Question about coil spacing effects.
-
KeithChard
- Posts: 41
- Joined: Tue Jul 08, 2008 9:16 pm
- Location: Banff,Scotland
Question about coil spacing effects.
On the Design forum it has become apparent that the size of the cooling system for the Magrid coils will require that there be a large separation between the individual coils, at least an order of magnitude larger than the small number of gyro radii that are talked about currently. The close spacings have been used on WB6 and presumably on WB7. I am not aware that this issue has been raised before and I think it would be wise to discuss the implications of this on the power output and power losses of the BFR, so I am opening this thread for the discussion to start.
Keith
-
Art Carlson
- Posts: 794
- Joined: Tue Jun 24, 2008 7:56 am
- Location: Munich, Germany
increasing separation should not degrade confinement
What does the separation have to do with the cooling system?
I think the (cusp) losses will be dominated by the cusps with the smallest fields. If the coils are close together, the field between them will be high. If they are moved apart a moderate amount, the field there will still be comparable to the field elsewhere, and the field elsewhere will not change much. I thus would expect that the separation between coils could be a substantial fraction of the coil radius (say 10%), and much larger than any gyro-radii, without compromising the confinement.
I think the (cusp) losses will be dominated by the cusps with the smallest fields. If the coils are close together, the field between them will be high. If they are moved apart a moderate amount, the field there will still be comparable to the field elsewhere, and the field elsewhere will not change much. I thus would expect that the separation between coils could be a substantial fraction of the coil radius (say 10%), and much larger than any gyro-radii, without compromising the confinement.
Keith has observed that the thermal protection system being discussed is perhaps 1.5 to 2.5 coil radii thick, so the coil separation will be not 10% but 150-250%. Then what?
The way I look at it, if the field that is produced at the 2.5 radii is as good or better than that which would have been created by copper that 2.5 radii thick, why should we care? The field doesn't care what creates it, and we should only care that it is strong enough and steady state enough at the surface of the coil can to make useful power.
The way I look at it, if the field that is produced at the 2.5 radii is as good or better than that which would have been created by copper that 2.5 radii thick, why should we care? The field doesn't care what creates it, and we should only care that it is strong enough and steady state enough at the surface of the coil can to make useful power.
Last edited by KitemanSA on Fri Apr 03, 2009 12:12 am, edited 1 time in total.
-
Art Carlson
- Posts: 794
- Joined: Tue Jun 24, 2008 7:56 am
- Location: Munich, Germany
Another constant source of misunderstanding. I was talking about the radius of the current loop, the plan form of the coil, the major radius of the toroidal coil. You are talking about the radius of the conductor itself, the minor radius of the toroidal coil. Can we agree on a terminology here? Major radius vs. minor radius?KitemanSA wrote:Kieth has observed that the thermal protection system being discussed is perhaps 1.5 to 2.5 coil radii thick, so the coil separation will be not 10% but 150-250%. Then what?
I imagine putting these toroidal coils, major radius R and minor radius r, on the faces of a cube with side s. Since I think field strength is more important that transparency, I would make R and r pretty big, maybe R=s/5 and r=s/10. The free bore would then be 0.2s and the closest approach of the coils (across the vertices of the cube) would be (3sqrt(2)-2)*s=0.225s.
Bussard claimed that the maximum attainable field scaled with the linear dimension. I'm not sure if that was for copper coils or superconductors or both, and if there is a discontinuity when you go from one material to the other. I don't understand why he didn't just dump all his notes in the public domain once the embargo was lifted, but he didn't. In lieu of doing real engineering designs, can somebody check catalogs of commercially available superconducting coils, e.g. for medical use? Is there a scaling relation between maximum field and bore size? Of course, the magnet you want to put an arm in will look a bit different from the magnet you put a source of copious 14 MeV neutrons in, but it would at least give us a handle.
I'm positive he expected a large performance jump between copper and superconductors. I'm also sure he intended to use superconductors on all the large reactors, if at all practical. Copper was only for the little lab experiments, and would seriously rob from the power balance.
As for why he never got around to releasing more of the information, the burst of information released in 2006 occurred between his two serious bouts of illness, when he was in tolerable health but desperately looking for financial support for the program. Following the second bout, he was receiving some attention in this regard, so that became his priority. I'm sure a lot of technical information was provided to interested parties, but not more widely disseminated in order to keep potential funders interested.
