Is There an Optimal Size for Magrid Casings?

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

Moderators: tonybarry, MSimon

MSimon
Posts: 14334
Joined: Mon Jul 16, 2007 7:37 pm
Location: Rockford, Illinois
Contact:

Post by MSimon »

KitemanSA wrote:
Billy Catringer wrote:This is about as close as I can get the magnets together with this cooling system. I will do some more work on it in an effort to reduce its size and weight.
You may consider trying LO2 rather than LN, and Fluorinert rather than the inner water course. they both have significanly greater liuid temperature ranges which lowers the pumping and volume requirements. Also, LO2 has a lower melting point than LN, so it can be chilled more. This will reduce the thermal load to the LHe circuit markedly. Along that vein, a ~20%N80%O eutectic has an even lower freezing point. Something to think about.
Second, the 1+cm walls an all those courses look WAY excessive unless you are trying to restrain boiling (see first comment).
Third, I don't see any real thermal radiation shielding, but that may just be because it is hard to depict graphically. What I mean here is multiple very thin sheets of foil.
Finally, look up truss core. This is a very light weight, pyramidal truss core panel that makes an excellent and extremely light weight heat exchanger. It is also reasonable easy to manufacture in tubular shapes using metal glass brazing.
I suspect you might be able to get the heat control system much smaller than it is now. For one thing, you are working with evacuated systems in a vacuum chamber. Not a lot of external, buckling type, pressure loads.
LOX is an Oxidizer. LN2 is safer. I have looked into Fluorinert. It produces sludge in a high neutron flux environment and it is tremendously expensive esp. given the volumes required. And then you have the loose fluorine when neutrons break the chemical bonds. Not good.

The LN2O2 idea is interesting. However the MRI guys use LN2. So at minimum it is OK and eqpt to deal with it is available.
Engineering is the art of making what you want from what you can get at a profit.

Billy Catringer
Posts: 221
Joined: Mon Feb 02, 2009 2:32 pm
Location: Texas

Post by Billy Catringer »

KitemanSA wrote:
Billy Catringer wrote:Second, the 1+cm walls an all those courses look WAY excessive unless you are trying to restrain boiling (see first comment).

Were it just the fluids, Kiteman, I would take chances with s/10 pipe.
KitemanSA wrote:[Third, I don't see any real thermal radiation shielding, but that may just be because it is hard to depict graphically. What I mean here is multiple very thin sheets of foil.

Well, there is no insulation other than the vaccum jackets. The idea is to use running water to catch and conduct the heat out of the magnet before it can get to the LHe core. The LHe would be worth recycling and given that we'd be recycling the LHe, recycling the LN2 would not be that much of an add-on.
KitemanSA wrote:[Finally, look up truss core. This is a very light weight, pyramidal truss core panel that makes an excellent and extremely light weight heat exchanger. It is also reasonable easy to manufacture in tubular shapes using metal glass brazing.

I have already advised MSimon that we should consult with All-Clad about having them make the outermost jacket.
KitemanSA wrote:[I suspect you might be able to get the heat control system much smaller than it is now. For one thing, you are working with evacuated systems in a vacuum chamber. Not a lot of external, buckling type, pressure loads.

You keep making this assertion, but you have yet to explain why. What makes you think there are no major structural loads on these magnets?

tombo
Posts: 334
Joined: Fri May 09, 2008 1:10 am
Location: Washington USA

Post by tombo »

Ummm... I had a bad thought. (repeatedly)
Steel is the usual first choice for the strength member but being magnetic it will short circuit the magnetic field if it surrounds the conductor and keep it from getting outside the coil assembly.
I may have popped off again before thinking hard enough.
Do you see it that way too?
-Tom Boydston-
"If we knew what we were doing, it wouldn’t be called research, would it?" ~Albert Einstein

93143
Posts: 1142
Joined: Fri Oct 19, 2007 7:51 pm

Post by 93143 »

Stainless isn't ferromagnetic. Even if it was, I'm pretty sure it would saturate rapidly in a full-scale device.

