[FLASH-USERS] Boundary Conditions for 3D Spherical Geometry

[Benny Tsang]

Dear FLASH users and developers,

Short version:
What should be the boundary conditions in the y (theta) direction for a 3D spherical geometry?

More background:
I am trying to set up a simulation for an expanding spherical shell using the 3D spherical geometry. As an initial test step, I started with a static sphere with constant density and temperature. Without gravity, radiation, nor magnetic field, there shouldn’t be any dynamics.

However, once it started, there were spurious velocities developed near the y (theta) boundaries which grew and became very large (>1e7 after a few time steps), and eventually crashed the runs. I tried the setup with (ymin = 0.0; ymax=180.0) and without (ymin = 45.0; ymax=135.0) including the poles, and was using time steps that are relatively small (~1000x smaller than dt_hydro). 

I have tried both ‘reflect’ and ‘outflow’ y boundary conditions, and alternated between various Riemann solvers and slope limiters (I’m using the unsplit hydro solver). They all crashed with the same issue. The fact that the spurious velocities developed near the y boundaries led me into thinking that I may not be using the proper boundary conditions. 

I appreciate any comments and suggestions. Thank you very much for your time.


Best,
Benny