Grid refinement for time dependent problem

[Yiyang Zhang]

Hello,

This might be a quite general question.

When we run a time-dependent problem, the solution might "move around", thus we need to use grid refinement as time goes on. However, I do not want the total DoF keep increasing. Ideally, the total DoF should fluctuate within some range that the computer can deal with. Usually how do people do with such problem?

Thanks!

[Wolfgang Bangerth]

On 09/27/2017 02:15 PM, Yiyang Zhang wrote:
>
> When we run a time-dependent problem, the solution might "move around",
> thus we need to use grid refinement as time goes on. However, I do not
> want the total DoF keep increasing. Ideally, the total DoF should
> fluctuate within some range that the computer can deal with. Usually how
> do people do with such problem?

Many of the GridRefinement functions allow you to specify an upper bound
for the number of cells they produce. Alternatively, you can cap the
level of refinement -- that's what step-26 and 32 do, for example.

Best
W.


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

[Yiyang Zhang]

Hello Prof. Bangerth,

Thank you for your answer!

It turns out I asked a stupid question.. I never noticed this parameter...

Best,

Yiyang