KitemanSA
Posts: 6179
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA

Post by KitemanSA »

Billy Catringer wrote:
KitemanSA wrote:Second, the 1+cm walls an all those courses look WAY excessive unless you are trying to restrain boiling (see first comment).
Were it just the fluids, Kiteman, I would take chances with s/10 pipe.
Conversely to your last statement, I am concerned that you are trying to react loads inefficiently. It is bad enough that the gravity loads on the SC core need to be carried past the thermal barriers to ground. But it seems that you want to carry the magnetic ones too. This seems unwise if there is an alternative, and there is.
BC also wrote:
KitemanSA wrote:Third, I don't see any real thermal radiation shielding, but that may just be because it is hard to depict graphically. What I mean here is multiple very thin sheets of foil.
Well, there is no insulation other than the vaccum jackets <<that is my point, there should be - KSA, aka DRJ>>. The idea is to use running water to catch and conduct the heat out of the magnet before it can get to the LHe core. The LHe would be worth recycling and given that we'd be recycling the LHe, recycling the LN2 would not be that much of an add-on.
I dont understand how this answers the question. To a first order, the heat that must be extracted by the second (the LN, LO, LNLO, whichever) course of HEX is a function of how much radiates to it from the outside. That amount can be GREATLY reduced by multiple standoff layers of foil. Once the amount of heat is reduced, the size of the HEX can go down. Do this for all the layers, and the overall size and weight of the thermal protection system can go down. Lighter weight, thinner structure...
BC also wrote:
KitemanSA wrote:Finally, look up truss core. This is a very light weight, pyramidal truss core panel that makes an excellent and extremely light weight heat exchanger. It is also reasonable easy to manufacture in tubular shapes using metal glass brazing.
I have already advised MSimon that we should consult with All-Clad about having them make the outermost jacket.
That is good, but I was talking about it for the LN and first water courses. I believe it would result in a MUCH thinner, lighter structure.
BC also wrote:
KitemanSA wrote:I suspect you might be able to get the heat control system much smaller than it is now. For one thing, you are working with evacuated systems in a vacuum chamber. Not a lot of external, buckling type, pressure loads.
You keep making this assertion, but you have yet to explain why. What makes you think there are no major structural loads on these magnets?
I think that the magnet's magnetic loads should be reacted out within the inner-most course and be done with. Then, the structural loads should be reacted out directly to ground with a high strength, low conductivity material (perhaps SiC?). The thermal protection and neutron conversion/protection system should be made as light and thin as possible and reacted out thru one of the layers of that system.

KitemanSA
Posts: 6179
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA

Post by KitemanSA »

MSimon wrote:LOX is an Oxidizer. LN2 is safer. I have looked into Fluorinert. It produces sludge in a high neutron flux environment and it is tremendously expensive esp. given the volumes required. And then you have the loose fluorine when neutrons break the chemical bonds. Not good.

The LN2O2 idea is interesting. However the MRI guys use LN2. So at minimum it is OK and eqpt to deal with it is available.
True, LOX is an oxidizer, but it is handled constantly and safely by facilities all around the world. I mention this because there is another reason to use it that I neglected to mention first time around. The specific heat capacity(SHC) of LOX is about three times that of LN. At least it is if my ancient CRC Handbook is to be believed.

All else being equal, 3x the SHC allows 1/3 the flow, and a smaller, lighter thermal protection system (TPS). Small TPS means lower intercepted flux and a lower thermal load and a still smaller lighter TPS...

I kind of like the LN2O2 eutectic myself. But it must remain in question because I don't know what it's boiling point is. I am not sure whether the marginally lower T(f) is countered by a loss of liquid T range.

Regarding the fluorinert, I shant argue about it, but it would be very nice to be able to lower the third course temperature below 0C. Any neutron safe liquids you know of that can do that? Does ethane or propane have the same sludging issue?
Last edited by KitemanSA on Sat Mar 14, 2009 7:59 pm, edited 1 time in total.

Billy Catringer
Posts: 221
Joined: Mon Feb 02, 2009 2:32 pm
Location: Texas

Post by Billy Catringer »

KitemanSA wrote:Conversely to your last statement, I am concerned that you are trying to react loads inefficiently. It is bad enough that the gravity loads on the SC core need to be carried past the thermal barriers to ground. But it seems that you want to carry the magnetic ones too. This seems unwise if there is an alternative, and there is.

Okay, what do you propose as an alternative? So far, you have not explained what you see as an alternative means of dealing with all the loads. Mind you, we are talking about some pretty big loads, even if we go back to magnets 2 meters in diameter. I am ready, Kiteman. I am all eyes, waiting willing to read what you have in mind.
KitemanSA wrote:I dont understand how this answers the question. To a first order, the heat that must be extracted by the second (the LN, LO, LNLO, whichever) course of HEX is a function of how much radiates to it from the outside. That amount can be GREATLY reduced by multiple standoff layers of foil.

