But thinking about it, what I imagine you want to say is not that the
heat flux is discontinuous, but that the gradient of the temperature is
discontinuous, right?
You're completely right. Sorry for not saying what I wanted to say, and
thanks for guessing what I wanted to say from what I said!
But so coming back to your original question: is the heat flux or the
temperature gradient one of your variables that you need to discretize?
Because if it is the former, you don't need discontinuous elements (the
Raviart-Thomas element would be the right choice) whereas if it's the latter
then I think you probably aren't using the best formulation.
Well, in a first time, I have a simple "bi-material" problem: at the
interface, the jump in temperature and normal flux must be zero, so the
jump in the normal temperature gradient must be non-zero.
And in a second time, the interface will be moving and some heat will be
generated, so neither the jump in the normal temperature gradient nor
the jump in the normal flux will be null.
I will definitely have a closer look at the formulation of
Raviart-Thomas elements (sorry for my very limited knowledge).
Many thanks for your help,
Martin.
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
