Some time ago I posted this topic. For very low flows, 0<Pe<100, I have favorable results IN 2D (axisymmetric) with species transport compared to accepted (zero velocity) theory. This is for diffusion from a spherical droplet into water.

I tried modeling similarly in 3D, but with poor results compared to theory. Previously, it was suggested that I refine my 3D grid in order to approach the effective (fine) grid of the 2D-axisymmetric model. What I've found is that the finer my 3D grid, the worse the zero velocity species tranpsort model compares to theory. In 3D, I'm running unsteady, but it's very inaccurate. I tried steady but with abysmal results.

Are there other ways to improve a 3D model for species transport so that it better mimics its 2D-axi brother?

My post, above, was somehow clipped, so here are clarifications. This is species transport from a spherical droplet into water.

I'm looking at Peclet numbers ranging from 0 to 100. Trying to match zero-velocity theory with Peclet close to zero (veloc ~ 1e-9 m/s).

The 2D-axisymmetric models run reasonably close to theory in the unsteady analysis. I've tried running 3D in both steady and unsteady, but oddly the steady case is very poor.

So my questions regard adapting mesh in 3D to better match 2D-axi results. And why wouldn't the near-zero flow case be a candidate for steady analysis?

Thanks.