videos of polywell phase space

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

Moderators: tonybarry, MSimon

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

videos of polywell phase space

Post by happyjack27 »

video of evolving potential energy, kinetic energy, and force distribution of a 3 meter polywell at 7 tesla, shown in 3d.

the orders of the "modes" have been change so that the ke and force views are grouped together:

0-normal
1-potential energy
2-kinetic energy
3-radial kinetic energy
4-axial kinetic energy
5-force
6-radial force
7-axial force
8-fusion cross section

this is a video of a slice in the center, as shown in the beginning of the video. watch the "z-mode" slider on the left to see what the z-axis represents. (look up the integer part in the table above).

at the beginning i show both electrons and ions. then i turn off the electrons so you can see the ion behavior more clearly. it's always the same center slice.

it's a long video. these things are so fun to watch!

btw, it looks to me like i might want to change kinetic energy (mv^2/2) to something not squared in v, say mv (is that "inertia"?).

i will post a link to it below when it's done uploading.

like i promised, worth the wait.

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

here it is:

http://www.youtube.com/watch?v=aqyBA4eCt6c

be sure to watch it in HD:)

Aero
Posts: 1200
Joined: Mon Jun 30, 2008 4:36 am
Location: 92111

Post by Aero »

happyjack27 wrote:here it is:

http://www.youtube.com/watch?v=aqyBA4eCt6c

be sure to watch it in HD:)
I don't suppose you could add a narrative describing your control actions, could you? It might help me understand what I'm seeing.
Aero

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

Aero wrote: I don't suppose you could add a narrative describing your control actions, could you? It might help me understand what I'm seeing.
i would love too, but i don't have the software. or the hardware, for that matter. i need to get a mic. and then find software, preferably free, to add a voice-over.

in fact, i'm often making a narrative in my head when i record them (rather pointless, eh?)

for now, though most important thing to watch is the "z-mode" slider value, on the left. the integer part of that corresponds to the numbered list of modes i posted above.

i signed the radial inertia view now, and it tells an unexpected story. uploading now.

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

here's the signed radial velocity component of the ions. bear in mind this is still very soon after the electron wiffleball formed. the ions have yet to make a single pass through the center.

presumably the first few passes will for the most part be coherent oscillations, and then after a while they'll start to thermalize and transport will be more continuous (as opposed to oscillatory) this is all supposition, though. one would actually have to run it through, which at this rate could take a while.

positive (up) is increasing radius, i.e. away from the center. negative (down) is towards the center.

you can see the ions inside the electron core are going outward. presumably from mutual repulsion (the electron core is hollow, so gauss' law says there's no net e-field from the electrons inside it) the ones outside are still heading towards it. the ones at the very edge are travelling outward, and will probably be lost.

on that note, i should note that these are all with the coil charge off. turning the charge on might help to contain these edge ions better, but once they're out they're out. also it would draw back in electrons, reducing electron losses. in theory, at least.

also the plasma density is very low. you can see where it is with the net charge and ion charge logarithmic sliders. i've noticed that if i set it too high i don't get a core or a well. it may just be too much pressure, but i think that's a limitation set by my particle resolution (only 14k particles representing many more); i.e. i think i need a finer approximation to simulate the MHD of higher density plasmas. in any case, the sim is limited that way.


oh, and the colors: blue are duetrium, cyan are tritium, and the green are just a few selected deutriums, colored to make them easier to visually track.

oh, and the electron and ion sources are both uniform throughout a sphere of half the magrid radius, centered at the origin. this roughly simulates injection by ionization of a neutral gas, with a few extra electrons shot in from the outside.

i call this one "W is for Wiffleball".

http://www.youtube.com/watch?v=JBObkl0EQGg

DeltaV
Posts: 2245
Joined: Mon Oct 12, 2009 5:05 am

Post by DeltaV »

happyjack27 wrote:
Aero wrote: I don't suppose you could add a narrative describing your control actions, could you? It might help me understand what I'm seeing.
i would love too, but i don't have the software. or the hardware, for that matter. i need to get a mic. and then find software, preferably free, to add a voice-over.

in fact, i'm often making a narrative in my head when i record them (rather pointless, eh?)

for now, though most important thing to watch is the "z-mode" slider value, on the left. the integer part of that corresponds to the numbered list of modes i posted above.

i signed the radial inertia view now, and it tells an unexpected story. uploading now.
You could put some text in the display showing the current mode:
http://www.opengl.org/resources/features/fontsurvey/

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

DeltaV wrote: You could put some text in the display showing the current mode:
http://www.opengl.org/resources/features/fontsurvey/
thanks. i just might do that.

i let it run overnight so i could get a better idea of thermalization.

this looks like bad news for continuous operation, but my sim starts to get inaccurate at these plasma densities, and this is only a small fraction of a second, and without a charged grid. so.... not to be fully trusted.

anyways, this is a new view. i've put radius cubed on the x-axis, so you can see an accurate depicition of relative density. the y axis is radial velocity times mass, and the z axis is axial velocity times mass.

you can see firstly that axial thermalization is fairly ubiquitos. secondly you see a sad story of poor ion confinement. the radial velocity is mostly positive and it grows as you go away from the center.

good news is the core is pretty dense.

also you can clearly see the radial velocity separation between deutrium (blue) and tritium (cyan).

