magrid configuration brainstorming

[rcain]

Icarus - thanks for the technical correction - yes, prolate spheroids.

TonyBarry - re: mass inertia in the plasma wiffleball interior - or inertial gradient - isnt this a function of our well-gradient and plasma density equations?

it seems also that we would be in the territory of MHD so would need to bring viscosity, speed of sound in plasma, etc into the equation i think.

Tombo - yes I was thinking of POPS to induce/regulate the spin. But, does the plasma ball/torus velocity need to cut B-lines?

The idea came from looking at Reversed Field Configurations and available analytic techniques, but retaining a magrid concept.

[tombo]

I can't imagine how you could avoid cutting B lines.
Some of the induced currents are what you want because they smooth out the lumps. (???) But too much would be problematical.

Why prolate? Why not keep it spherical?
Is that to bring it back to spherical from the oblate-ness caused by the centrifugal force? (tidal bulge) (Is this even an effect in plasmas? The mass is so low and the fields are so high.)

It might not have to spin. It could oscillate rotationally along with the pops.
Sort of a shimmying dance.

[MSimon]

I can't imagine how you could avoid cutting B lines.
Some of the induced currents are what you want because they smooth out the lumps. (???) But too much would be problematical.

Why prolate? Why not keep it spherical?
Is that to bring it back to spherical from the oblate-ness caused by the centrifugal force? (tidal bulge) (Is this even an effect in plasmas? The mass is so low and the fields are so high.)

It might not have to spin. It could oscillate rotationally along with the pops.
Sort of a shimmying dance.

I could see making it rotate by phasing the POPS drive to the various grids.

You might want to look up the Phasotron Tube (sp?) for a look at how something like that might be done. Do a search here or over at NASA SF. It has been discussed.

Simon

[rcain]

I can't imagine how you could avoid cutting B lines.
Some of the induced currents are what you want because they smooth out the lumps. (???) But too much would be problematical.

indeed a particular type of lumpiness is desirable i think - turbulent flow optimizing a further boundary layer.

i also wonder whether gyro effects could become significant within some hypothetical Pollywell regime.

Why prolate? Why not keep it spherical?

to improve controlled flow in one direction.

It might not have to spin. It could oscillate rotationally along with the pops.

the various possible 'modes' of sheer, spin, etc are interesting; they can help to stabilize and regulate the system and also provide additional non-linear mechanisms (ie. amplification of confinement).

i have a hunch that this might be one of those a topics where making things apparently more complicated has the over all effect of simplifying and improving eventual outcomes. eg: inherent instabilities incorporated into supersonic aerospace design. we end up with a highly tuned machine.

I am wondering how the the geometry and scale of a MHD model compares with magrid and IEC energy boundary models. (I have never used Reynolds Numbers or Navier–Stokes equations - anyone a dab hand?)

[93143]

Just a quick drive-by answer:

I have some experience with fluid mechanics. One thing to consider is that the Knudsen number is probably pretty high (electrons are supposed to make 1e5 transits without thermalizing), so concepts like the Reynolds number aren't really applicable because the action of 'viscosity' is nonlocal. Under these circumstances, the formation of a boundary layer in the conventional sense is not plausible. MHD assumes small perturbations from LTE, and is thus probably not a good way to model the Polywell overall.

In short, compared to familiar fluids like air or n-octane, the stuff is moving extremely fast, at an extremely low density, with a viscosity I can't even guess at (mostly because I don't have my gas kinetics notes handy), and the plasma particles don't tend to hit each other much, with the result that with everything else that's going on, the phase space distribution is highly non-Maxwellian, and the Navier-Stokes equations (or the 20-moment MHD equations, or what have you) don't apply.

I'm not a plasma physicist, but the spinning wiffleball idea strikes me as something that might not have much effect unless it were strong enough to disturb the potential well and/or mess up the cusp geometry (remember, there's a superimposed magnetic field involved). Don't take my word for it, though...

[tombo]

Yes I think messing up the cusp geometry is the idea.
With the intent of skewing the path somehow to tmake it more difficult for the electrons to get out.
I kind of see the the inside portal of the cusp being dragged around with the plasma ball (maybe even being stretched out like taffy) requiring the electron to take a curved path to get out the exit.
(This is entirely intuitive and off the cuff and not mathematical.)

[drmike]

From Francis F. Chen, Introduction to Plasma Physics "A plasma is a quasineutral gas of charged and neutral particles which exhibits collective behavior."

