CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Flow through porous medium using CFX: lack of convergence (

Chander February 14, 2011 11:33

Flow through porous medium using CFX: lack of convergence
1 Attachment(s)

I am simulating flow through porous medium. In my domain, the fluid enters an inlet duct vertically downwards (-z direction) and impinges on a porous medium vertically , then turns and flows horizontally (y-direction) through porous medium and then at the other end, it again turns upwards (+z direction) and exits through outlet duct. The outlet duct is taken to be long so as to avoid an opening type of boundary condition. Please refer to the attached figure for an explanation of the computational domain.
I am using uniform mesh of equal cell size in x-y-z direction. The porous medium has anisotropic permeability so as to inhibit the flow in x-direction. There is also a small fluid gap above the porous medium.
Now I am not able to get convergence. My inlet Reynold number based on hydraulic diameter for inlet rectangular duct is 3023. Inlet flow is 1 l/min and normal to inlet boundary. I have tried the following:
a) Using a k-omega , and SST model for turbulent flow instead of using laminar model
b) reducing the timescale factor in solver settings so as to under-relax the solution
c)Using a finer mesh

However, I see that the residuals just fail to reduce below certain levels (1e-2 to 1e-3) and thus solution does not converge. Many of the maximum residuals are located near porous-medium-fluid interface. Please suggest a remedy for this issue.

ghorrocks February 14, 2011 17:04

A good guide for general tips on convergence is here:

What version of CFX are you using?

There have been some issues with high residuals at the edge of porous regions in past versions, I think more recent versions are improved.

stumpy February 15, 2011 09:16

Try the expert parameter:

FLOW: Flow Analysis 1
porous cs discretisation option = 1

shubham jain July 23, 2014 06:32

2 Attachment(s)
hi, i am simulating a Brush seal using Porous medium approach in CFX.

My inlet (INLET) and outlet (OPENING) channels are very long. CFX says that wall is placed at a portion of Inlet, then it stops the simulation.

;My mentor told me to try it with shorter channel length. I tried with shorter inlet and outlet channels and it worked. But i could not understand, why the shortening of channel length makes it (wall places at INLET) better??

Picture attached: Green region in right is inlet.... Dark yellow in the middle is the porous medium ..... light yellow in the left is outlet

Thanks in advance

All times are GMT -4. The time now is 10:42.