First Simulation Success
(6) CommentsA joint team from LPP and Focus Fusion Society has produced a first numerically stable run of 70 ns simulating the formation of a single filament as part of the plasma sheath in a DPF.
The simulation shows a filament pinching itself down from a radius of about 300 microns to around 50 microns, a first step in compressing his plasma and its magnetic field.
While the simulation has only single grid dimension—along the radius of the filament—at each point velocities and magnetic fields in all three dimensions are calculated.
The simulation team consists of John Guillory, Jeff Schoen, Henning Burdack and Luis Angulo. FFS volunteer Burdack, who lives in Germany, provided the implementation that actually ran.
Algorithm
The simulation uses an original algorithm developed by the team and LPP President Lerner.
We had previously found that neither conventional particle-in-cell (PIC) simulations (which model plasma as clouds of particles) nor magnetohydrodynamic (MHD) simulation which model plasma as a fluid, could work for the DPF.
PIC codes, no matter how clever, required astronomical amounts of computing time and MHD codes made the assumption that electrical resistivity was the same in all directions. But in highly magnetized plasma, like in the DPF, current flows much more easily along magnetic field lines than across them. So our original algorithm considers the plasma as a fluid, but inserts realistic formulae for the resistivity in all directions.
Previous versions of the algorithm had shown numerical instability—that is the values started to oscillate wildly in an unrealistic manner. The new runs are the first to be stable and appear to show the filamentation process.
More work needs to be done before the simulation can be relied on. At the same time, we will be stepping up work on the next stage, a two-dimension simulation of the entire sheath run-down.
In time such simulations will complement our experimental observations and analytical theory, allowing us a more detailed look at the DPF process. Our new algorithms may also find applications in other fields of plasma physics.
Simultaneous Firing at Last, First Pinch above 1 Mega-Amp, Record-High Pressure
Two Paths to Fusion
Comments
For a more in depth discussion, start a thread in the forums.I suspect that the values and vectors that work out will suggest important things about plasma physics. This is valuable work!
Congrats! Doing the computer simulation work and the empirical work simultaneously should enable them to complement one another as things move forward.
Outstanding work! Congratulations to all who contributed. I’m pleased to see the recent steps in progress, and heartened thereby. Press on!
I have 2 questions for LPP.
1. Why not make the source openly available from time to time, so that those of us who are interested in such things can offer suggestions. I’m not talking about constantly updating a website, just releasing your current code on ad hoc occasions, when you reach a milestone or get stuck. (The slow PIC code would be interesting to see as well.)
2. Given that the conditions have to at some point become sufficient for fusion, why is it OK in the first place to model it as a non-collisional plasma rather than using Fokker-Planck or a higher level of BBGKY?
I say “in the first place” meaning that PIC is here presumably a “kinetic” simulation corresponding to Vlasov… (?)
I don’t have a very good sense of the time line on a single pulse. A single pulse from initiation until the plasmoid spits out the last alpha particle is 120 nano seconds? Is that right?
So at 7 shakes the plasma sheath is towards the end of the electrode? Filaments are forming? I don’t think I have seen this posted anywhere.
A couple of mileposts would be helpful.
Post a Comment