It is the collective part that makes plasmas different than normal gas dynamics. The whole fluid has viscosity that comes from interactions with all of itself, compared to the viscosity of a fluid which comes from near neighbors. It certainly has boundary layers, and these are usually referred to as "sheaths". Debye length, plasma frequency and cyclotron frequency become the fundamental parameters - just like any fluid has fundamental parameters that help characterize its behavior.

I'm not done with the Dolan paper, but it sure seems like a lot of this ground has been covered for the past 50+ years. We have very fast reaction computers these days, so we can do things now which could not be done a decade ago, or even last year. It seems like dynamic control would be a good way to go compared with static control attempted in pulsed systems of the past 50 years. The problem is you need to understand how the plasma will respond so you can properly change things - and our fundamentals are lacking.

Time for me to build stuff....

[93143]

Yes I think messing up the cusp geometry is the idea.

No, I mean in the sense of something like partly cancelling the magnetic field on a near-parallel cusp (and of course reinforcing the one on the other end of the wiffleball). That could be bad.

It certainly has boundary layers, and these are usually referred to as "sheaths".

That's why I said "in the conventional sense". For starters, in the conventional sense, if it's not touching a solid wall it's not a boundary layer.

One thing to remember (it might not be obvious from drmike's post) is that plasmas have more fundamental parameters than neutral fluids, even if you try to stick to the continuum assumption. If you go strongly non-LTE things get very messy very fast.

[drmike]

That's why I said "in the conventional sense". For starters, in the conventional sense, if it's not touching a solid wall it's not a boundary layer.

Agreed. You can create a boundary layer in a plasma using a stiff field,
it appears as a "solid wall". You can't do that with a conventional fluid.

One thing to remember (it might not be obvious from drmike's post) is that plasmas have more fundamental parameters than neutral fluids, even if you try to stick to the continuum assumption. If you go strongly non-LTE things get very messy very fast.

Yup. Lot's of fun there!

[rcain]

Just a quick drive-by answer:

that sounds like a pretty good fly-by to me.

I have some experience with fluid mechanics. One thing to consider is that the Knudsen number is probably pretty high (electrons are supposed to make 1e5 transits without thermalizing), so concepts like the Reynolds number aren't really applicable because the action of 'viscosity' is nonlocal. Under these circumstances, the formation of a boundary layer in the conventional sense is not plausible. MHD assumes small perturbations from LTE, and is thus probably not a good way to model the Polywell overall.

re: 'conventional sense' - and boundary layers vs sheaths - i see picked up by Dr Mike above.

in general i see it as an analogue - also, i see different spaces within the system and different views of the system modeled in different ways, i see local as well as nonlocal phenomenon; one question is do we reach a unified model first or do we end up with a patchwork model.

there is no 'flow' in the current pollywell model (if we can say there is such a model), excepting ion and flux modeling and statistical transports. also system and metric boundaries are coarsely defined.

Maxwellian assumptions of 'fluid' based approaches certainly a good point, Maxwellian distribution functions are a special case. more promising seem to be models based on Vlasov- Poisson solvers and Yang-Mills/Guage Theory.

Low perturbation requirement also a good point. though fluid dynamics can make a pretty good job at transporting mass where the distribution is more Maxellian.

so i suppose put another way, my question might be put as 'under what conditions/in which spaces/phases' does the Ploywell config behave like a fluid?' - i am thinking particularly of the core, the wiffallball boundary, and cusps.

the other question i'm asking is 'how can it be made to behave like a fluid' - potential benefits are both physical and analytic.

In short, compared to familiar fluids like air or n-octane, the stuff is moving extremely fast, at an extremely low density, with a viscosity I can't even guess at (mostly because I don't have my gas kinetics notes handy), and the plasma particles don't tend to hit each other much, with the result that with everything else that's going on, the phase space distribution is highly non-Maxwellian, and the Navier-Stokes equations (or the 20-moment MHD equations, or what have you) don't apply.

take your point. but would this be true in all regions of the system? in particular high ion density regions with lower kinetic energy?

i also wonder whether there should be a spectral 'trace' of fluid phase behavior within the model - whether in fact there is anything preventing fluid mechanics, or whether its just happening at a different scale.

