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

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

Post by Billy Catringer »

Thanks, Betruger.

Here is the first one. The SC Core has a radius of 1m from the center of the torus to the center line of the coil. I may revisit this after making some changes to the support design I had in mind.

Image

This version of the beast has a two meter radius. Notice the thickness of the Cool Water jacket. The outermost jacket in this version is solid copper, 1.9cm thick. That is going to change as well.

Image

Here is the two meter beast without the core in place so that you can see how the shells nest together:

Image

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

Post by KitemanSA »

MSimon wrote: BTW I see unsupported "V"s in your dwg. If you intend to support them with struts going to the walls you should make that clearer.

And the piping can not be stepped. It will cause high electrical field gradients.
I don't have a drawing. I sketched on two of tombo's wonderful pieces (almost felt like defacing a De Vinci!)
The first showed ONCE how all the intersections would look (poorly rendered, but, hey that's what you get with powerpoint!)
The second was there to show where the single loop MPG should be cut into a dual loop BEFORE the 4 stratum process is applied. After it is applied, ALL the intersections would look something like the first sketch.
Stepped? Where is it stepped?

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

Post by MSimon »

KitemanSA wrote:
MSimon wrote: BTW I see unsupported "V"s in your dwg. If you intend to support them with struts going to the walls you should make that clearer.

And the piping can not be stepped. It will cause high electrical field gradients.
I don't have a drawing. I sketched on two of tombo's wonderful pieces (almost felt like defacing a De Vinci!)
The first showed ONCE how all the intersections would look (poorly rendered, but, hey that's what you get with powerpoint!)
The second was there to show where the single loop MPG should be cut into a dual loop BEFORE the 4 stratum process is applied. After it is applied, ALL the intersections would look something like the first sketch.
Stepped? Where is it stepped?
Maybe I'm mistaking the joints for steps. It is not clear.
Engineering is the art of making what you want from what you can get at a profit.

imaginatium
Posts: 55
Joined: Mon Jan 05, 2009 10:46 pm

Post by imaginatium »

MSimon wrote:
KitemanSA wrote:
MSimon wrote:Probably. I don't have the link anymore. A few pages were developed and then it died.
One of the problems was that the size was limited.
Do you remember who set it up?
No. Had it been what I consider useful it would have gotten a link on my fusion sidebar. If you want to start one "Wiki Spaces" is what you want.

I found a link:

http://iecfusion.wikispaces.com/IEC+Fus ... ganization
that wiki says it is no longer to be used and the link it points to is defunct.

If people are willing to put their content and designs on it, I will create a wiki and host it for free. I just need need to see at least 3 people commited to being actively involved in creating maintaining the wiki content. A forums and blogs are a lousy format, to work on and refine a design ( or a group of alternate designs.) is anyone willing to be an active contributor?

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

Post by tombo »

All I'm seeing is donuts not supports or tension cage structure that was referred to.
If you don't have it drawn up yet I understand. Bludgeoning these computer models into doing what YOU want them to do, not what the bit pushers that built them thought you wanted to do takes time, especially if you want to be precise.

Feel free to sketch away on my pictures. that is what I'm throwing them up on the screen for. It's not like they are the only vellums or anything.

Here is another thought.

