Thank you Wolfgang.  I think having an extension that will return zero if it is 
outside the old mesh would be the easiest to implement.  I will have a go at 
writing this over the next few weeks for a test problem with a known solution, 
and see where I get.

Thanks for your help.

Katie


Katie Leonard

DPhil student in Computational Biology,
The University of Oxford.
________________________________________
From: Wolfgang Bangerth [[email protected]]
Sent: 18 June 2012 11:47
To: Katie Leonard
Cc: [email protected]
Subject: Re: [deal.II] A moving boundary problem

> My problem is that I need to interpolate the solution onto the new
> mesh - but the two meshes don't necessarily have the same boundary
> (consider a domain that is advecting with constant velocity, for
> example, in which case the two meshes would overlap) and then I don't
> think I can use FEFieldFunction?

Possibly. Like I said, the question is what you want this class to do if
you interpolate at a point that happens to be outside the old mesh. I
don't know what you need in this case -- continue the function by zero?
continue it as a constant value on radial lines away from the domain?
take the value from the closest point inside the domain? throw an error?
It all depends on what you want to do.

If you're ok with extending the function by zero outside the old mesh,
then it should be easily possible to extend the class in a way that
allows for that, and writing such an extension would likely be the
easiest way.

Best
  W.


------------------------------------------------------------------------
Wolfgang Bangerth               email:            [email protected]
                                 www: http://www.math.tamu.edu/~bangerth/

_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

Reply via email to