CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Non-reflecting boundary conditions (http://www.cfd-online.com/Forums/fluent/80060-non-reflecting-boundary-conditions.html)

droberts September 14, 2010 03:40

Non-reflecting boundary conditions
 
Hey,

I have just started to use a non-reflecting boundary on a basic 2D rectangular mesh. The boundary forms a pressure outlet with a fluctuating static pressure being applied at an opposing pressure inlet. The problem I am having is that the solver diverges almost instantly. I have reduced the courant number and the relaxation constants however non of these seem to stop the divergence. Please can someone give me some ideas on how I can solve this issue as I would like to use the boundary condition on a larger more complex grid.

Thanks

jwillie2000 September 15, 2010 08:18

nonreflecting boundary condition
 
Hi,
What case are you simulating? This is because if you are using the species transport model in fluent then it does not work with non-reflecting boundary condition. In such a case you have to use a udf. If acoustics is what you are looking at, especially with combustion i think using CFX might do a better job for you. Fluent has issues, especially with compressibility. I am not sure how compressible your flow is?

Good luck!

Jimmy

droberts September 15, 2010 08:25

Hey,

Eventually I will be simulating acoustic/pressure waves impinging upon a resonator coupled with a grazing flow velocity. However, the simple mesh at the moment is a 2D tube just to set up the boundary conditions. I dont know if this helps with any suggestions.

Dave.

jwillie2000 September 15, 2010 09:39

non-reflecting boundary condition
 
Hi Dave,

It means you are simulating just air and so it should work for you? The first thing that comes to mind wold be to play around with your mesh. What is the maximum skewness for your mesh for example? Try to make the mesh very fine and check the skewness to make sure it is not high.

Sometimes the trick in fluent is to do the first few iterations as incompressible before switching to compressible. This could also be tried.

Let me know how it goes.

Jimmy

droberts September 16, 2010 05:01

Hey,

I tried altering the mesh so it now has a maximum aspect ratio of 6 and getting convergence whilst incompressible. However, once I changed back to the density based solver and ideal gas model for the non-reflecting boundary, the solver diverged again as it did previously. Do you have any other ideas?

Thanks for your help on this matter.

Regards
Dave

jwillie2000 September 17, 2010 05:47

non-reflecting boundary condition
 
Hi Dave,

I guess you are using the explicit solver? Did you try the implicit one as well.

Plus I would suggest you consider doing a steady simulation first before switching to unsteady, to see whether that works. Plus the time step size is critical and so reducing it might help. You want to make sure that the time step size u are using can resolve the acoustic field.

Hope this helps!

Jimmy

droberts September 17, 2010 05:55

Hey,

Thanks for your quick replies. I did try the implicit solver, but with similar results. The solver diverges after about 5 iterations so the time step used cant be an issue as it doesnt even get that far. I've never used CFX before, but I might give that a try as I really cant see what the issue is with fluent.

Cheers

Dave.

jwillie2000 September 17, 2010 06:07

non-reflecting boundary condition
 
Hi Dave, if in deed you want to do acoustics, i think CFX is better than Fluent. So go for the CFX. Fluent has issues with how it handles compressibility, which is a key issue in acoustics.

Good luck!

Jimmy


All times are GMT -4. The time now is 01:26.