 30 cm in radius, this simple scaling would allow for half of the magnetic field strength advertised (8X).
Having tighter control of the sequence of events (turn on magnets, turn on e-guns, turn on gas puffer or ion guns) then perhaps the relevant tests could be done ~twice as fast. In that case the magnets would be on for ~ 1/2 the time as in WB6 (1-2 seconds?). The total heating would thus be the same without further effort needed to cool the magnets. I had originally thought that they would use liquid nitrogen cooling as that increases the conductivity of copper ~ 8X. Now, I believe that if they did this they could increase the magnetic field ~ 64X (ignoring the volume needed for the cooling plumbing). Doubling the magrid radius again to 60 cm could double this again to ~ 12 Tesla for the same time period before the magnets had to be shut down to prevent overheating. Of course the liquid nitrogen flow would have to be great enough to maintain the 8X conductivity advantage, and of course at these coolant flows the magnets could be maintained indefinitely. This may have been obvious to some, but it was not to me. I don't know how much coolant flow would be needed, though I believe some estimates have been given by M. Simon and possibly others.

*The B field strength also involves the radius of the magnet coils and the resulting distance to the point cusps where the field is weakest.

ps: Based on this unrealistic model (ignoring coolant space and structural concerns) a 120 cm radius machine could have B-fields of ~ 24 Tesla. If coolant space, etc occupies ~ 50% of the volume of the magrid casing, then ~ 12 Tesla could be reached- easily surpassing the 10 Tesla strengths Bussard used for his Demo reactor.
Lets see, the copper wire electromagnets power consumption be ~ 20,000 W (my estimate of WB6 magnet power input) X 120 fold increase in magnetic strength = ~ 2,400,000 W. I believe that the 8X advantage in conductivity with liquid nitrogen temperatures would allow thinner wires with resultant ~8X increase in wire turns, so the current (or) voltage could be reduced the same amount. So, the power required to run the magnets might only be ~ 400,000 watts. This would still be a small percentage of the total power needed to drive the system as the electron drive power would be ~ 64 times greater than WB6, which I believe was ~ 480,000 W. With r^2 scaling (for a 120 cm radius machine), this input would be ~64X larger, or ~ 30,000,000 watts or more

Dan Tibbets
To error is human... and I'm very human.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am
I will add that the 30,000,000 W input power estimate for a breakeven Demo Polywell is based on my attempt to directly scale the WB6 claimes ( if indeed it ends up with a radius of 120 cm). If the electron drive voltage is increased some, the efficiency will improve. I'm guessing that R. Nebels reported 10 MW input power for (I assume) a brerakeven machine incorporating this along with other possible gains from modifying or eliminating the nubs, and/ or increasing the number of faces, Increased fusion efficiencies, etc. Improved electron confinement would decrease input power needed and thus the size needed for breakeven with a synergistic feedback.

While I don't believe R. Nebel gave a size for this hypothetical machine, I assume that he arrived at 10 MW input power when he considered the various needs and resultant costs (B field strength, losses, drive voltage, size, etc).

Dan Tibbets
To error is human... and I'm very human.

MSimon
Posts: 14332
Joined: Mon Jul 16, 2007 7:37 pm
Location: Rockford, Illinois
Contact:
At this point I would build experimental machines (3 T) from MgB SC wire. Cooled to 2 K (a standard these days).

By the time scale up to full power comes - go to the new Fe SCs which should be in production.

With SC magnets machine power scales inversely with size: within limits.

Instead of a 2 m (bore) machine we may want a .5 m bore machine.

So the deal is: with SC magnets we want to build the smallest possible machine.
Engineering is the art of making what you want from what you can get at a profit.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am
Indeed, superconductors have potentially large advantages. But, I have heard concerns about maintaining liquid helium temperatures in this harsh environment with large heating loads and neutron loads (in D-D reactors) on the magrid casings and conductors. High temperature superconductors ease the thermal considerations some, but are they less resistant to neutron damage?

My point is that, the problems with copper electromagnets are not as intimidating as I previously thought, and provide a more proven technology for obtaining and maintaining the needed magnetic field strengths. It can be done today, instead of waiting 5-10 years for superconductor technology to hopefully mature. Obviously, in a demo machine that is not expected to last long, current superconductor technology may serve, provided the cooling challenge can be met. Which is harder - to keep a superconductor at ~ 4 degrees K with only the external thermal loads, of to keep copper cooled to ~ 80 degrees K against external and internal heat loads?

Dan Tibbets
To error is human... and I'm very human.

MSimon
Posts: 14332
Joined: Mon Jul 16, 2007 7:37 pm
Location: Rockford, Illinois
Contact:
Dan,

I have worked out ways to overcome the neutron problems. It just makes the coil casings fatter i.e. limits the minimum size. Also I've taken a first cut at the cooling issues and I'm not particularly concerned.
Engineering is the art of making what you want from what you can get at a profit.

Posts: 6204
Joined: Thu Sep 17, 2009 11:18 pm
Location: North East Coast
How about getting away from metals...
Use ceramic tubing as a central core to wrap the SC around to build the coils? Then have this construct in another ceramic tube. I have no idea how ceramics hold up to n-flux. I do not recall seeing it at all on the S5G and S6G cores/vessels/rod mechanisms. Maybe inside the connectors, but I think they were phenolic...
Hmmm....

After typing this and reading it again before posting, I am thinking the construct must be encased in something to control outgassing.

MSimon
Posts: 14332
Joined: Mon Jul 16, 2007 7:37 pm
Location: Rockford, Illinois
Contact:
Thermal conductivity is actually a desired property if you are doing liquid cooling. And metals yield under delta temp stress. i.e. stresses equalize (within limits).
Engineering is the art of making what you want from what you can get at a profit.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am
MSimon wrote:Thermal conductivity is actually a desired property if you are doing liquid cooling. And metals yield under delta temp stress. i.e. stresses equalize (within limits).
Ditto.