I am going to assume that what you mean with the extemporaneous use of the acronym "HEX" is heat exchange. As it happens, I agree with you that layers of foil could cast a useful shade. Now, how do you propose to one, suspend these layers of foil, and two, cool these layers of foil?
KitemanSA wrote:Finally, look up truss core. This is a very light weight, pyramidal truss core panel that makes an excellent and extremely light weight heat exchanger. It is also reasonable easy to manufacture in tubular shapes using metal glass brazing.
Billy Catringer wrote:I have already advised MSimon that we should consult with All-Clad about having them make the outermost jacket.
That is good, but I was talking about it for the LN and first water courses. I believe it would result in a MUCH thinner, lighter structure

Just nesting one piece of pipe inside the other is not what I am doing. As I explained before, every jacket must have a set of shoes inside. Those shoes transfer the mechanical loads imposed by the magnets to the rest of the structure. The Cool Water Jacket will be the jacket that transfers all of those loads to the supports holding the entire thing up. Honeycomb or pyramidal truss core materials might work for the inner jackets, but they will be thicker than the solid 316 I am proposing to use in the cryogenic layers. Sandwich materials will not work at all for the Cool Water Jacket. We need it to absorb heat very quickly AND be the strongest part of our structure.

The outer jacket, the Hot Water jacket, will be dealing only with the water pressure induced by heating. That pressure will be quite high and we will have to put a lot of water through that jacket. Using the materials All-Clad uses to make skillets might, MAYBE, let us get a way with a much thinner jacket than the one I have drawn. Right now, it is 1.9cm of copper and I am not at all sure that 1.9cm of copper would endure under these temperatures and pressures.

Now, there are two problems with this thick outer jacket. The obvious one is the surprising amount of weight it adds to the structure. The other one though, has to do with same problem your foils have. The surface of it will start turning to vapor before the water can cool it. Honeycomb materials for this jacket would fail almost as quickly as fusion began.
KitemanSA wrote:I suspect you might be able to get the heat control system much smaller than it is now. For one thing, you are working with evacuated systems in a vacuum chamber. Not a lot of external, buckling type, pressure loads.

Yes, there is. We have six 10Tesla solenoids that will be compressing the fields together in a rather small space for such intensities. Can you show me what the magnitude of the resultants will be? Have you done the math? I've been digging and have yet to get an answer that I think is reliable.
KitemanSA wrote:I think that the magnet's magnetic loads should be reacted out within the inner-most course and be done with.

Really? How?
KitemanSA wrote:Then, the structural loads should be reacted out directly to ground with a high strength, low conductivity material (perhaps SiC?).

Silicone carbide holds up wonderfully well to high temperatures and abrasion. That is why we use it for the teeth, mind you, just the teeth of saw blades. It is brittle. It will not do for a material used in structural members unless those members are stacked up like bricks so that they never see anything but compression loads that are spread out evenly.
KitemanSA wrote:The thermal protection and neutron conversion/protection system should be made as light and thin as possible and reacted out thru one of the layers of that system.

Okay, again, explain to me how you propose to do this? Simply asserting that we should do thus and so is not good enough. You have to give me some idea of how you think that this can be done.
Last edited by Billy Catringer on Sat Mar 14, 2009 7:14 pm, edited 1 time in total.

MSimon
Posts: 14334
Joined: Mon Jul 16, 2007 7:37 pm
Location: Rockford, Illinois
Contact:

Post by MSimon »

Anything with carbon has sludging issues. With water the O2 and H2 have to be recombined. But that is well proven technology in the nuclear industry.

And yes O2 can be handled safely. But it complicates the safety rqmts. Apollo 1 fire?

Flow is not a big problem with the LN2 channel. It won't be much. Loss rates for MRIs with total loss systems are on the order of 10 liters a week.

In lieu of thermal shields the vacuum sides of the piping will be flashed with silver or aluminum.
Engineering is the art of making what you want from what you can get at a profit.

KitemanSA
Posts: 6179
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA

Post by KitemanSA »

MSimon wrote: Anything with carbon has sludging issues. With water the O2 and H2 have to be recombined. But that is well proven technology in the nuclear industry.
I guess this then becomes an optimization issue, sludging vs lower third course temperature and less thermal energy flow.
MSimon wrote:And yes O2 can be handled safely. But it complicates the safety rqmts. Apollo 1 fire?
Indeed, pure gaseous O2 in a warm, fuel filled volume repleat with ignition sources is not a great place to be. But does this reference to a tragedy really apply here? Come on, really?
MSimon wrote: Flow is not a big problem with the LN2 channel. It won't be much. Loss rates for MRIs with total loss systems are on the order of 10 liters a week.
MRIs don't have mega-degree plasmas spitting HE neutrons at them. I suspect anything we can do to improve will be beneficial.
MSimon wrote: In lieu of thermal shields the vacuum sides of the piping will be flashed with silver or aluminum.
Silvering the vacuum walls should be a given, but each layer of foil will greatly reduce the flow of heat unless there is too much solid matter cross talk.

