Markus,

Thank you so much! This fixed my problem perfectly--everything is working beautifully! Thank you!

Best,
-J

PS: I sent the previous reply to this thread while the mail list was down. I didn't know that you had replied already!

On 04/09/2012 10:02 AM, Markus Bürg wrote:
Hello Justin,

the problem are the boundary values. In your code you set for every exterior face its first dof to the midpoint value of the boundary function on this face. This is just fine for lowest order elements (degree 0), because there the degrees of freedom are defined this way (i.e. choosing the midpoint value). However, for elements of degree one we have two degrees of freedom per face. You set the first one to the midpoint value again (line 320). This is correct, but you do nothing for the second one. To make your programme work correctly you have to set this degree of freedom, too.

Anyway, you do not have to write a function for setting the boundary values by yourself. There is already one in the namespace VectorTools: VectorTools::project_boundary_values_curl_conforming.

Best Regards,
Markus



On 05.04.2012 22:38, Justin Droba wrote:
On 04/05/2012 02:17 PM, Markus Bürg wrote:
Hello Justin,

I didn't know there was one already implemented! (It's not in the FEValues class, right?) Does it compute the curl differently?
No, it is in the FEValuesExtractors::Vector namespace.

Could you send me a minimal version of your code which reproduces the problem? Then I will have a look and check what is going wrong.


Markus,

Thank you so much! I copied the code of my implementation below. It's currently set to order = 1 (this can be adjusted in line 469 in the main function). The action occurs in assemble_system_chunk (this code is multithreaded), starting on line 329. Assembly of the system matrix is centered around line 370 or so.

I am now using the curl function from FEValuesViews as suggested.

Thanks for your help!

Best,
-J
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

Reply via email to