Raviart-Thomas elements not correct on manifold

Running the following test:

mesh2d = UnitSquareMesh(2, 2)

# Convert 2D mesh to 2D mesh in 3D space
mesh3d = Mesh()
e = MeshEditor()
e.open(mesh3d, 2, 3)
e.init_vertices(mesh2d.num_vertices())
for (i, x) in enumerate(mesh2d.coordinates()):
    e.add_vertex(i, x[0], x[1], 0.0)
e.init_cells(mesh2d.num_cells())
for (i, cell) in enumerate(mesh2d.cells()):
    e.add_cell(i, cell[0], cell[1], cell[2])
e.close()

# This makes no difference here
mesh3d.init_cell_orientations(Constant((0, 0, 1)))

mfile2 = File("mesh2d.xdmf")
mfile2 << mesh2d

mfile3 = File("mesh3d.xdmf")
mfile3 << mesh3d

Q2D = FunctionSpace(mesh2d, "RT", 1)
Q3D = FunctionSpace(mesh3d, "RT", 1)
F2D = Function(Q2D)
F3D = Function(Q3D)
F2D.interpolate(Constant((0.0, 1.0)))
F3D.interpolate(Constant((0.0, 1.0, 0.0)))

print F2D.vector().array()
print F3D.vector().array()

v2D=F2D.compute_vertex_values().reshape(2,-1).transpose()
v3D=F3D.compute_vertex_values().reshape(3,-1).transpose()

print v2D
print v3D

output2d = File("f2d.xdmf")
output2d << F2D

output3d = File("f3d.xdmf")
output3d << F3D

results in spurious output. Both should be the same, as the only difference is the geometric dimension of the Mesh.

[ 0.  -0.5 -0.5 -0.5  0.   0.  -0.5 -0.5 -0.5 -0.5  0.   0.  -0.5 -0.5 -0.5
  0. ]
[-0.   0.5  0.5 -0.5 -0.  -0.   0.5  0.5 -0.5  0.5  0.  -0.  -0.5  0.5 -0.5
  0. ]
[[ 0.  1.]
 [ 0.  1.]
 [ 0.  1.]
 [ 0.  1.]
 [ 0.  1.]
 [ 0.  1.]
 [ 0.  1.]
 [ 0.  1.]
 [ 0.  1.]]
[[ 0.  1.  0.]
 [ 0.  1.  0.]
 [ 1.  0. -1.]
 [ 0.  1.  0.]
 [ 0.  1.  0.]
 [ 1.  0. -1.]
 [ 1.  0.  1.]
 [ 1.  0.  1.]
 [ 0.  0.  0.]]

Comments (8)