BoundaryMesh projection

Hi,

Not sure what's the best way to do this, but here's a hard and not so
elegant way:-)

from dolfin import *

mesh = UnitSquareMesh(4,4)
V = FunctionSpace(mesh,'CG',1)

mesh_e = BoundaryMesh(mesh,'exterior')
V_e = FunctionSpace(mesh_e,'CG',1)
v_e = interpolate(Expression("x[0]"), V_e)

v = TrialFunction(V)
u = TestFunction(V)

V_dofmap = V.dofmap()
u = Function(V)
for facet in facets(mesh):
    if facet.exterior():
        c = Cell(mesh, facet.entities(2)[0])
        local_facet = V_dofmap.tabulate_facet_dofs(c.index(facet))
        facet_coors = V_dofmap.tabulate_coordinates(c)[local_facet]
        facet_dofs = V_dofmap.cell_dofs(c.index())[local_facet]
        for x, dof in zip(facet_coors, facet_dofs):
            u.vector()[dof] = v_e(x)

plot(u)

Mikael

On 14 March 2013 14:11, Miro <email address hidden> wrote:

> New question #224282 on DOLFIN:
> https://answers.launchpad.net/dolfin/+question/224282
>
> Hi everyone,
>
> I'd like to project a function defined on a function space over a boundary
> mesh to a function space defined over the full mesh.
> I thought this could be achieved by integrating over the boundary of the
> full mesh but I get 'matrices are not alligned error '.
> The error can be reproduced by the code below.
>
> from dolfin import *
>
> mesh = UnitSquareMesh(4,4)
> V = FunctionSpace(mesh,'CG',1)
>
> mesh_e = BoundaryMesh(mesh,'exterior')
> V_e = FunctionSpace(mesh_e,'CG',1)
> v_e = Function(V_e)
>
> v = TrialFunction(V)
> u = TestFunction(V)
>
> L = v_e*u*ds
> assemble(L)
>
> In terms of math, I think the integral makes sense, as both v_e and u are
> defined over the domain of intergration. Is there a
> way to achieve this? Thanks for help, Miro
>
> --
> You received this question notification because you are a member of
> DOLFIN Team, which is an answer contact for DOLFIN.
>