I'm not a plasma physicist, but the spinning wiffleball idea strikes me as something that might not have much effect unless it were strong enough to disturb the potential well and/or mess up the cusp geometry (remember, there's a superimposed magnetic field involved). Don't take my word for it, though...

i was particularly interested in that 'superimposed' field also; are your speaking of the 'imaginary'/diamagnetic field proposed at the heart of the wifleball model by Indrek recently?

ps. here is a nice picture of the config i am musing around:

here is a nice picture of the config i am musing around: http://depts.washington.edu/rppl/progra ... intro.html

[rcain]

From Francis F. Chen, Introduction to Plasma Physics "A plasma is a quasineutral gas of charged and neutral particles which exhibits collective behavior."

It is the collective part that makes plasmas different than normal gas dynamics. The whole fluid has viscosity that comes from interactions with all of itself, compared to the viscosity of a fluid which comes from near neighbors. It certainly has boundary layers, and these are usually referred to as "sheaths". Debye length, plasma frequency and cyclotron frequency become the fundamental parameters - just like any fluid has fundamental parameters that help characterize its behavior.

- but is still amenable to spectral analysis, and also capable of 'complex' behavior/structure.

I'm not done with the Dolan paper, but it sure seems like a lot of this ground has been covered for the past 50+ years. We have very fast reaction computers these days, so we can do things now which could not be done a decade ago, or even last year. It seems like dynamic control would be a good way to go compared with static control attempted in pulsed systems of the past 50 years. The problem is you need to understand how the plasma will respond so you can properly change things - and our fundamentals are lacking.

Time for me to build stuff....

just try to make sure you build it the right shape - i dont know, but i'm guessing it might be important

btw: do you have a link to the Dolan paper?

[Solo]

So someone thinks an FRC could be stuffed inside a polywell? Well, I guess if you have a toroidal cusp you might as well try an FRC/cusp combo.

Dolan's paper is here:
http://mr-fusion.hellblazer.com/pdfs/ma ... nement.pdf

[rcain]

So someone thinks an FRC could be stuffed inside a polywell? Well, I guess if you have a toroidal cusp you might as well try an FRC/cusp combo.

Dolan's paper is here:
http://mr-fusion.hellblazer.com/pdfs/ma ... nement.pdf

thanks Solo.

[Solo]

No problem! I've kind of adopted the project of getting everyone to read it, because it's very helpful in my opinion.

In fact, Dolan mentions someone's discovery that ExB rotation in a spindle cusp can create an effective potential that retards ion loss from a point cusp. (I wonder if that isn't just another way of describing the wiffleball's purported effect on cusp losses?) But I guess ExB rotation is not what we are talking about here.

And I wonder whether rotation would really be defined for a plasma like the one in a polywell? But I think the original proposal was to rotate the fields to keep the plasma guessing where the cusp was, right? That ought to be possible, though maybe not effective.

[rcain]

No problem! I've kind of adopted the project of getting everyone to read it, because it's very helpful in my opinion.

In fact, Dolan mentions someone's discovery that ExB rotation in a spindle cusp can create an effective potential that retards ion loss from a point cusp. (I wonder if that isn't just another way of describing the wiffleball's purported effect on cusp losses?) But I guess ExB rotation is not what we are talking about here.

And I wonder whether rotation would really be defined for a plasma like the one in a polywell? But I think the original proposal was to rotate the fields to keep the plasma guessing where the cusp was, right? That ought to be possible, though maybe not effective.

Hi Solo.
I'm still reading the Dolan paper, in fits and starts - very busy elsewhere.
Dolan looks very interesting so far; lots of ideas, and some accessible maths by the look of it.

Re your last points:

Wiffleball vs ExB rotation - I presume that in so far as they share identical dimensions (/fields), they are indeed related by a common algebra. Where they differ however is topology.

Cusp losses Vs ion plugging - yes, personally I think this is a strong possibility, funnily enough along Art Carlsons previous arguments. The trick, it would seem to me, if there is one, would be in configuring/inducing/sustaining an asymetric ion popluation in those regions, to advantage.

I was further postulating, whether, if in ion rich cusp, sufficient density could be achieved to support critical collisional probability, ie. in common with the core.

I think that 'spinning' alone might not reduce the probability of 'escape' since the particles being contained are essentially moving randomly - therefore its just down to exposed area. However, I am wondering how local gyro-magnetic effects might alter the probability distribution of charge velocity.

I am also wondering whether the Wiffle ball boundary itself and/or the cusps , effectively a charged, magnetic 'shell', might become our actual Polywell core?

(For this reason also, I am interested in Plasma 'sheer' effects and local dynamics, etc. - Gestalt effects)