[joha]

Hi,

I'm trying to model the imbibition of water in a porous domain. Therefore I set up a simple "1D" model.

My geometry has the following dimensions: 5x5x25 mm, the mesh consists of 1000 stacked cells with symmertries to left+right, front+back, an inlet at the bottom and an opening on the top.

The pressure drop in the porous domain is modelled by setting the appropriate permeabilty. In addition to that I configured a subdomain to model the capillary and the gravity force. Both are take into account as a bulk source by the following term in z-direction:

g*density -C [N/m^2]*WATER.vfs.grad z

C is a material constant (=f(porosity, surfacetension,.....))

The simulation works fine for a few thousand iterations. After that I get a verry blurred asymmetric interface (before the interface gets blurry the pressure field becomes asymmetric):

Detailed view with visible mesh.



I already tried various settings without success:

• smaller timesteps
• higher mesh resolution
• switching inferface compression (0, 1, 2)
• a different mesh (2x2x1000 cells, same geometry) with a free slip wall instead of a symmetrie boundary
• serial run (CFX documentation mentions possbile instabilities if the interface crosses the partitition boundary)

The simulations ends with no errors if I remove the capillary source term and set the same pressure at the inlet.

How is it even possible to get this kind of asymmetry in a "1D" mesh?

I would appreciate any kind of help. Thanks!