ffc team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Mon, Feb 16, 2009 at 05:47:28PM -0600, Brandon Chabaud wrote:
> Hello,
> 
> Here is the problem that I'm having.  I want to implement a discontinuous
> Galerkin method for a 2D elasticity problem.  The boundary is divided into two
> subsets: Gamma_0 and Gamma_1.  One of my unknowns is the vector u, the
> displacement.  Boundary conditions are imposed weakly.  In my formulation of
> the method, I have a term of the form
> 
> <u_i*n_j,v_i*n_j>_{Gamma_0},
> 
> where n is the facet normal and < , >_{Gamma_0} means the integral over
> Gamma_0.  Using the built-in exterior facet integral measure *ds in my .form
> file will cause the compiler to integrate the term over both Gamma_1 and
> Gamma_0.  How do I restrict the integration to Gamma_0 only?  Thanks.
> 
> Brandon Chabaud

Use ds(0), ds(1) etc to specify integration over sub sets of the
boundary. Then you need matching mesh markers (MeshFunction over the
facets) to indicate the different sub sets when assembling.

-- 
Anders

Attachment: signature.asc
Description: Digital signature