Had he not received that interest, it is likely he would have simply presented all the data, just to insure it survived him. As things turned out, he managed to pass on the information, and obtain funding, without simply dumping it to the world at large.
The magnets should be a straightforward engineering problem, not some particularly dark secret. The exact mathematical workings of the model are the real key to why he was so convinced this was the right approach, and we only know these in a general sense. If the thing works, I'm sure all this will come out. If it doesn't, then the model is wrong and who cares what is in it.
As for why he never got around to releasing more of the information, the burst of information released in 2006 occurred between his two serious bouts of illness, when he was in tolerable health but desperately looking for financial support for the program. Following the second bout, he was receiving some attention in this regard, so that became his priority. I'm sure a lot of technical information was provided to interested parties, but not more widely disseminated in order to keep potential funders interested.
Had he not received that interest, it is likely he would have simply presented all the data, just to insure it survived him. As things turned out, he managed to pass on the information, and obtain funding, without simply dumping it to the world at large.
The magnets should be a straightforward engineering problem, not some particularly dark secret. The exact mathematical workings of the model are the real key to why he was so convinced this was the right approach, and we only know these in a general sense. If the thing works, I'm sure all this will come out. If it doesn't, then the model is wrong and who cares what is in it.
Bussard claimed that the maximum attainable field scaled with the linear dimension. I'm not sure if that was for copper coils or superconductors or both, and if there is a discontinuity when you go from one material to the other.
They do appear to be planning on that. Rick said they are looking at 5-10T, which would be 50-100 times the WB-7 field of .1T. I bet radius isn't 50-100x bigger!I'm positive he expected a large performance jump between copper and superconductors.
Heh, talk about points of disagreement. In the traditional, non-skeptical Polywell model, there are no cusp losses per se, only a requirement to keep a density ratio.I think the (cusp) losses will be dominated by the cusps with the smallest fields.
Some comments... some of which may even be reasonable.
The magnetic coil casing (minor radius) does not need to be round. It can be elliptical (?). There could be a layer or two of water, liquid nitrogen slieves/ jackets on the inside surface (side that is exposed to the most high energy fusion particles and x-rays) while the lateral sides (importantant to keep the lateral surfaces within the magnetically shielded area) and outer surface would not need as much thermal shielding- less or thiner slieves, so the magnetic wire could be packed into these areas more (to increase the amp turns/ field strength). I'm guessing this might result in an increase in effective internal volume for the magnet wires without increasing the crossectional area exposed to the heat does not increase as much, decreasing the realitive heat load on the coil and allowing for more 'transparency' of the grid in the case of p-B11 for direct conversion of the outward streaming alpha particles by more periferal structures.
As A. Carlson mentioned, increasing tha relative minor diameter realitive to the major diameter would increase the aviable internal volume of the magrid for insulating/cooling layers, and wires. This could be a major engeneering advantage, at least in the thermal cycle based D-D reactor. I don't have any idea how the tradoffs between increased wire and insulater/ cooling volume and the standoff distance that the magnetic fields have to reach to shield the magrid would play out.
Keeping the same minor diameter to major diameter ratio as the WB 6 (~1/10) I figure that for each doubling in diameter, the Ohmic heating of each copper wire loop would double, but the aviable volume within the minor diameter of the magrid would quadruple. This would allow ~ twice the number of amp turms ( minus the volume used for cooling). If my math is right this could allow up to a 10 fold increase in the number of amp turns in a 3 meter diameter WB100 type machine (WB6 had 200 copper wire turns). Assumeing a 1.5 fold increase in amp turms for each doubling of size (rest is cooling space) and assuming liquid nitrogen cooling of the copper wires with a resultant 8 fold increase in conductivity- the magnetic field could be ~ 50 times greater in a 3 meter machine or ~ 5 Tesla for a few seconds (like the WB6) or longer if the coolent flow is high enough .
A increase in the minor diameter to major diameter ratio might increase this further. I'm guessing that the proportionatly greater internal volume aviable for cooling would offset the increased surface crossection of the magrid exposed to heating .
I'm unsure how the input power would scale with copper wires as opposed to the free power requirements of superconducters (ignoring startup costs and any increased refrigeration costs (might refrigeration costs decrease if no need to carry away the Ohmic heating from copper wires?).
