[chenu]

Hello,
I have a problem with my simulation. I would simulate flow over a sphere by icoFoam. It is a quite simple simulation. The model is built in Salome. I created unstructured mesh and there is no prismatic mesh on the sphere's surface. The sphere is enclosed in a cubic computational domain, whose sizes are 20 times larger than sphere diameter.
I defined in, out, wall and sphere. The boundary condition are:
Pressure:
in: zero-gradient
out: fixedValue, uniform value 0
wall: zeroGradient
sphere: zeroGradient
Velocity:
in:fixedValue, uniform(0.5,0,0);
out: zeroGradient
wall: noSlip
sphere: noSlip

The time step is 5e-4, it keeps maximum Courant number less than 0.2. This simulation goes well, but after 7.5 seconds(after 15000 timestep), divergence is occurred due to a large pressure near the out boundary. I have attached velocity and pressure distribution to this post.

My question is: Is this divergence due to bad mesh quality?
If I set wall as velocity zeroGradient, I think there is no problem. However, I have more interest in which boundary condition is the best to apply?
I will appreciate any recommendation. I will be also glad if you share your experience. Thank you