Benchmark vs EIRENE

[jwscook]

1. Compare to the time dependent neutral solution for a MAST-U dry run, in which D is puffed from the high field side into a vacuum and the neutral pressure at the outer mid-plane and subdivertor entrances is measured experimentally. I have Eirene simulations of this which have a B2 grid with a very small hole at the magnetic axis, on which I assume essentially zero plasma. The only thing I'd need to do to adapt to NESO particles would be to switch to specular reflection from all boundary surfaces, which is easily done. Outputs to compare: particle and energy density of D (no D2 will be present) as a function of time. Could compare with and without neutral-neutral collisions...
2. Compare to the steady state (Eirene time independent, tracking neutrals trajectories until they are ionised) neutral solution for a MAST-U simulation in which D atoms are emitted from a point source near the magnetic axis (not exactly because of the aforementioned hole in the grid), with a time-independent, spatially constant background plasma (plasma equations are not solved). Again, specular reflection of the neutrals from all surfaces. I don't have this but should be straightforward to run I think. Outputs to compare: particle and energy density of D in steady state, as well as plasma sources (of particles, ion momentum, electron and ion heat).

[jwscook]

Got the EIRENE grid working in nekmesh.
So far only one boundary label - will need one label per boundary condition type.

[jwscook]

Scaling factors have been included for fluids. Grabbing same from particle pusher app

[jwscook]

Added a way to produce the input from a given distribution.

[JamesEdgeley]

Scaling factors for particles added. Particles given curved trajectories via fictitious centrifugal force in semi-implicit Euler pusher to replicate the effect of EIRENE's "periodic surfaces". A particle moving in a straight line in a tokamak when projected onto a toroidal plane through its location will have a curved trajectory. Diagnostic projection field added.

EIRENE2.mp4

[SMijin]

Next steps: Quantitative comparison with the Eirene output from David before moving on to reactions.

[jwscook]

NESO-Tokamak calculates the same order of magnitude of weights integrated in small boxes across the domain.

10M particles and CG order 4 shows some noise. Try DG0 to remove the noise like low order codes do.