[aspect-devel] Temperature mesh refinement criteria

[Wolfgang Bangerth]

On 07/07/2017 01:58 PM, FELIPE ORELLANA ROVIROSA wrote:
> 
>     In that spirit, the variable u could be Temperature, or velocity 
> components (or the strainrate), etc. Then I guess the results or effectiveness 
> of this T-based mesh refinement would depend on what element order is in use, 
> cause their sensitivity to the second derivative is different.

Potentially. At least in theory, it would make sense to use third derivatives 
for quadratic elements (as we use for velocity and temperature) but in 
practice, the general observation in the computational science community is 
that using second derivatives leads to pretty good meshes.

Best
  W.

-- 
------------------------------------------------------------------------
Wolfgang Bangerth          email:                 bangerth at colostate.edu
                            www: http://www.math.colostate.edu/~bangerth/