Here I've added that test case I was talking about where we add in the velocity space advection term from the vlasov / transport operator which is responsible for the cyclotron motion (gyration) of particles around a magnetic field.
Admittedly, I am not sure as to why I have to have D on one side and N on the other for the boundary terms.
It does not (yet) have an analytic solution, although I could likely come up with one.
Time step error appears to dominate with numerical diffusion of the gaussian into a peanut shape unless dt is reduced by a factor of 10.
Here I've added that test case I was talking about where we add in the velocity space advection term from the vlasov / transport operator which is responsible for the cyclotron motion (gyration) of particles around a magnetic field.
Admittedly, I am not sure as to why I have to have D on one side and N on the other for the boundary terms.
It does not (yet) have an analytic solution, although I could likely come up with one.
Time step error appears to dominate with numerical diffusion of the gaussian into a peanut shape unless dt is reduced by a factor of 10.