code_saturne User's Forum

Hi everyone,
I am currently simulating flow through a circular pipe where an annular porous zone is implemented along the interior wall.
I have defined the physical properties of the porous region using the following formulas:

My results show a significant and localised pressure difference between a specific section of the annular porous region and the main flow area.It appears that the flow resistance in one particular part of the porous zone is excessively high, resulting in very low velocities in that area. This localized pressure gradient is forcing the fluid to divert toward the opposite side of the pipe, causing an asymmetrical flow pattern.
I have applied uniform physical parameters to the entire annular region but why only a small portion of the porous zone exhibits this high resistance while the rest behaves as expected. What could be causing this localized high resistance despite uniform property settings? (Could it be a mesh resolution issue?) Should I consider adding specific source terms to stabilize the pressure at this location, or is there a more fundamental way to address this? Any insights or suggestions would be greatly appreciated. Thanks!

Hello,

Coule you provide more information (pas per the forum usage recommendations) ?
Without any info on the code version, model used, type of mesh, setup, ... it is difficult to provide a relevant answer.

Best regards,

Yvan

Hello,

Thank you for your reply. I will provide more relevant information. The version of Saturne I used is 8.0.1. The mesh is hexahedral, and the turbulence model is RANS k-ε. The computational domain consists of a circular pipe containing three layers of annular porous media. Below are screenshots of my computational domain and the results.

MESH:

porosity and density, As shown here, the distribution of physical properties is uniform across each porous region.


PRESSURE: The pressure results are non-uniform: a high-pressure region is present at the bottom of the pipe.


Other calculation settings can be found in the attached setup file. Thank you!

Hello,

Does the high pressure region appear along the whole axis of the pipe or closer to the inlet or outlet ? This might be due to a combination of inlet or outlet boundary conditions and gravity.

Can you try using a regular outlet instead of an imposed pressure outlet ? When using an imposed pressure outlet, there are 2 ways of imposing this, one relative to the "working pressure" (which subtracts the hydrostatic component) and one including the hydrostatic component. Un user-defined functions, one can switch from one to another by changing the sign of the "icodcl" component, but I am not sure this is documented). Although imposed pressure outlet has some advantages (more robust) in many cases, it can be tricky when used on an outlet which is not orthogonal to gravity...

Also, can you test without the "pseudo-coupled velocity solver" option ? This is rarely used, interesting for larger time steps, but less tested/validated.

Best regards,

Yvan

Thank you for your suggestion.

The figure below shows the 3D pressure results for my pipe. The annular porous region I defined is located at a certain distance from both the inlet and outlet (the area covered by the blue brackets in the figure).

The high-pressure zone is not near the inlet or outlet, but rather confined to a single strip-shaped region(the location indicated by the black arrow) along the axis of the annular porous zone.



Following your advice, I modified my calculation settings (changing the pressure outlet to a regular outlet and disabling the “pseudo-coupled velocity solver” option). However, these changes did not significantly alter my results.

Maybe this high-pressure zone may be related to the material properties of the porous region. When I increased the porosity of the porous region near the wall, the pressure band persisted, but the overall pressure difference decreased.

Currently, I am trying to adjust the position of the porous region, as well as experimenting with different turbulence models and wall functions. I don’t think this high-pressure region should be present. Are there any other ways to improve the results?