On Fri, 23 Apr 2010, Liang wrote:
> Is that possible to dump the flux value at each node? Taking the thermal
> problem
> at ex10, the independent variable u (Temperature) could be generated in the
> postprocessing data, while how to obtain the other dependent variable,
> such as
> flux(gradient of the temperture).
>
> I have been thinking two methods without sure to solve this.
>
> One method is deriving the mixed weak form and the flux is the second
> variable, then the flux could be dumped as the independent
> variables. The cost is having to derive the mixed form and may
> leading to the instability if the constrains are ill set.
Yes, this would work, but it's overkill if you're just looking for the
output and don't have other reasons for solving with a mixed form.
> Another method is adding flux variables directly with the function
> system.add_variable ("gradT_x", FIRST), then plugging additional flux
> computing codes code in the element part.
This is almost right. Instead of adding an additional variable to
your system (which would then have to be taken into account in your
implicit solves), create a separate ExplicitSystem with a variable
there to store the data. Then write a postprocessing loop that
computes the flux of your main system and sets it into those variables
in the secondary system. If you're using FIRST order variables and
you want the same approximation for your fluxes look at
Zienkiewicz-Zhu.
---
Roy
------------------------------------------------------------------------------
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users
