Deformation gradient at face quadrature point for current traction

48 views
Skip to first unread message

Krish

unread,
Jun 26, 2023, 2:49:53 AM6/26/23
to deal.II User Group
Hello,

I am following step-44 to solve an elasticity problem. I am trying to implement current traction (as opposed to the dead load used in step-44) on a part of the boundary. For that I need to use the deformation gradient and its determinant at the f_q_point(s) (face quadrature points). I am interested to know what might be a suitable way to get these.

I have observed PointHistory being used to store and use similar things for cell quadrature points. Is there anything similar for the face quadrature points?

Thanks,
Krish

Wolfgang Bangerth

unread,
Jun 28, 2023, 1:55:19 PM6/28/23
to dea...@googlegroups.com
On 6/26/23 00:49, Krish wrote:
>
> I am following step-44 to solve an elasticity problem. I am trying to
> implement current traction (as opposed to the dead load used in step-44)
> on a part of the boundary. For that I need to use the deformation
> gradient and its determinant at the f_q_point(s) (face quadrature
> points). I am interested to know what might be a suitable way to get these.

If you have the deformation field as a solution vector, then computing
its gradient at quadrature points on faces is easily done using
FEFaceValues. Take a look at the computation of "fluxes" here:
https://dealii.org/developer/doxygen/deal.II/step_4.html#PostprocessingWhattodowiththesolution
In your case, the gradients are going to be tensors, and you will
compute them using code such as

std::vector<Tensor<2, dim>>
solution_gradients(face_quadrature_formula.size());

FEValuesExtractors::Vector deformation(...);
...

fe_face_values[deformation].get_function_gradients(solution,
solution_gradients);

There is a free function called 'determinant()' that can be used to
compute the determinant of a dim x dim tensor.

Best
W.

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

Abbas Ballout

unread,
Jul 11, 2023, 10:34:42 AM7/11/23
to deal.II User Group
Krish,

I could do this following step 18 and how history variables are constructed there.
It was straightforward in case you are interested.   



Reply all
Reply to author
Forward
0 new messages