Talk-Polywell.org

ekribbs wrote:Now everyone who has ever played with household magnets knows that pushing two "norths" or two "souths" together causes not only large axial forces, but also large lateral forces that make the magnets slip to the side of each other.

Good point. The equilibrium forces in a symmetrical system are easy - everything gets pushed radially outward. But I believe you are right that lateral displacements will be unstable. Figuring out the 3-D forces on the 6 coils for finite displacements will be a bear of an engineering problem. I'm not smart enough to solve that one, so I'll let the engineers play with it. As far as a potential polywell reactor goes, I'm willing to assume that the engineers can find a solution, as long as the physics works.

Art Carlson wrote:

ekribbs wrote:Now everyone who has ever played with household magnets knows that pushing two "norths" or two "souths" together causes not only large axial forces, but also large lateral forces that make the magnets slip to the side of each other.

Good point. The equilibrium forces in a symmetrical system are easy - everything gets pushed radially outward. But I believe you are right that lateral displacements will be unstable. Figuring out the 3-D forces on the 6 coils for finite displacements will be a bear of an engineering problem. I'm not smart enough to solve that one, so I'll let the engineers play with it. As far as a potential polywell reactor goes, I'm willing to assume that the engineers can find a solution, as long as the physics works.

I agree. Now if we can only get more details on the physics.

MSimon wrote:I would build a tension cage inside the vessel to support the ceramic/metalic stand offs. Then your standoffs need not reach the wall.

Of course ultimately an integrated structure is desired.

I think COMSOL (sp?) which does integrated physics/mechanics can handle the problem. It is rather pricey - for an individual.

Cage within the pressure vessel. That would tend to make the standoffs shorter and consequently more stiff to resist bending, but I had imagined that the standoffs would be constructed of some metallic piping and coated with ceramic so that electrical power and LN2 or some other coolant could be supplied directly through those pipes from outside the pressure vessel to the magnets. Making an inner structure just complicates things. A carefully designed structure to the outer pressure vessel which is constrained to eliminate vibration is all that is really needed.

That would tend to make the standoffs shorter and consequently more stiff to resist bending,

That is the idea. It is due to the pressure vessel not being optimized for the function required.

Yeah. It is a slight complication. As long as there is minimal compression/tension on the feeds that exit the cage I don't see any serious difficulty.

Because, as you well know, shear is a bitch. Not to mention the strength loss from excessive l/d ratios.

MSimon wrote:

That would tend to make the standoffs shorter and consequently more stiff to resist bending,

That is the idea. It is due to the pressure vessel not being optimized for the function required.

Yeah. It is a slight complication. As long as there is minimal compression/tension on the feeds that exit the cage I don't see any serious difficulty.
l/d ratios
Because, as you well know, shear is a bitch. Not to mention the strength loss from excessive l/d ratios.

In all my 27 years of education, I never heard of an "l/d ratio" for strength. A truss tower/standoff can be designed wherein the pipe shear forces are unimportant. If you are working with equipment which is inconvenient for an optimal design, well, God bless you. My suggestions are for the next generation of the design of the Polywell so that potential design flaws my be avoided. Best Regards, Ed Kribbs

Never forget the buckling issue with struts in compression.
It is a very nonlinear effect often leading to catastrophic mechanical failure.
(one more highly non-linear interaction for the full-up model to deal with. What fun!:) )

I don't have a good image of your "cage".
What shape is it?
What potential is it?
How are you addressing its effect on the arc breakdown issue.
Is it doubling as a "shield grid" mentioned in a direct conversion thread?

I envision the struts as stout enough metal to do their job with whatever insulation is needed covering their outsides.
Perhaps each strut could be a tripod to keep their internal forces exclusively in compression. This avoids that pesky shear and reduces the buckling initiation loads, effectively widening he column without filling the space with too much stuff.
No, I'm not sure how to keep it all inside the shadow. But, the full shadow is a very small zone anyway (see a layout I showed on a direct conversion thread) so I'm not going to worry too much about it once we are beyond about 1 magrid thickness outside the magrid.

tombo wrote:Never forget the buckling issue with struts in compression.
It is a very nonlinear effect often leading to catastrophic mechanical failure.
(one more highly non-linear interaction for the full-up model to deal with. What fun!:) )

I don't have a good image of your "cage".
What shape is it?
What potential is it?
How are you addressing its effect on the arc breakdown issue.
Is it doubling as a "shield grid" mentioned in a direct conversion thread?

I envision the struts as stout enough metal to do their job with whatever insulation is needed covering their outsides.
Perhaps each strut could be a tripod to keep their internal forces exclusively in compression. This avoids that pesky shear and reduces the buckling initiation loads, effectively widening he column without filling the space with too much stuff.
No, I'm not sure how to keep it all inside the shadow. But, the full shadow is a very small zone anyway (see a layout I showed on a direct conversion thread) so I'm not going to worry too much about it once we are beyond about 1 magrid thickness outside the magrid.

Hello tombo,

The "cage" as suggested by MSimon would have to be a spherical thing inside whatever pressure vessel he is dealing with that may be cylindrical, or whatever. The need for a spherical internal structural cage would be to remove some problems related to pipe shear and bending of base supports of the magnets in a new design wherein the current inter-magnet nubs may be eliminated. MSimon's cage must necessarily be quite a stout thing, not only to absorb the hoop stresses, but also to absorb the induced bending from lateral magnetic forces, if no other structural considerations (constraints) are added. as I previously suggested. On the other hand, a spherical pressure vessel sufficient to support a near vacuum should be plenty stout enough without an internal spherical cage. Certainly, buckling could be an issue, but that is also a consideration of a proper and carefully designed structure, as I mentioned before.

