Thanks Thomas!
It works now, but I suggest to add some clarification about this subtle
point to the tutorial.
In Step-1 it is stated:
"By default, all boundary parts have this number, but you can change
this number for some parts of the boundary..."
So there is nothing about disjoint boundary parts...
Moreover, you can solve Poisson equation on 2D hyper_shell - it also has
2 disjoint boundaries, but in this case both boundaries get the same 0
boundary indicator,
and everything works fine.
So that I still do not understand the peculiarity of 1D case.
Ivan
On 09.03.2011 21:19, Thomas Wick wrote:
Hi Ivan,
In 1D, the boundary is not any more connected like in 2D or 3D. Therefore,
you have to pass two boundary colors to the interpolate_boundary_values
function.
Thomas
P.S. You will get the same "errors" in step-3 and step-5, too.
Hi guys,
I'm passing through your very nice tutorial and have encountered the
following problem in Step 4.
It is claimed (and seems to be) that step-4 is a truly dimension
independent code to solve Poisson equation.
But If I set dim=1 then I get a very strange solution which does not
even satisfy boundary values and is not symmetric around x=0.
In 1D the step-4 equation is: -u'' = 4 x^4, u(-1) = u(1) = 1. Its
solution must be: u(x) = -4/30 x^6 + 68/60.
What is wrong with dim=1 in step-4 code?
Thanks!
Ivan
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
