Scaling

So it seems, the units of the scales read in by astrobear and defined as "rscale" and the like have pre-set unit(s). I.e. rho scale (rscale) is in g/cc. lscale in cm., etc.

These scales are divided into your problem variables (that have the same units of the scales) to create scale-less quantities called computational units (c.u.) that are compatible with astrobear.

It is useful to assign a value to these scales that will result in a small computational unit after division.

This means, choose physically appropriate scales, here is an lscale that might be relevant to your problem:

lscale = 7.4d15 !cm

Here, 1 c.u. (computational unit) of length is equivalent to 7.4 x 1015 cm.

So if you have a clump say that has a length on the order of this scale, you can divide it's length by this scale to get something close to 1 that astrobear uses.

If you want to use units other than CGS, you may personalize them in problem.data, as long as you convert them to CGS before dividing by the scales (remember the goal is non-dimensionalization) The flow of info. should go something like:

vars in problem.data → read into problem.f90 → populate q-arrays with var/varScale

Comments

No comments.