Stokes Flow for a moving bubble
 

Hello Foamers,

I am trying to simulate stokes flow (low reynolds number) for a bubble (or droplet) and compare the results to the analytical solution.

I am using the interFoam solver with the following fluid properties (all units are SI units):
rho = 1000, nu = 1e-2
rhoB = 900, nuB = 1.11e-2

The radius of the bubble is 5mm. The Reynolds is 0.01.

As I increase the mesh resolution the average bubble speed converges onto to the analytical, but if I continue to increase the resolution the average speed falls below the analytical (see attached image, R/4 indicates that a dimension of a square mesh element is 1/4 that of the radius).

I am using a fixed time step and the maximum Courant number for all simulations is less than 0.09.

Could anyone help me as to why the increasing the mesh resolution would cause the solution to go past that of the analytical value? Is it possible to have to fine of a mesh? Any insight on to what I may be doing wrong would be much appreciated.

Sorry that the question is so long, but I wanted to add as much relevant information as I could.

Thanks,
Paul