EMC2's EIXL code and state-space modeling

Discuss how polywell fusion works; share theoretical questions and answers.

Moderators: tonybarry, MSimon

Post Reply
rcain
Posts: 992
Joined: Mon Apr 14, 2008 2:43 pm
Contact:

EMC2's EIXL code and state-space modeling

Post by rcain »

I did raise this on another thread, but think it may have been passed over.

Does anyone know, can we get access to a copy EMC2's EIXL code to inspect?

I am interested in how it modeled:

a) distributions
b) collisions
c) parameterisation

I have an idea to produce a simplified state-space model using z-transforms, looking for stabilities/instabilities, attractors, sinks, etc, by region/domain. (Art's 'simple' 9D optimisation problem)

anyone have any input, thoughts here?

rnebel
Posts: 144
Joined: Mon Dec 24, 2007 12:15 am

Post by rnebel »

EIXL is a 1.5-D Vlasov-Poisson equilibrium solver. The fact that it's Vlasov means that it doesn't have collisions. It assumes 1-D spherical potential wells. The .5-D is that it assumes finite angular momentum which is a conserved quantity.

The fact that it is an equilibrium code generally means that it is an elliptic set of p.d.e.s which will require iterative solutions. My understanding is that it requires some user intervention and doesn't always converge. I don't have a copy of it.

drmike
Posts: 825
Joined: Sat Jul 14, 2007 11:54 pm
Contact:

Post by drmike »

Here is a great list of software and information on many packages. I've used GSL with excellent results, and have written my own code (with not enough time to debug it all fully....)

Both Octave and ]url=http://www.scilab.org/]Scilab[/url] can do PDE's and they are Matlab look alikes.

If you have a hell of a lot of money and want really pretty pictures, check out Comsol. I keep getting e-mail brochures from these guys and all I can do is drool. :D

The main problem is getting enough time to focus on getting the model right.

rcain
Posts: 992
Joined: Mon Apr 14, 2008 2:43 pm
Contact:

Post by rcain »

Gentlemen,

Thank you very much for your feedback.

I had wondered about the Vlasov Vs collisions issue - then came across...
Bussard wrote: ..in which... particle collisions are estimated from density and energy distributions...
in his Valencia paper (p7).

It does surprise me that no one seems to have access to the EIXL code: over 10 years in development one would have thought it would be quite useful/robust. Bussard sounded quite proud of it.

I'm afraid I dont have loads'a money either DrMike; I think if I did I would rather invest it in keeping Dr Nebel gainfully employed in experimentation a while longer.

As for pretty pictures - I think anyone would be hard pressed to beat Indreks fantastic creations thus far; I would be very happy to produce a smidgen of theoretical data in refutation of Art's back of an envelope calcuations (he seems to have pretty large envelopes - plus or minus a few orders of magnitude ;) ).

drmike
Posts: 825
Joined: Sat Jul 14, 2007 11:54 pm
Contact:

Post by drmike »

I started working on it a while back and got some theory down, but I didn't have time to keep running with it. Way too many projects, and building real things is more fun than pure theory for me. But I have to go back to theory eventually because I'll want answers to things that I observe that don't make sense - mostly because I don't actually understand the theory!

Given the processing power available today, I think it makes sense to build models that are more sophisticated and can get closer to the mark. Actually being able to model non-linear unstable processes is no simple task, but they can overcome the past 60 years of hand waving arguments and remove unwanted surprises in experiments.

Both good experiments and good models can help us determine if this will really work well enough to be profitable. I sure hope we get the time to find out!

rcain
Posts: 992
Joined: Mon Apr 14, 2008 2:43 pm
Contact:

Post by rcain »

Hi Dr Mike

Just taking a look over your BFR work at http://www.eskimo.com/~eresrch/Fusion/ - looks very impressive. Hope you don't mind if I scavenge?

Know what you mean about time constraints - luckily this subject seems to possess an addictive quality.

Rather contrarily, I actually prefer the theoretical side of things (notwithstanding a real 'physical' exclamation of 'Eureka!' from deep within Dr Nebel's laboratory).

Its true, experiment with no theory, like theory with no experiment is likely to be dismissed ... possibly rightly so.

Anyway, I think my only hope of making some small positive contribution to this lofty debate may come from left-field lateral thinking.

... now, wheres that Holy Grail ... beauty, simplicity, truth ... :)

apouliot
Posts: 13
Joined: Thu Oct 09, 2008 2:30 pm
Location: Canada
Contact:

Post by apouliot »

Don't know if it will help, but for computer analysis of a polywell, if I had to code it I would make a grid who use the spherical coordinate system. But with a twist the resolution would grow from the center. Something like the first sphere radius would be 50pm the next would be 150pm(100pm+50pm), etc... That would give a good resolution in the center of the system and the farther you go you decrease the resolution, but since the density is smaller you can proceed to do group calculation. Also for the angle covered you could begin with pi/2 section and each few shell reduce the angle by a factor of 2.

Post Reply