Thanks Daniel!

On Mon, Oct 19, 2009 at 12:02 PM, Daniel Wheeler
<[email protected]> wrote:
>
> On Mon, Oct 19, 2009 at 12:46 AM, weasky <[email protected]> wrote:
>>
>> Hi,
>>
>> I noticed there are a couple of examples in the folder about
>> cylindrical cords, and my question is about the the BCs:
>> mesh.getFacesLeft(0) is the boundary for r=r_{inner}
>> Right would be r_{outer}
>> top would be z_{highest}
>> bottom would be z_{lower}
>>
>> Is my understanding correct?
>
> Yes, perhaps we should add in stub methods that are named something
> more appropriate (inner, outer, upper,lower???).
>
>> Basically, I want to solve a 2D diffusion
>> eqn in a geometry of (r,z) and use the antisymmetry of the center
>> line.
>
> Antisymmetry? I'm not sure exactly what you mean. Is that just a typo?

Sorry, I mean axial symmetry. Do I need to add a no flux BC for that?
I think if I do nothing ,Fipy will automatically give me that BC then.

>> so for the BC part I will do
>>  >>> boundaryConditions = (
>>     ...  FixedFlux(faces=mesh.getFacesLeft(), value=0), #center-line, flux 
>> is 0
>> FixedValue(faces=mesh.getFacesRight(),value=valueRight),
>> FixedValue(faces=mesh.getFacesTop(),value=z1),
>>
>> FixedValue(faces=mesh.getFacesBottom(),value=z0),
>>      ...     )
>>
>> is my understanding correct?
>
> I think that is all good, but I don't think you need a BC for the left
> faces. Their surface area vanishes.
>
> I have had issues with the corner cells for flow problems, I shift the
> mesh away from 0 by some very small amount. You may or may not need to
> this. I think it is only for flow problems that I found this to be
> necessary.

Thanks! I may try this and see the differences. BTW: is there a way to
visualize the mesh? or export the mesh to other format (such as
tecplot)? or better to visualize the mesh in 3D cylinder?

Thanks a lot for this great app!

Tim
>
> Cheers
>
> --
> Daniel Wheeler
>
>


Reply via email to