[gera]

Does anybody know how to assess quality of the mesh at interface? In particular, mesh expansion factor and volume change. ICEM CFD shows these parameters only in the domain on one side of the interface. I noticed that convergence can highly depend on interfaces in the computational domain. Although, quality of the mesh can be ideal on both sides of the interface still convergence can be far from ideal. At the same time, convergence can be improved if the mesh is built without interfaces even when the quality of the mesh is worse.

What does CFX exactly do at fluid-fluid interface? How does CFX construct a control volume on interface? Is there any literature to read about algorithms for computation through interface?

[ghorrocks]

The GGI interfaces are complex and described in the documentation.

Yes, your simulation is more numerically stable and accurate if the size of the elements is constant across the interface. But how significant this is will depend on what you are modelling and the physics you are using. Some will be highly sensitive and some not very sensitive at all. So your comments may well be correct for your application, but be careful when you think about generalising it to all applications.

[gera]

Originally Posted by ghorrocks

The GGI interfaces are complex and described in the documentation.

Could you please give the direct reference to the section of the documentation that you mentioned? I found something, however, it is still not so clear to me.

Thank you!

[ghorrocks]

I am at home and don't have access to the software so cannot. But it is not hard - just do a search for General Grid Interface.