How to plot just the values in a SubDomain?

Question

Hello,

I'm working on a mixed problem in which the values of the variables are in different order of magnitude for different subdomains so it is not possible to visualize the lower values with regular plot.

I would like to know if there is a easy way to plot just the values of a single subdomain. For example, if I can use SubDomain objects marked by CellFunction to select which points should be ploted.

Thanks in advance,
Best regards

Answer 1

Also you can "extract" your solution only at the dofs that belongs to a given cell. Something like this:

from dolfin import *

mesh = UnitSquareMesh(100, 100, "crossed")
V = FunctionSpace(mesh, "CG", 2)
u = interpolate(Expression("x[0]*x[0] + x[1]*x[1]"), V)

class half(SubDomain):
  def inside(self, x, on_boundary):
    return x[0] > 0.5

cell_f = CellFunction("size_t", mesh)
cell_f.set_all(0)
half = half()

# mark subdomains
for cell in cells(mesh):
  if half.inside(cell.midpoint(), True):
    cell_f[cell.index()] = 1

def extract_values(u, cell_function, subdomain_id, V):
  dofmap = V.dofmap()
  mesh = V.mesh()
  for cell in cells(mesh):
    # Preserve only the dofs in cells marked as subdomain_id
    if cell_function[cell.index()] != subdomain_id:
      dofs = dofmap.cell_dofs(cell.index())
      for dof in dofs:
        u.vector()[dof] = 0.0
  return u


u_left  = Function(V); u_left.vector()[:]  = u.vector()
u_right = Function(V); u_right.vector()[:] = u.vector()
u_left  = extract_values(u_left, cell_f, 0, V)
u_right = extract_values(u_right, cell_f, 1, V)

plot(u_left, title="left")
plot(u_right, title="right")
interactive()

Comments

Thank you! This really solved my problem.

---

For those interested in extracting the solution on a set of nodes defined as a facet mesh function, see MiroK's answer (thanks again to him) to this question: https://fenicsproject.org/qa/11876/extract-solution-at-a-set-of-nodes. Martin

Answer 2

Unfortunately, this is not possible today. However, you can adjust the value range plotted like this:

plot(u, range_min=0., range_max=1.)

Comments

Thank you for attention! But this doesn't remove the undesired points. It just set the value of them to range_max when they are greater than range_max.