i think i'm going to switch to smaller polywells, maybe as small as 10cm. no bigger than 1m. that way i can get more realistic plasma densities without hurting the accuracy, and since i'll have lower b-fields and densities, i can increase the time step so it'll run faster. i presume the phase space geometry just scales with the other parameters, so presumably we'll know what the big ones look like from looking at the small ones.

http://www.youtube.com/watch?v=4ZQSteAMGO8

ladajo
Posts: 6258
Joined: Thu Sep 17, 2009 11:18 pm
Location: North East Coast

Post by ladajo »

You should run two, a 30cm at .1T and a 30cm at .8T, replication of the WB6, 7 and 8 runs more or less.
Be interesting to see, and then compare to actual data later... ;)

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

so i set up vnc on my desktop computer this morning. took me a while to realize that time warner was blocking port 5900. so i had to change the port number. so now i can run these simulations from anywhere.

i brought the size down to 15cm radius, then calculated the amp turns for a 0.8 tesla field strength. my sim uses powers of ten so i put it at 10E5.292 amp turns which turns out to be 0.82 tesla.

then i futzed with the particle counts a little and found a good operating regime. still probably a pretty low particle density compared to real life, but orders of magnitude closer, and a this scale the simulation is much more accurate.

i uploaded the radial momentum view of a 9cm slice of the center. you can clearly see it's higher in the center, as it should be. if you bring down the scale, though (as i do near the end of the short video), you can see that a lot of ions are escaping out the sides at high energies, though.

i don't know if i run it long enough to let these particles die and get recycled a few times if the loss rate will go down. that would certainly take some time to test.

still no coil charge. i'm going to have to revisit my equations for that soon enough.

note these particles were all started out uniformly up to 60% of the magrid radius. so when you see the variance in their momentum, that's a direct result of the variance in their initial electric potential energy. it does not neccessarilly suggest radial thermalization.

to get a better idea of radial thermalization, i'd have to start them out at the same radial distance from the center. but i don't know how realistic that is. that would certainly be ideal in real life. anyone know how finely the radial distribution of ionization can be controlled in real life?

http://www.youtube.com/watch?v=ksh3SWhYECc

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

here's the electric potential energy for the same system:

http://www.youtube.com/watch?v=8AlEiByWl1Y --disregard this video

interestingly, it looks pretty chaotic.

the periodic short pauses are probably when it's sending screenshots to my vnc connection.

err... i shrunk my time scale down and now it's acting differently. seems that was the fault of a poor time resolution. disregard. at least now i know how to check. i'll make a new one.
Last edited by happyjack27 on Thu Dec 02, 2010 9:34 pm, edited 1 time in total.

ladajo
Posts: 6258
Joined: Thu Sep 17, 2009 11:18 pm
Location: North East Coast

Post by ladajo »

To date, the only info we have revolves around puff gas feeding of fuel. That does not tend to lead to any control. The radial ionization in the machines would probably be dependant on the electron population and injection methods. We know they have improved the injection of both electrons and ions, but have no details on how. In short, no idea.

Did you do a comparision run for a .1T machine yet?
Scaling is currenlty the explored issue, and even more so with Joel Rogers latest study which says power scales to the 3 on radius vice Bussard's 5. Critical difference. I am curious to see what you model produces in regard to B Field scaling. Will your numbers follow Bussard's model?

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

ladajo wrote:To date, the only info we have revolves around puff gas feeding of fuel. That does not tend to lead to any control. The radial ionization in the machines would probably be dependant on the electron population and injection methods. We know they have improved the injection of both electrons and ions, but have no details on how. In short, no idea.

Did you do a comparision run for a .1T machine yet?
Scaling is currenlty the explored issue, and even more so with Joel Rogers latest study which says power scales to the 3 on radius vice Bussard's 5. Critical difference. I am curious to see what you model produces in regard to B Field scaling. Will your numbers follow Bussard's model?
i don't have any "numbers" yet. just visualizations. i haven't written the code to add up the numbers yet.

i've lowered the time step and i'm going to see how that looks after an hour or so. if it looks good, i'll record those parameters and make a video. then i'll run with same parameters (and visualization scales) except at .1T and make videos for that.

then i think i might move on to the inverse wb-6 with those params.

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

i might be wrong on that time-scale potential energy thing - that video may be accurate. the chaotic behavior might just be the ions oscillating back and forth through the center.

i'll have to find some video acceleration utilitizes to find out. record at low time step, then speed it up 100x. that's my next mission.

for now, here is the phase space view of a 15cm radius 0.82 tesla polywell:

15cm radius, 0.82 tesla polywell
phase space evolution


http://www.youtube.com/watch?v=BKuWlbm3tbQ

in order:

1. electric potential energy
2. total momentum
3. radial momentum
4. axial momentum
5. total force
6. radial force
7. axial force

(this is listed on the video description, too)

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

15cm radius ~0.1 tesla, same settings above. same views as above.

http://www.youtube.com/watch?v=trtyWzvzoWA

looks like mostly it doesn't contain ions as well. probably leaks electrons a lot faster too.

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

i'm making a "reference simulation": i'm running it for hours at a very small time step, and then i'm going to speed it up w/video editing software.
then i can run at larger time steps and compare to see if the resolution is good enough or i have to run lower.

this is going to take a few days, so it means i won't be simulating any fancy configurations for a while. but when i do, it'll be much more accurate. and take a lot longer.

Post Reply