Dear Arjun, The ordering of the extracted data should align with how the components are internally unrolled. Both the Tensor and SymmetricTensor classes have two functions to assist with this type of operation, namely unrolled_to_component_indices <https://www.dealii.org/developer/doxygen/deal.II/classSymmetricTensor.html#a4c052ca00daee61e07cec3b550bda6eb> and component_to_unrolled_index <https://www.dealii.org/developer/doxygen/deal.II/classSymmetricTensor.html#a8a3806a78029858205695c5abb1d2f72>, as well as the static member data n_independent_components <https://www.dealii.org/developer/doxygen/deal.II/classSymmetricTensor.html#adf501ef6f23cafa2ff79081afbd8478d> for help looping over the correct number of components that each structure holds. In particular, the TableIndex <https://www.dealii.org/developer/doxygen/deal.II/classTableIndices.html> object returned by unrolled_to_component_indices <https://www.dealii.org/developer/doxygen/deal.II/classSymmetricTensor.html#a4c052ca00daee61e07cec3b550bda6eb> will tell you exactly which stress/strain component you're interrogating.
I hope that this helps! Regards, J-P On Tuesday, November 29, 2016 at 3:48:03 AM UTC+1, Arjun Narayanan wrote: > > Hello everyone, > > I am trying to follow Wolfgang's suggestion here ( > https://groups.google.com/forum/#!topic/dealii/KrSp6k0FSx8) to implement > a mixed formulation for elasticity. I'm experimenting with the Hu-Washizu > 3-field variational form. > > The most straightforward approach appears to be to declare dim basis > functions for displacements, and dim*(dim+1)/2 basis functions for > strains and stresses respectively. > > I am not exactly sure how the FEValuesExtractors::SymmetricTensor orders > the elements it extracts into the symmetric tensor object. > > For instance, for a rank 2 tensor, does it assume that the first dim entries > are the diagonal entries, and the next dim*(dim-1)/2 entries are the off > diagonals? > As an example for the stresses: > > [sigma_11, sigma_22, sigma_33, sigma_23, sigma_13, sigma_12] > > I believe this is the typical Voigt notation rules, but I'm not sure if > dealii does it this way. I did try to look into the documentation but > couldn't quite put my finger on it. > > Thanks for any inputs! > > Arjun > -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