In all my 27 years of education, I never heard of an "l/d ratio" for strength. A truss tower/standoff can be designed wherein the pipe shear forces are unimportant. If you are working with equipment which is inconvenient for an optimal design, well, God bless you. My suggestions are for the next generation of the design of the Polywell so that potential design flaws my be avoided. Best Regards, Ed Kribbs

I'm all for that. In the mean time we have what we have and we may have to figure out how to use it.

BTW not being an ME I may get the terminology wrong. But I do know that the force a given member can resist is dependent on the l/d ratio. i.e. generically strength. Maybe you call it buckling resistance. Or perhaps you can give me the preferred term. It still amounts to effective strength. Or if you prefer - safe loading.

Note: I don't expect you to be able to do a rate monotonic communication/control system. Or figure out worst case set up and hold times for data/clock streams that are asynchronous. Or the difference between black box and white box testing of software modules. Or the advantages of doing a PID loop in the time domain vs frequency domain.

The truss needs to be effectively spherical. I'm not sure it can be fully latticed.

OTOH. If it interests you do an optimal design. Have fun.

ekribbs,

Please check your inbox.

Thanks.

MSimon wrote:

In all my 27 years of education, I never heard of an "l/d ratio" for strength. A truss tower/standoff can be designed wherein the pipe shear forces are unimportant. If you are working with equipment which is inconvenient for an optimal design, well, God bless you. My suggestions are for the next generation of the design of the Polywell so that potential design flaws my be avoided. Best Regards, Ed Kribbs

I'm all for that. In the mean time we have what we have and we may have to figure out how to use it.

BTW not being an ME I may get the terminology wrong. But I do know that the force a given member can resist is dependent on the l/d ratio. i.e. generically strength. Maybe you call it buckling resistance. Or perhaps you can give me the preferred term. It still amounts to effective strength. Or if you prefer - safe loading.

Note: I don't expect you to be able to do a rate monotonic communication/control system. Or figure out worst case set up and hold times for data/clock streams that are asynchronous. Or the difference between black box and white box testing of software modules. Or the advantages of doing a PID loop in the time domain vs frequency domain.

The truss needs to be effectively spherical. I'm not sure it can be fully latticed.

OTOH. If it interests you do an optimal design. Have fun.

My Goodness, aren't We important! It may interest you that I do know something about time domain and frequency domain analysis! PID loops may be of of interest to IT types, and signal processing types, asynchronous or otherwise, but it has nothing to do with the interaction of magnetic field forces and structural response that I have been trying to point out today.

I/d ratio or whatever that you may call something you remember is inconsequential for a land based or ocean ship based power system. Polywell will not (in any near future) be used in Airplanes where weight is important. Make the blessed truss as stout as necessary. If you must deal with a cylindrical pressure vessel you may have to make the Polywell a structure an axi-symmetrical thing instead of spherical symmetrical thing. All I will say on that subject is to imagine a PolyWell in a WWI or WWII 6 cylinder radial airplane engine form.

If you get this, you may also get the possibility of stacking layers of radial Polywells.

Make that a 9-cylinder radial engine and I'll buy it. There are 5- and 7-cylinder radial engines, but I've never seen 6.

The forces on the magnets on all the Polywells I worked with were substantial, but not challenging in an engineering sense. You would not want to hold them together with your hands or with duct tape, but hardware-store bolts could handle the loads. Engine design would be more challenging.

Tom Ligon wrote:Make that a 9-cylinder radial engine and I'll buy it. There are 5- and 7-cylinder radial engines, but I've never seen 6.

The forces on the magnets on all the Polywells I worked with were substantial, but not challenging in an engineering sense. You would not want to hold them together with your hands or with duct tape, but hardware-store bolts could handle the loads. Engine design would be more challenging.

Tom,

I suggested a "6 cylinder" arrangement because , as you know, the Polywell requires co-linear opposing magnetic fields, not exactly a thermo-cycle gas engine thing. So an axi-symmetric Polywell engine would be a 4, 6, 8 or any even numbered radial design, with two opposing magnets on the engine axis to close the magnetic bottle.

If you are not boosting the B field to take advantage of the (B^^3)*R scaling law, then that is something that you all need to be addressing first.

A balloon doesn't need a "cage" to hold it in. The pressure loads the material with simple membrane stresses.

A polywell doesn't need a cage either. Given a layup like being discussed in the design topic, the repulsive forces can be taken out with simple tensile loads in the coil backbone. No cage needed. This rather light structure can be suspended with the four radial i/o leads.

With regards to the lateral forces on the magnets

Art said:

Good point. The equilibrium forces in a symmetrical system are easy - everything gets pushed radially outward. But I believe you are right that lateral displacements will be unstable.

I agree. Any slight asymmetry of alignment of the coils will produce lateral forces upon the coils, I don't see finding load member configurations for those forces to be a huge problem.

However, what I do wonder about, and have done for some time now, is how much torque these inevitable lateral forces will apply to the plasma?

And obviously what will be the implications of this for the bulk motion of the plasma?

I think this thing is gonna want to spin up and rather quickly.

It will be near on impossible to get the symmetry so good that there will be no resultant magnetic torques and if there are resultant torques there are few dissipative mechanisms to slow the plasma down. Anybody want to do a viscosity derivation for this plasma? (As this would be the ultimate determinant of terminal total vorticity, i.e. when applied torque and angular momentum dissipation are in balance.)

Who knows, if it is advantageous, for example for hydromagnetic stability or turbulence reduction, maybe a prescribed asymmetry could be designed in to intentionally spin the heck out of the wiffleball. The centrifugal forces would tend to make the ball more spherical, than lumpy, albeit oblate(?).... just musing here.

Hello? Membrane? No moments, no torques. What up?