Please someone give me a tip! I have to model a density current being released by a wall. The wall has to be lifted vertically and slowly (let's say 10 second to move across a 1m vertical displacement) allowing the density current to flow beneath it. I have tried with a changing porosity with time but with low values of volume porosity the solver get stuck after a while. I tried with mesh deformation (representing the wall as a gap in the domain and moving the mesh of its boundary), but the mesh get stretched too much and from a certain point on it goes outside the domain resulting in negative volumes. Does anyone an idea on how to overcome these problems or any other way to solve hte simulation?

please please...

Francesco

Hi,

How about representing the wall as a moving mesh with a GGI on either side of the wall so the meshes can slide past each other without distortion.

Glenn Horrocks

thanks Glenn, I was really hoping in a tip from you...by the way, I have never tried anything like that, do you have any examples on how to implement it? I mean...does the general domain has to have a gap where the wall is supposed to be, or can it be full, with the wall simply overlapping over the domain mesh? I am sorry but I didn't manage to find anything that could help me...

Hi,

Have a gap in the domain where the sliding wall is located. Have a moving mesh domain in the gap which you stretch up and down to model the gate opening and closing with a GGI on either side to connect it to the rest of the domain. The GGI will need to be a transient rotor-stator. In older versions of CFX you had to activate an expert parameter to allow this sort of sliding GGI to work but I don't think it is required now.

This stuff is not really covered in the tutorials in CFX but I have done similar things and it works quite well.

Glenn Horrocks