I'm getting my feet wet with OpenFoam and am trying to run a free jet code in 2D cylindrical coordinates. The messages I get about convergence seem fine during the run (pisoFOAM), but I'm not sure that continuity is satisfied terribly well. This seems unlikely, so I hope someone can point out where I'm going wrong.
I was suspicious, because I'm taking results I get from "sample" and applying them to a PDF code I wrote, but my particles are tending to aggregate in certain regions.
As a sanity check, I calculated the velocity gradients in paraFoam and then calculated the divergence (since I've got a cylindrical mesh, I had to take ux/x + dux/dx + duz/dz). This was disconcerting, because the calculated values were not terribly small in many regions, especially where my PDF code was congregating particles.
However, just a visual examination of a contour plot of each gradient field makes it look like something is working; they qualitatively look like negatives of each other, as they should.
My front and back boundaries are symmetry (wedges were giving me issues), so I shouldn't be losing any flow tangentially. I also did some set sampling of faces along lines. If I sample horizontally and vertically and then check the in and outflow of the faces at the intersection, the inflow and outflow were of order 1e-7 or 1e-8, but the net flow in/out was order 1e-9. This seems like it should be much smaller; relative to the inflow and outflow anyway.
Does anyone have any idea where OpenFoam or myself are going wrong? Do I just need to tighten up the U convergence criteria (trying that now)?
Thanks in advance!
|All times are GMT -4. The time now is 05:00.|