Image

http://i299.photobucket.com/albums/mm31 ... Corner.jpg

http://i299.photobucket.com/albums/mm31 ... Radial.jpg
-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,
Zoom in on the funny cusp region and look closer at the current directions and fields.
Using the picture you just posted:
(BTW What program did you use to mark up my drawing? It looks useful.)
Powerpoint.
tombo wrote:BEFORE adding the cross connects sketch in the current directions.
The field above and below the cross connect points (say) into the sphere and the field left and right of the cross connect points out of the sphere.
In the area that will be cross connected the field points out of the sphere just like it does in the connected regions left and right. All fields change continuously and there are no field nulls.
AFTER adding the cross connects sketch the currents in the cross connects. Now look at the field within the small cross-connected square.
Some fields point into the sphere and some point out of the sphere.
That means there are field nulls in that region. Actually it looks to me like one "X" shaped one. And, of course we all know that field nulls are like a short circuit for the plasma.
No, we know that field nulls are a path for the electrons. The ions should never climb out of the electro-static well that far.
tombo wrote:Yes, it does support the the V's very nicely, but the price in field nulls is too high.
Unless somehow the electrons decide to magically veer sideways, they should head on out the "funny" cusp, get caught by the electro-static field, turn around and come right back in. No harm, no foul. The only real question is the extent of the null field. My suspicion is that we can adjust the coil positioning to where it will not be an issue.
tombo wrote: Dr Mike has pointed this type of problem out to me several times when I had what I thought was a good idea. Welcome to the club.

Another way to look at it is that you have traded an even vertex (per Dr B) with four odd vertices (not per Dr B).
Hunh? I have created the exact field that DrB called the "Funny cusp"
tombo wrote:The funny cusp area does not mean field nulls. In fact those are the highest fields in the device.
As MSimon says above, yes it does.
tombo wrote:The funny cusp comes from its similarity to 1/4 of the circumference of the cusp machine field geometry.
I have no idea what you mean here. The "Funny" cusp as used by DrB was the null field point you ALWAYS get when an even number of alternative direction fields come together. And unless there is a hunk of stuff their way, it is no big deal.

ekribbs
Posts: 13
Joined: Thu Feb 26, 2009 3:04 pm
Location: Laurel, MD

A structural design for Polywell with no tension supports

Post by ekribbs »

I am submitting a description of my vision of a design which does not have tension supports as requested by Billy Catringer.

Sorry, Billy, I have no pictures, so I will do my best to describe my idea.

You want six 2 meter magnets in WB-6 configuration. I assume the requirement of no tension supports means removal of the "nubs" discussed elsewhere. Form follows function, so I ask myself, where else can we hang things. The obvious answer is the vacuum chamber walls. To keep the six toroidal magnets in their proper place, each magnet needs a structure from the magnet to the wall.

I imagined this intermediate structure to be a truss. Imagine the truss to resemble a short, miniature oil rig. Imagine the magnet torus stuck to the top of the mini oil rig, and the base of the rig welded to the inside of the vacuum chamber. I imagined all the cooling, power and control wiring for each magnet to be fed through the legs of the truss piping to the outside of the vacuum chamber. I also imagined thick, free-floating ceramic insulators on the legs of the truss pipes, which must penetrate the outer negatively charged mesh, and thick insulators between the top of the truss to the magnets.

Now the length of the truss can be a problem. If the distance from the magnets to the vacuum chamber walls is long, you are going to have a vibration problem. If you have the funds to build a new vacuum chamber, make the truss as short as you can so as not to interfere with electron/Ion recirculation, and make the new vacuum chamber fit to the outside of that.

If you do not have the funds to build a new vacuum chamber, then we have to deal with a long Cantilevered truss which may have a vibration problem. Why a vibration problem? There will be lateral forces as well as the axial magnetic forces. Lateral forces deflect the magnets. Deflection changes the magnet spacing. That changes the magnetic field. That changes the lateral force. That re-changes the lateral deflection. On and on.

Form follows function. We need to stop the lateral deflection of the long truss. That means to make the truss be Not cantilevered. To solve that problem, my first idea was to place thick ceramic bumpers in each of the eight corners where each triplet of the magnet faces meet. These would also be free floating, and the trusses should be slightly preloaded to keep the bumpers in place.

Other folks might object to the bumper idea, so a second solution would be to insert cross braces between neighboring trusses just below th magnets. Add insulation as needed. That approach seems a kludge to me, but you have to deal with you own situation.

