Applying dirichlet condition using the class of nonlinear problem

Question

Hi! I'm trying to solve a non linear problem following the cahn hilliard demo but I have a system made up by three equations. I have dirichlet conditions for the first and the third variables.

mesh=Mesh("brain.xml")
V    = FunctionSpace(mesh,"Lagrange",1)
ME   = MixedFunctionSpace([V,V,V])

g_c  = Constant(0.0)
bc_c = DirichletBC(ME.sub(0), g_c, DirichletBoundary())
g_n  = Constant(1.0)
bc_n = DirichletBC(ME.sub(2), g_n, DirichletBoundary())
bcs = [bc_c, bc_n]

If I define the non linear problem as

class MyTumor(NonlinearProblem):
def __init__(self, L, a, bcs):
  NonlinearProblem.__init__(self)
  self.L = L
  self.a = a
  self.bcs = bcs
def F(self, b, x):
  assemble(self.L, tensor=b, bcs=self.bcs)
def J(self, A, x):
  assemble(self.a, tensor=A bcs=self.bcs)

It gives me annoying warnings about applying manually the BC.

If I apply them manually :

class MyTumor(NonlinearProblem):
def __init__(self, L, a, bcs):
  NonlinearProblem.__init__(self)
  self.L = L
  self.a = a
  self.bcs = bcs
def F(self, b, x):
  assemble(self.L, tensor=b)
  self.bcs.apply(b,x)
def J(self, A, x):
  assemble(self.a, tensor=A)
  self.bcs.apply(A)

It gives me this error:
AttributeError: 'list' object has no attribute 'apply'

I recall the problem with

problem = MyTumor(L,J,[bcs])

How can I apply my Dirichlet conditions ??

Answer 1

Hi, apply is a method of the DirichletBC class and not the list class. To apply each condition from a list of DirichletBC objects the call could be for example

bcs = [bc_c, bc_n]
[bc.apply(A) for bc in bcs] 

Note that [bcs] is a list with one list object.

Comments

Is there any different between

problem = MyTumor(L,J,[bcs])

and

problem = MyTumor(L,J,bcs)

?
Thanks again!

---

It is your own class so it is only up to you how the function arguments are handled. But in my opinion there is no need to nest the lists, that is, I'd prefer MyTumor(L,J,bcs).

---

ok.. thanks a lot!!!