Question #105304 on DOLFIN changed:
https://answers.launchpad.net/dolfin/+question/105304
Status: Open => Answered
Andy R Terrel proposed the following answer:
I believe this is the paper describing the methods used in the demo.
http://dx.doi.org/10.1016/j.jcp.2006.03.010
The split is done to separate the equations based on the Mixed Space.
This comes from the UFL specification so the manual there could give
you more information. It is written this way so that a and L can be
automatically linearized for the Newton-Raphson method.
The CahnHillardEquation class is a NonlinearProblem. This is the
simple way to use Newton-Raphson method that is coded in dolfin. In
this style you need to specify the application of F and its jacobian
J. The easiest way to do that is to define a form and have UFL/FFC
generate the derivatives for you. The newon solver will reconstruct F
and J as it needs to for the Newton-Raphson method.
I hope that helps.
-- Andy
On Tue, Mar 23, 2010 at 10:53 AM, cutejeff
<[email protected]> wrote:
> New question #105304 on DOLFIN:
> https://answers.launchpad.net/dolfin/+question/105304
>
> There is a demo code under src/demo/pde/cahn-hilliard/ to solve the
> Cahn-Hilliard equation, which is a fourth-order nonlinear time-dependent
> problem. Since my problem is also nonlinear initial value problem, I want to
> follow this code as an example.
> I have gone through the nonlinear and time-dependent prob parts in the
> tutorial. But I still have trouble to understand the approach the author
> did. He seems to split the original eqn into 2 eqns. Does anybody know the
> original eqn and the derivations?
>
> I also have trouble to understand the following code:
> dk, dc = split(du)
>
> class CahnHilliardEquation(NonlinearProblem):
> def __init__(self, a, L):
> NonlinearProblem.__init__(self)
> self.L = L
> self.a = a
> self.reset_sparsity = True
> def F(self, b, x):
> assemble(self.L, tensor=b)
> def J(self, A, x):
> assemble(self.a, tensor=A, reset_sparsity=self.reset_sparsity)
> self.reset_sparsity = False
>
> They haven't been covered in the tutorial. Are there any other documents to
> learn these?
> Any help will be appreciated.
>
> --
> You received this question notification because you are a member of
> DOLFIN Team, which is an answer contact for DOLFIN.
>
> _______________________________________________
> Mailing list: https://launchpad.net/~dolfin
> Post to : [email protected]
> Unsubscribe : https://launchpad.net/~dolfin
> More help : https://help.launchpad.net/ListHelp
>
You received this question notification because you are a member of
DOLFIN Team, which is an answer contact for DOLFIN.
_______________________________________________
Mailing list: https://launchpad.net/~dolfin
Post to : [email protected]
Unsubscribe : https://launchpad.net/~dolfin
More help : https://help.launchpad.net/ListHelp
