dolfin team mailing list archive

[Martin Sandve Alnæs <martinal@xxxxxxxxx>]

Hi Gerard.
Use index notation instead of the grad operator:

(u.dx(i)*n[i]).dx(j)*n[j]

And that should work fine.

Mixing of "compound operators" that create expressions
with a tensor shape, with index notation which deals with
"sets of scalars", will not always work. Some times it is
ambiguous. In this case, should it work? Could be a bug, not sure.

Martin

On 19 May 2011 00:35, Gerard Awanou
<question158115@xxxxxxxxxxxxxxxxxxxxx> wrote:
> New question #158115 on DOLFIN:
> https://answers.launchpad.net/dolfin/+question/158115
>
> I am running the Fenics demo written for C^0 Discontinuous Galerkin
> method for the biharmonic equation. I understood that the demo is for
> the boundary condition \Delta u =0. For a boundary condition \partial
> u/\partial b =0, it seems that one should use the formulation in
>
> C0 Interior Penalty Methods for Fourth Order Elliptic Boundary Value
> Problems on Polygonal Domains by Brenner and Sung
>
> To do this I need to compute n'(D^2 u)n (p. 100 of the paper).
> I tried n[i]*( grad(grad(v))[i,j]*n[j]) without success.
> How do you compute \partial^2 u/\partial n^2  where n is the unit
> outward normal to an edge?
>
> --
> You received this question notification because you are a member of
> DOLFIN Team, which is an answer contact for DOLFIN.
>
> _______________________________________________
> Mailing list: https://launchpad.net/~dolfin
> Post to     : dolfin@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~dolfin
> More help   : https://help.launchpad.net/ListHelp
>