So, if I'm not too far off, a liquid nitrogen cooled copper wire magrid design should be adiquite to test all of the scaling and geometry questions, and possibly even provide a larger, but workable solution for a commercial reactor. Again, all of this assumes the Polywell is promising enough to persue.
ps: I like A. Carlson's termonology of minor and major diameter to describe the magrid torus.
Dan Tibbets
The magnetic coil casing (minor radius) does not need to be round. It can be elliptical (?). There could be a layer or two of water, liquid nitrogen slieves/ jackets on the inside surface (side that is exposed to the most high energy fusion particles and x-rays) while the lateral sides (importantant to keep the lateral surfaces within the magnetically shielded area) and outer surface would not need as much thermal shielding- less or thiner slieves, so the magnetic wire could be packed into these areas more (to increase the amp turns/ field strength). I'm guessing this might result in an increase in effective internal volume for the magnet wires without increasing the crossectional area exposed to the heat does not increase as much, decreasing the realitive heat load on the coil and allowing for more 'transparency' of the grid in the case of p-B11 for direct conversion of the outward streaming alpha particles by more periferal structures.
As A. Carlson mentioned, increasing tha relative minor diameter realitive to the major diameter would increase the aviable internal volume of the magrid for insulating/cooling layers, and wires. This could be a major engeneering advantage, at least in the thermal cycle based D-D reactor. I don't have any idea how the tradoffs between increased wire and insulater/ cooling volume and the standoff distance that the magnetic fields have to reach to shield the magrid would play out.
Keeping the same minor diameter to major diameter ratio as the WB 6 (~1/10) I figure that for each doubling in diameter, the Ohmic heating of each copper wire loop would double, but the aviable volume within the minor diameter of the magrid would quadruple. This would allow ~ twice the number of amp turms ( minus the volume used for cooling). If my math is right this could allow up to a 10 fold increase in the number of amp turns in a 3 meter diameter WB100 type machine (WB6 had 200 copper wire turns). Assumeing a 1.5 fold increase in amp turms for each doubling of size (rest is cooling space) and assuming liquid nitrogen cooling of the copper wires with a resultant 8 fold increase in conductivity- the magnetic field could be ~ 50 times greater in a 3 meter machine or ~ 5 Tesla for a few seconds (like the WB6) or longer if the coolent flow is high enough .
A increase in the minor diameter to major diameter ratio might increase this further. I'm guessing that the proportionatly greater internal volume aviable for cooling would offset the increased surface crossection of the magrid exposed to heating .
I'm unsure how the input power would scale with copper wires as opposed to the free power requirements of superconducters (ignoring startup costs and any increased refrigeration costs (might refrigeration costs decrease if no need to carry away the Ohmic heating from copper wires?).
So, if I'm not too far off, a liquid nitrogen cooled copper wire magrid design should be adiquite to test all of the scaling and geometry questions, and possibly even provide a larger, but workable solution for a commercial reactor. Again, all of this assumes the Polywell is promising enough to persue.
ps: I like A. Carlson's termonology of minor and major diameter to describe the magrid torus.
Dan Tibbets
To error is human... and I'm very human.
I'm not sure weather you are referring to the minor radius or the major radius shape/ warpings. If recirculation is actually 100% due to osselation of an electron in a single in a single cusp as opposed to a portion being due to orbiting around a peticular field line , then possibly having a spiky shaped magnet might work(?), but I have no idea how you would wind such a magnet. Perhaps you refer to the rounded corner square magnets that I believe Bussard mentioned.KitemanSA wrote:Technically speaking, the magnet casing could look like the wiffleball itself, couldn't it?D Tibbets wrote: The magnetic coil casing (minor radius) does not need to be round. It can be elliptical (?).
Based on other threads, Ive wondered if external magnets (outside the vacuum vessel walls) might actually work despite Bussard's conclusions. The magnets would have to rest in groves in the vacuum vessel surface, with outward bulging surfaces of the vacuum vessel walls where the cusps are located so that there would be space for recirculation of the electrons (ossilation only). Essentially the vacuum vessel walls would have to be close enouth to the magnets that they would be shielded. This would reduce the thermal loads on the magnets, might simplyfy some of the engeenering for the magnet cooling and support, but it would require a strange porqupine shaped vacuum vessel that would probably be a nightmare to build, cause all sort of arcing and vacuum pumping problems, and would only work with a thermal cycle. Then there are the charges on the magrid to consider.....
Dan Tibbets
To error is human... and I'm very human.