Billy Catringer
Posts: 221
Joined: Mon Feb 02, 2009 2:32 pm
Location: Texas

Post by Billy Catringer »

Okay, my first pass at a design for this thing turns out to have more problems than even I had thought about. Ed Kribbs pointed out to me that my shoes-in-the-pipe idea will suffer from significant sag. We can't have that, so I'll start over. And, Kiteman, you are not that far off on your suggestion about the sandwich materials. I'm going to redesign these toroidal tubes as though they were airplane wings. This should save on weight and size, although I don't know how much yet. The last wing I fooled with was made out of balsa.

I'm goin' back to the drawin' board. See ya later.

MSimon
Posts: 14334
Joined: Mon Jul 16, 2007 7:37 pm
Location: Rockford, Illinois
Contact:

Post by MSimon »

Indeed, pure gaseous O2 in a warm, fuel filled volume repleat with ignition sources is not a great place to be. But does this reference to a tragedy really apply here? Come on, really?
You ought to get to know Murphy. Because Murphy definitely wants to know you.
Engineering is the art of making what you want from what you can get at a profit.

MSimon
Posts: 14334
Joined: Mon Jul 16, 2007 7:37 pm
Location: Rockford, Illinois
Contact:

Post by MSimon »

MRIs don't have mega-degree plasmas spitting HE neutrons at them. I suspect anything we can do to improve will be beneficial.
That is the purpose of TWO water jackets. So the LN2/LHe section is facing a 300K section. Just like MRIs. And I have a trick or two up my sleeve for neutrons.
Engineering is the art of making what you want from what you can get at a profit.

tombo
Posts: 334
Joined: Fri May 09, 2008 1:10 am
Location: Washington USA

Post by tombo »

Here is the square planform truncube with the legs moved out of the funny cusp area.
Image
http://i299.photobucket.com/albums/mm31 ... sy2top.jpg

Then I started to try to calculate the loads on the coil.
I started with Indrek’s equation at http://www.mare.ee/indrek/ephi/force/ but quickly discovered that he solved the hard part numerically for only 3 configurations and that his maximum radius was 1 meter and we are modeling 2 meter radius..
Then I slowly discovered that I could not extrapolate to 2 meters with any confidence. (I tried using excel’s multiple regression tool and found the resulting formula to be so sensitive to the coil spacing that it actually gave negative force predictions for some quite reasonable coil spacings.) This was not surprising given that the integrations he did must be quite non-linear.
So, I started trying to get my own handle on the force instead of using his black box.
I need an approximate model that is simple enough for me to solve.
To calculate (estimate) the forces on the bottom coil (Antarctica) I chose to substitute 2 concentric coils (tropic of cancer and tropic of capricorn) for the 4 coils on edge by reconnecting the corners at the equator (in my imagination) then including the top (arctic circle) as a third coil. The diameter is the one that will give the same area as the square coil. The spacing is some kind of average between the min and max of the actual coil.

BUT I stumbled over a problem: The load is not evenly distributed along the beam (of the coil supported by the legs).
The distance between the coil and the adjacent coil at the corner is 0.5 meter or less and in the middle it is 2 meters. Since the effect varies as 1/R the force varies by a factor of 8 or so.
This means that THE LEGS CANNOT BE ANYWHERE EXCEPT VERY CLOSE TO THE CORNERS for structural stress reasons. (I apologize for shouting but this is an important point.)
-Tom Boydston-
"If we knew what we were doing, it wouldn’t be called research, would it?" ~Albert Einstein

tombo
Posts: 334
Joined: Fri May 09, 2008 1:10 am
Location: Washington USA

Post by tombo »

Ethylene glycol automotive antifreeze for both the water jackets can be used from 120C (higher at 15 psi) down to -40C = 233K which is only 156K above 77K which is not a huge temp difference. It is small enough in my mind to allow use of solid insulators. If sludge is a problem then just add filters to take out the sludge. That should work long enough for a proof of concept machine. It’s cheap, it’s safe (as long as you don’t drink it) and it has a high heat capacity.