In the case of copper coils , there would be considerable internal heating. Holding that heat in or limiting its conduction/ convection into a cooling liquid would be counter productive. Having a non outgassing ceramic layer on the outside might be better than stainless steel, but vacuum layers between inner and outer structures is the best insulation against incident heat from the plasma.

I know M. Simon has thought about and evaluated the potential use of using MRI scanner superconducting magnets. If these can be obtained and adapted to a Polywell configuration, it could serve as a relatively cheap (in time and money) path to a superconducting magrid. I'm not sure this could be considered as a Demo reactor as I'm guessing it would not be breakeven, but it would serve as an intermediate device that would test many of the critical assumptions about containment, and scaling laws. I'm also not sure that it would be cheaper than liquid nitrogen cooled copper magnet assemblies pushed to similar Teslas.

Dan Tibbets

Dan Tibbets
To error is human... and I'm very human.

Posts: 6204
Joined: Thu Sep 17, 2009 11:18 pm
Location: North East Coast
Ahh....if I won the lottery in the near future...

KitemanSA
Posts: 6114
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA
Seems to me that SCs have the distinct advantage that they only have to baricade the external heat out while the copper magnet needs to contend with both external heat and internal too. (Ok, they'll both have some Xray and neutron energy deposition to worry about to) Providing proper coolant flow throughout the entire bulk of the electro-magnet seems a MUCH more difficult packaging requirement.

Thoughts?

DeltaV
Posts: 2245
Joined: Mon Oct 12, 2009 5:05 am
Cryo-cooled, high-purity nonsuperconductors are also a possibility:

NASA Glenn Research Center Program in High Power Density Motors for Aeropropulsion
http://gltrs.grc.nasa.gov/reports/2005/ ... 213800.pdf
The electrical resistance of high purity aluminum or copper near liquid hydrogen temperature can be 1/100th or less of the room temperature resistance. These conductors could provide higher motor efficiency than normal room temperature motors achieve. But much more importantly, these conductors can carry ten to a hundred times more current density than copper conductors do in normal motors operating at room temperature (refs. 1 to 5). This is a consequence of the low (or zero) resistance and of good heat transfer coefficients to boiling LH2 or to flowing supercritical H2. Thus the conductors can produce higher magnetic field strengths and consequently higher motor torque and power.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am
DeltaV wrote:Cryo-cooled, high-purity nonsuperconductors are also a possibility:

NASA Glenn Research Center Program in High Power Density Motors for Aeropropulsion
http://gltrs.grc.nasa.gov/reports/2005/ ... 213800.pdf
The electrical resistance of high purity aluminum or copper near liquid hydrogen temperature can be 1/100th or less of the room temperature resistance. These conductors could provide higher motor efficiency than normal room temperature motors achieve. But much more importantly, these conductors can carry ten to a hundred times more current density than copper conductors do in normal motors operating at room temperature (refs. 1 to 5). This is a consequence of the low (or zero) resistance and of good heat transfer coefficients to boiling LH2 or to flowing supercritical H2. Thus the conductors can produce higher magnetic field strengths and consequently higher motor torque and power.
H'mm, interesting. I didn't know cooling much below liquid nitrogen temperatures helped. I have read that impurities in even high grade copper limited further improvements in conductivity. This must be super pure copper (or aluminum). Of course, if you are cooling to liquid helium or hydrogen temperatures anyway, I don't see the advantage of avoiding superconductors. Perhaps supercooled copper systems would fail more gracefully if there was a coolant failure.

Dan Tibbets
To error is human... and I'm very human.

Tom Ligon
Posts: 1871
Joined: Wed Aug 22, 2007 1:23 am
Location: Northern Virginia
Contact:
Intriguing ... search for 'liquid nitrogen copper electromagnets' and the first hit is TalkPolywell.org. You would think this subject would provoke more general interest as a practical means of making high fields without resorting to superconductors.

The search was provoked by the posting about a new record magnetic field.

I've been toying with the idea of building an LN2-cooled copper tubing or ribbon coil with the goal of making my own mag-grid, around WB3 size. The fans are getting tired of seeing my old fusor ... time to start planning Dog and Pony III.

IntLibber
Posts: 747
Joined: Wed Sep 24, 2008 3:28 pm
clearly using liquid hydrogen cooling is a plus for any space propulsion application of polywell fusion, and having a more graceful failure mode would reduce risks of catastrophic loss of craft, so IMHO high purity copper, possibly copper tubes (since hydrogen is also a metal, how would hydrogen flowing through the core of copper tubes used in a coil improve electromagnetic performance?).

KitemanSA
Posts: 6114
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA
Tom Ligon wrote: I've been toying with the idea of building an LN2-cooled copper tubing or ribbon coil with the goal of making my own mag-grid, around WB3 size. The fans are getting tired of seeing my old fusor ... time to start planning Dog and Pony III.
Tom,
How big is your vacuumn chamber? If it can hold a WB3 I'd like to discuss actually doing some research with you. With that size capability (20cm major diameter) I think we could get some valuable data.

I'd be willing to help build a series of units. My thought is that the series should go something like:

WBX-3 Triple (2/3) turn WB, LN, Round plan-form coil, X Cusp?
MPGX-1 Single turn MPG, LN, Square plan-form coil, Open Cusp
MPGX16 16 Turn MPG, LN, Square plan-form coil, X-Cusp
MPGB16 16 turn MPG, LN, Bowed Square plan-form coil, X-Cusp

Thoughts?