All of those reinforcements to make the trusses Not Cantilevered, of course, would be unnecessary if the vacuum chamber were designed to fit the magnet truss assembly. For your case, such a vacuum chamber would be spherical. Make the trusses as Short and stiff as possible to allow the trusses to be Cantilevered and also by shortness and stiffness resist vibration. Making the base of the "oil rig" truss much wider than the top will also help to resist lateral forces and vibration. At least have a first natural eigenvalue (frequency) of the truss and magnet assembly to be so high that it will not resonate with the magnet forces. A full finite element model of the spherical vacuum chamber with all six trusses and magnets included may be necessary to obtain the real modes of vibration. Short of that, you may make a model of one magnet on its truss and a portion of the vacuum chamber. Then enforce symmetrical boundary conditions at the edges of the portion of the vacuum chamber wall.

Pre-loading is important! You need to do a FEA vibration problem with all the expected forces included, magnetic and structural pre-loading, if applied. The pre-load is important because it influences the natural frequencies. That is how you tune a violin, remember? You tension the strings with the knobs on the end.

Good luck.

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

Post by Billy Catringer »

ekribbs wrote:I am submitting a description of my vision of a design which does not have tension supports as requested by Billy Catringer.

Sorry, Billy, I have no pictures, so I will do my best to describe my idea.

Ed, I greatly appreciate you're taking a hand in this, make no mistake about it.
ekribbs wrote:You want six 2 meter magnets in WB-6 configuration. I assume the requirement of no tension supports means removal of the "nubs" discussed elsewhere.

Absolutely. They are, as I understand it, in the way of particle circulation and I can see no way to cool them reliably. As much as this thing needs supports in tension, I have yet to find a way to install them without causing major trouble.
ekribbs wrote:I imagined this intermediate structure to be a truss. Imagine the truss to resemble a short, miniature oil rig. Imagine the magnet torus stuck to the top of the mini oil rig, and the base of the rig welded to the inside of the vacuum chamber. I imagined all the cooling, power and control wiring for each magnet to be fed through the legs of the truss piping to the outside of the vacuum chamber. I also imagined thick, free-floating ceramic insulators on the legs of the truss pipes, which must penetrate the outer negatively charged mesh, and thick insulators between the top of the truss to the magnets.

I started off thinking that we would be required to build these with an absolute minimum of clutter inside the vacuum vessel. I would imagine that keeping it as clean as practicable is still a requirement. However, when I first began to imagine what this thing would look like I expected them to be standing inside the vessel on twin stalks side by side coming out of the edge of each torus, rather like a lollipop with two sticks. Two obvious flaws with this approach became apparent after I thought about it. One is that assembling the structure with the necessary coolant piping would be exceedingly difficult. The other one was that the "lollipop" configuration would be unlikely to properly deal with all the forces imposed by and on the magnets.
ekribbs wrote:Now the length of the truss can be a problem. If the distance from the magnets to the vacuum chamber walls is long, you are going to have a vibration problem. If you have the funds to build a new vacuum chamber, make the truss as short as you can so as not to interfere with electron/Ion recirculation, and make the new vacuum chamber fit to the outside of that.

I think that we will need to build a vacuum chamber around the physics. That means we will be building a new vacuum vessel and a rather large one at that. However, we cannot do much on the design of it until we have a fair idea on two things. First is the magnet assembly the other is what I think of as the "alpha catcher". The two questions that come to my mind are; "How far away from the center of the magnet assembly does the "alpha catcher" have to be and how big is the "alph catcher"?

My best estimates suggest that each magnet with its coolant jackets and its coolants will weigh over 55,000kg. This figure does not include the supports that attach the torii to something solid and it does not include the weights of the shoes transferring the loads from the SC coils to the structure of the torii. While trying to model a likely shoe last night, it dawned on me that the shoes for the LHe jacket, the LN2 jacket and both water jackets will add some significant weight. More on the weights later, perhaps in another post, but for now let's say we are talking about magnets that will have operational weights of about 60,000kg.
ekribbs wrote:Why a vibration problem? There will be lateral forces as well as the axial magnetic forces. Lateral forces deflect the magnets. Deflection changes the magnet spacing. That changes the magnetic field. That changes the lateral force. That re-changes the lateral deflection. On and on.