I found a formula for B at the center of a square 1 turn coil B=(sqrt(2)*u0*I)/(pi*(sideofsquare/2))
From: http://www.tutorvista.com/content/physi ... cal-02.php
It is a higher field than for a circular one which reduces our max current and therefore forces.

Here is an interesting point about designing Bitter Coils:
“high magnetic fields” by Fritz Herlach, Noboru Miura very long url says about Bitter coils: ”… properly subdividing a thick coil into a few concentric coils is of great advantage to the stress management.” It also refers to polyhelix & poly bitter magnets which I will look up sometime. But for now I want to stay focused on stresses.
-Tom Boydston-
"If we knew what we were doing, it wouldn’t be called research, would it?" ~Albert Einstein

KitemanSA
Posts: 6179
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA

Post by KitemanSA »

tombo wrote:Kiteman:
I agree it will have an X shaped region of zero field.
I think it will be more of a concave sided diamond than an X.
tombo wrote:I disagree that zero fields are fine.
If someone can point out the distinction in overall effect between a hole (cusp) due to 1 field (point cusp), due to two fields (line cusp), due to three fields (quasi-line, or maybe better put "terminating" line cusp) or due to four fields (funny cusp) I would greatly appreciate it. Until then, I maintain that a hole is a hole is a hole, and the smaller the better for all of them.
tombo wrote:DrB never drew it this way to my knowledge. [Although, I vaguely remember that one of the enclosed machines had vertex coils and did not work very well for reasons that Art Carlson was very eloquent and adamant about a few months back.]
DrB said it didn't work because of the metal in the way. Did I mis-read him?
tombo wrote: When I look at it I see 48 vertexes, each with an odd number (3) of faces.
Image
First, MARVELOUS image. I wish you had done it with the bowed coils, but I agree that your bottom image is what DrB proposed. Beautiful!
As to 48 vertices, that is interesting because I see 12 vertices of even number (4). I can also stretch it to see 48 with even number (2), but as I say, it is a stretch.
tombo wrote:Of interest there are irreconcilable geometrical issues with the shape.
If the hole is square then the straight to curved transitions cannot be in the same place for the two curves. If they are the same then the hole is very much longer than it is wide. I tried to make that work for hours.
Hmm. I never anticipated the hole to be square. I find that the centers of curvatures of the "round corners" will be situated ~2.5 and 3.5 coil x-section diameters away from the true vertex, resulting in a hole that is ~3 times as tall as wide. This results from taking lines normal to the incoming and outgoing arms of the "square" coil (when using the bowed, spherical layup in your MPG varient). Of course, the further the centers of rotation are away from the vertex, the more squarish the rectangle becomes.
If you would take your spherical MPG and simply copy and past with 90 degree rotations about the north pole, you will see it come out naturally.
tombo wrote:Without the crossovers there are no zero field points:
Image
True, but there are quasi-line cusps (maybe even true line cusps) that stretch from virtual point cusp to virtual point cusp. The void resulting from these long cusps seems much larger, and therefore worse at keeping relative density, than a small point like void.
tombo wrote: I believe that DrB referred to square plan form, round cross section coils like these. (Although maybe closer to touching than shown in this plot.)
I agree completely, but in order to hold these together there must either be a forest of supports outside the MaGrid, or nubs between the coils. My slight modification of his plan has coils (instead of nubs) between the coils and ALL metal is protected. It also provides for adequate cooling, etc.
tombo wrote:IMO the funny cusp is an artifact of approximating the ideal infinitely sharp vertex points with the long rounded almost kissing regions of round coils like in WB6.
Actually, the funny cusp is what you get with my configuration (4 fields meeting together) and a quasi-line cusp is what the WB6 gives as the approximation. The WB6 is best approximated by 24 vertices of odd number (3).
tombo wrote: I further believe that the square plan form reduces it greatly.
This I agree with completely because it greatly shrinks the distance between each pair of the 24 three- field vertices, making them meld into effectively 12 four-field vertices.
tombo wrote:The sharper the bends can be made, the smaller the funny cusp becomes.
Abso-dang-lutely! (or more acurately, the less linear, more funny the cusp becomes.
tombo wrote:It reduces in size to approximately what it would be in a machine with round coils the same radius and spacing as the radius of the curved corners of the square coils.
Gotcha, I think.
tombo wrote:
Here without my notes for your marking pleasure: http://i299.photobucket.com/albums/mm31 ... sOver2.jpg

P.S. Here it is up close:
http://i299.photobucket.com/albums/mm31 ... sOver4.jpg
As I said, BEAUTIFUL!!! No can you do the bowed MPG as I described? Should be fairly easy. :D

Post Reply