Martin's update, 12 feb 2013

Polloidal collimation.

Running the high res sim. Field lines seem to be fixed to the boundaries. However, I see spurious features which seem to develop from protections, see attachment (gray scale rim). I'm working no getting rid of this. http://www.pas.rochester.edu/~martinhe/2012/4feb2013.png
Found that the top part, and not the bottom —as we've been thinking—, of the lines outside the spherical wind BC (circular-ish gray region) move towards the axis, despite the fact that I'm using normal mag field BC. This needs to be fixed. http://www.pas.rochester.edu/~martinhe/2012/5feb2013.png

The above way of plotting the field lines is misleading. I've been playing around with the boundary conditions and with visit too.

  1. killing the radial flow velocity in the bottom domain boundary introduces spurious flows.
  2. the stream lines that we use to see mag fields in visit are misleading: I tried using different seed distributions in the grid to plot these lines. Some cases shows that the upper part of the lines moved towards the axis, hence giving the impression that the wind pushed the lines radially. Other seed configurations, on the other hand, showed that the bottom part of the lines were the ones moving.

So, I went back to use simple BC, where bz is forced to be parallel to both the bottom and top boundaries of the domain, and I do not touch the radial velocity at the bottom boundary. Here's a movie showing the results, descent resolutions, 2d cut (but the sim in 3d), and with log(number density) and magnetic field arrows. The fields seem to bend the way we want: stay frozen to the bottom and go around the wind. They only show a radial B component once the wind is well collimated (close to the end of the movie).

​http://www.pas.rochester.edu/~martinhe/andrea/6feb2013.gif

Comments

No comments.