On Jul 12, 2011, at 3:06 AM, Marc Saudreau wrote:
> You are right, by using the DiffusionTermNoCorrection the system converges.
> Unfortunately, and that's what we could expect with skewed cells, the solution
> I get is not really the right solution.
To what extent is not not right? Is it off by a little or is it completely
wrong? I ask, because in your original email, you said
>> I've decided to try a more simple case: adiabatic conditions on lateral and
>> bottom faces, and T = Tair on the top face
but the code you sent had
UpBC = FixedValue(faces=maskUpSurface, value= Tair)
LowBC = FixedValue(faces=maskLowerSurface, value= Tair)
> I guess there is no solution unless I can use GMSH meshes composed of
> quadrangles and hexahedrons !
> Do you plan to make this available in a next release of Fipy ?
The next release has support for quadrangles and hexahedrons.
