Add Raviart-Thomas grad-div example#4061
Conversation
| // | ||
| // This example uses Raviart-Thomas elements to solve a model grad-div problem | ||
| // in H(div) in both 2d and 3d: -\nabla (\nabla \cdot \vec{u}) + \vec{u} = f. | ||
|
|
There was a problem hiding this comment.
Would you mind adding some comments about how this example compares/contrasts with vector_fe_ex6? Obviously they are solving different PDEs, but e.g. ex6 uses a "mixed" (u,p) formulation, whereas in this case we don't do that. I guess this information is fairly basic, but since the examples are meant for explanation/teaching, I think it would be appropriate to include additional information.
There was a problem hiding this comment.
Yeah, that's an improvement... I was also just wondering what practical applications are related the grad-div problem? Obviously div-grad (aka Poisson) has many applications in physics so we don't need to elaborate on them, but I have never run into grad-div before. If it's just a standard/convenient test for RT elements, that's fine too of course.
There was a problem hiding this comment.
Yeah, that's an improvement...
I realise I've added more to the description of ex6 than to ex10, but if the point is distinguishing them, I hope that works?
If it's just a standard/convenient test for RT elements, that's fine too of course.
Yeah, it's to H(div) what the curl-curl problem is to H(curl), i.e. it's the problem whose differential operator naturally realises the inner product on H(div).
|
This already passed before the comment update so I'm just going to merge without waiting for the latest round of testing to finish. |
4 participants
The (real) motivation is that of other PRs, to be opened soon.
However, since this example is self-contained and should have some value in and of itself, I thought I'd alleviate those other PRs of these changes (including the always massive re-bootstrap and new html).