Add a prestress contribution to stress tensor results output

[ktbolt]

Problem

The output of stress tensor results does not currently add a prestress contribution to the 2nd Piola-Kirchhoff stress tensor.

Solution

Add a prestress contribution to the 2nd Piola-Kirchhoff stress tensor for the output of stress tensor results.

Viscous stress also not currently added so add that too.

Additional context

No response

Code of Conduct

• I agree to follow this project's Code of Conduct and Contributing Guidelines

[divya-adil]

It seems like the stress used during assembly and the stress written to the results were not being computed in the same way.

During structural assembly, the code computes the 2nd Piola-Kirchhoff stress with mat_models::compute_pk2cc(...), then adds viscous stress and prestress before using the stress in the residual (see sv_struct.cpp, lines 666-687).

In post.cpp, we are recomputing S with mat_models::compute_pk2cc(...) and then writing that value to the output stress field. This means that the reported stress did not include prestress. The same section also had TODOs where viscous stress should be added.

I am updating post.cpp so the reported stress follows the assembly path more closely. Prestress is already stored in com_mod.pS0, so I interpolate those nodal values to the current Gauss point before adding them to the reported stress tensor.

Viscous stress is not stored, so I recompute it at the Gauss point using the existing assembly routine mat_models::compute_visc_stress_and_tangent(...).

See my branch for changes.

There aren't any test cases with prestress with struct/ustruct - do you have any ideas on how to test this @ktbolt ?

[ktbolt]

@divya-adil The CMM pipe_3d CI test computes and uses Prestress_file_path to read in the prestress results but it does not output stress.

We should add a prestress CI Test for at least ustruct.

@dseyler Do you have a simple example for this ?

[dseyler]

We should add a prestress CI Test for at least ustruct.

@dseyler Do you have a simple example for this ?

I don't. Maybe one from the passive struct / ustruct LV cases

[dseyler]

@divya-adil Thanks for addressing this!

Just a heads up, my open PR will impact the way the viscosity calculation is called so may collide with your branch.

I also think it would be nice to be able to output viscous stress as its own field for debugging

[divya-adil]

@dseyler thanks for letting me know! I will wait for the PR to be merged and update my code before testing it.

@ktbolt I could add a cylinder prestress case as described in issue #296. We tested it but never added the test case.