dolfin team mailing list archive

["Johan Hoffman" <jhoffman@xxxxxxxxxxx>]

> On Fri, Dec 09, 2005 at 03:11:28PM +0100, Johan Hoffman wrote:
>
>> >>  2. Fix Navier-Stokes module?
>> >
>> > Are we still waiting for a fix of the Navier-Stokes module, or can we
>> > release anyway tomorrow Friday (and add a comment that the
>> > Navier-Stokes module is broken)?
>>
>> We are still waiting for the fix... I have been away a few days, but I
>> hope to fix it early next week (depending on how nasty the bug is).
>> Sure,
>> we can release if we really want to, but I think it would be better to
>> wait until this is fixed.
>>
>> /Johan
>
> ok, I can wait.
>
> But maybe if it takes more than a day to fix the Navier-Stokes module,
> we can release anyway early next week. I have some things lined up
> that I want to wait adding until after the release. And also, the
> latest version of DOLFIN does not work with the latest version of
> FFC so we should release pretty soon.
>
> /Anders
>

Ok. I've had a quick look at it, and it seems the fix-point iteration does
not converge. The criterion is that k/delta should be sufficiently small
(with delta the stabilization parameter). So typically one can fix by
reducing k or increase delta, but that did not seem to fix it. Another
common problem is that the boundary conditions are not applied correctly.

I have not installed Paraview on this computer yet so I could have a look
at the solution. I'll try again tomorrow.

/Johan