Hi Joshua,

yes D is fixed, and the value on the polygon is always positive; Indeed,
here is what I do;

I have an old concentration field at one time step "OLDC" (all concentration
is between 0 and 1). then I update the geometry of the polygon (it increases
its size a little bit)

I create a new mesh, and a new variable "NEWC" . I put the value 0. at the
creation of NEWC, and then I copy the values of OLDC in NEWC.
so, so far no negative values.

and then I iterate one timestep, the diffusion equation on NEWC.  and
sometimes, I get negative values !!! or values > 1




On Tue, Feb 15, 2011 at 3:33 PM, Joshua Bush <[email protected]> wrote:

> Julien,
>
>     Would you mind giving us more information?
> Is D a fixed value? What is the center polygon's initial concentration
> value?
> On Feb 15, 2011 6:29 AM, "Julien Derr" <[email protected]> wrote:
> > Hi all,
> >
> > I am still bothering you with my simple diffusion problem between an
> outer
> > circle (Boundary condition value=1)
> > and an inner polygon (Boundary flux=0)
> >
> > I am getting negative values on the concentration field !!
> > a simple diffusion process ( eq = TransientTerm() ==
> > DiffusionTerm(coeff=D)) should not allow that , no ?
> >
> > even with a fixed flux boundary ?
>

Reply via email to