Hello Soon Hoe,
what finite elements do you use? For continuous finite elements this is
natural behaviour, because they are continuous -- as the name already
says -- and, thus, you get an approximate continuous function if you
interpolate a discontinuous function with it. Basically there are two
ways to deal with this. The first one is to use discontinuous elements
which resolves the problem in a natural way. The second one is to use a
finer grid or a higher polynomial degree of the continuous finite
element. Then the approximation gets better, though not discontinuous at
all.
Best Regards,
Markus
On 23.06.2012 13:52, Soon Hoe Lim wrote:
Hi all. For solving time dependent problems, at t=0 we need to project the
initial conditions into the finite element space. But when we have a
discontinuos initial condition, say u(x,y,0)=1200 on [0.4,0.6]x[0.4,0.6],
u(x,y,0)=0 elsewhere, where the domain is [0,1]x[0,1], I observed that the
VectorTools::project function tries to smooth the discontinuity at t=0.
This causes the solution to violate the initial condition because say at
(0.4, 0.4) the solution at t=0 is no longer 1200 as desired but some
smaller value. In this case, how should I deal with the initial condition?
Thank you for any help.
Sincerely,
Soon Hoe
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
