Thank you very much for the pointers, Daniel!
- Including the templates header did the trick. I no longer need to copy
the data over
- I was able to implement the penalty method and the DoF constraint method;
the subspace projection method seems unambiguously best in terms of
accuracy, iteratio
Corbin,
- Using ARPACK might indeed be an option. Do you really need to know the
whole spectrum, though?
- Yes, you would just call AffineConstraints::add_line() to add constrain a
given DoF to zero. Note that it's a no-op if the DoF is already
constrained. This doesn't change the size of the line
I'm working on a similar pure Neumann problem with a rank-1 deficiency
(pressure known only up to a constant). I've implemented the subspace
projection method Konrad mentioned, with good results. However, I'm
interested in comparing it to the single DoF constraint method, as well as
an alternat
On 12/26/20 3:06 AM, Konrad Simon wrote:
What one can also do is just constrain one DoF to a specific value (this would
also remove rigid motion in elasticity). But think about your solution
variable: If it is in the Sobolev space H^1 then point evaluations may not be
defined for dimension larg
Thanks, I'll give it a try!
On Saturday, 26 December 2020 at 08:45:04 UTC-5 Konrad Simon wrote:
> Little correction: I wrote "In the vmult() function remove the mean value
> (i.e., project the rhs on the orthogonal complement of the kernel of the
> kernel)", I meant "In the vmult() function rem
Little correction: I wrote "In the vmult() function remove the mean value
(i.e., project the rhs on the orthogonal complement of the kernel of the
kernel)", I meant "In the vmult() function remove the mean value after you
multiply (i.e., project the rhs on the orthogonal complement of the kernel
Hi,
On Saturday, December 26, 2020 at 6:43:21 AM UTC+1 smetca...@gmail.com
wrote:
> Hi all,
>
> Does anyone here have any experience applying mean value constraints
> (specifically with periodic boundary conditions)? I'm having some trouble.
> As far as I can tell, there are two approaches (bo