I have been worried about vibration from the git-go. On top of the magnetic forces wanting to shove everything around, we are also pumping coolants through the whole thing. There are many sources of vibration in the system and all of them could cause major trouble when you stop to think that we would be jiggling 60 Teslas worth of magnetic fields around. Moving fields of that magnitude would likely generate enough current in your fillings to give you an electrically induced heart attack.

Whether or not we use bumpers, or use additional supports would be determined by the physics. We dasn't put anything in the way of our quantum fuzzy-wuzzies because that would defeat the purpose of our machine. I would not have a problem with bumpers if the physicists do not cry foul. That seems to be the simplest solution, even if it proved necessary to mount them on water cooled spring cans.
ekribbs wrote:All of those reinforcements to make the trusses Not Cantilevered, of course, would be unnecessary if the vacuum chamber were designed to fit the magnet truss assembly. For your case, such a vacuum chamber would be spherical. Make the trusses as Short and stiff as possible to allow the trusses to be Cantilevered and also by shortness and stiffness resist vibration. Making the base of the "oil rig" truss much wider than the top will also help to resist lateral forces and vibration.

So it shall be. I will model it per your advice.
ekribbs wrote:At least have a first natural eigenvalue (frequency) of the truss and magnet assembly to be so high that it will not resonate with the magnet forces. A full finite element model of the spherical vacuum chamber with all six trusses and magnets included may be necessary to obtain the real modes of vibration. Short of that, you may make a model of one magnet on its truss and a portion of the vacuum chamber. Then enforce symmetrical boundary conditions at the edges of the portion of the vacuum chamber wall.

Okay, I'm caught. I have been trying to pull a Tom Sawyer and lack the nerve to con you into white washing my fence.

I did just a wee bit of this back in the late seventies when you paid for time on a mainframe and talked to the mainframe with a printer, a 300 baud modem with a phone coupler--no monitor. I entered the data for my engineer boss and gave him the print outs. I also walked down the steam system we were analyzing and marked up as built drawings. I translated the drawings into the descriptive code the software needed to do the analysis. So, while you might say that I know what needs to go in, the outputs will tell me nothing and I have not worked with a modern FEA package. Yes, I need help whitewashing the fence, or will need it once I have a complete model.
ekribbs wrote:Pre-loading is important! You need to do a FEA vibration problem with all the expected forces included, magnetic and structural pre-loading, if applied. The pre-load is important because it influences the natural frequencies. That is how you tune a violin, remember? You tension the strings with the knobs on the end.

And the pre-loading helps to kill the vibrations before they can get into step. Gotcha.
ekribbs wrote:Good luck.

Thanks, Ed. Now I know which directions to look in. I still have a long, long worry list, but worrying is one of the things I am typically paid to do.

Do stop by and air your own ideas whenever the mood takes you. I know you have something completely different in mind and, if you would like, I'll even take a shot at putting a digital model together for you.

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

Post by MSimon »

This is what we need:

http://www.comsol.com/products/
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
Posts: 6179
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA

Post by KitemanSA »

MSimon wrote:This is what we need:

http://www.comsol.com/products/
How much?

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

Post by KitemanSA »

MSimon wrote:This is what we need:

http://www.comsol.com/products/
How much?

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

Post by MSimon »

KitemanSA wrote:
MSimon wrote:This is what we need:

http://www.comsol.com/products/
How much?
I think it starts at $10K per year per seat. But the do have prices at their site (I think) so I'll have to verify.

You can request a price list here:

http://www.comsol.com/contact/request/
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 »

You will want to load this onto fresh new hardware, preferably 64 bit.

Roger
Posts: 788
Joined: Fri Jul 06, 2007 2:03 am
Location: Metro NY

Post by Roger »

Whats wrong with standoffs like the 4 that held up Wb-6? Tapered cylinders IIRC.
I like the p-B11 resonance peak at 50 KV acceleration. In2 years we'll know.

Post Reply