[kpax]

hello everyone,

as i posted in another thread, i am currently trying to simulate the flow in a branching vessel tree. In addition, i want to simulate the transport of a tracer in this domain by solving the convection-diffusion equation. for this purpose, i modified the icoFoam solver according to http://openfoamwiki.net/index.php/Ho...ure_to_icoFoam

i encountered some problems (i.e., unphysical results) when using a quad mesh that were probably due to the poor mesh quality. So i thought why not try a tri mesh - it is prone to numerical diffusion, i know, but it might give me a first idea of what happens in the domain.

So here's what's weird: first, the velocity field does not develop a parabolic profile as I would expect in a tubular region (see picture in attachment). but even stranger, the tracer is not transported along the domain at all! I can see an increase of concentration at the inlet (according to the BC there) - but that is it, the rest of the domain has c=0, as you can see in the attached picture.

I already tried different discretization schemes, but that did not yield any different results.


anyone has an idea whats going on here? maybe i missed something?



for the sake of completeness: i used deltaT = 0.0001 (different values did not affect result), the mesh has around a million cells (again, coarser meshes gave same results). the solver worked for different domains and hexa meshes, so i don't think it's the problem.
if more information is needed, i can post it.


thx in advance!

[kmooney]

Could post a quick picture of your mesh?

[kpax]

sure. mesh at the inlet looks something like this..