On Fri, Nov 21, 2014 at 4:42 PM, Kyle Briton Lawlor <[email protected]>
wrote:
>
> Hi, Jonathan and FiPy.
>
> Thanks! This really helped. I can now implement some various Neumann BCs.
Very cool.
>
> var will need to be a FaceVariable, but yes, you should be able to do this
>
> Do I need to declare a CellVariable and a FaceVariable?

I'm confused by the code above, "phiBoun" is created but never used. In
general FaceVarialbes are used for coefficients and CellVariables are used
for solution variables. I think "phi.faceValue" is all you need.

> Or do I need to use a FaceVariable instead of CellVariable?

You definitely need a CellVariable if you want to solve anything.
"phi.faceValue" is a FaceVariable. I don't think you need anything beyond
that.

>
> I am confused with why I need to do this...
>
> You mean create a FaceVarialbe?

I don't think you do need to, you've created one, but it's not used so your
code would run equally well without it. There is no black magic in FiPy.


>  Is this interpretation correct?
> #ROBIN BC
> phi.faceGrad.constrain([phi.faceValue*mesh._faceNormals],where=outerFaces)
> phi.faceGrad.constrain([1*mesh._faceNormals],where=innerFaces

I don't believe that you need to feed a list into the value for the
boundary condition. Does it work? If it doesn't work, I just answered a
question with a similar issue and implemented an implicit version of a
Robin boundary condition just using source terms. It might be an easier
approach if you can't get the constraints to work correctly.

http://nbviewer.ipython.org/gist/wd15/f04b22c6ec80b1eda8e3

>
> Does using a FaceVariable mean the value of phi.faceValue would change?

Yes. It should be automatically updated as phi changes.

> I just want to say, thanks for all the help so far! I’ve been learning a
lot, so I really appreciate it!
> Look forward to hearing more.

No problem at all. I'm just sorry that I answer questions so slowly and
can't spend enough time making FiPy better.

Cheers,

Daniel

-- 
Daniel Wheeler
_______________________________________________
fipy mailing list
[email protected]
http://www.ctcms.nist.gov/fipy
  [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]

Reply via email to