[Acto de Lima Cunha]

Hi

I'm having a problem to set my problem involving a fluid domain and a porous domain (ceramic membrane) in the CFX pre.

The problem is that so far unable to set a model of the turbulent for fluid domain and laminar (Darcy) for the porous domain.

I created the two domains (porous and fluid) in the CFX Pre and a interface porous fluid in Domain interface, I configured option in the model Additional Interface Model of mass transfer and momentum-type Conservative Interface Flux, the method used was the GGI Connection.

With the domains and the interface created, I'm able to simulate with the same turbulence model in the two domains, but want to know how to separate the models.

Getting separates the two models using the Enable Beta Features in Case Options > General Options and disabled Constant Domain Physics. Thus, achieving enable a model the turbulent to fluid domain and laminar to porous domain, the problem is that in CFX-solver error appears before the first interaction "Stopped in routine memerr".

I think that is missing enable a feature, or am I skipping a step, and do not know which, can anyone help me?

I am using the package ANSYS CFX 12 and generating the mesh in ICEM-CFD.

I thank anyone who can help me.

[ghorrocks]

You cannot have a laminar region ajoining a turbulent region. Can you use a fully turbulent model and define your porous region using the resistance coefficients as it is normally done? When the resistance dominates the flow I suspect one of the resistance options will boil down to Darcy flow but I have not done the maths to find out which one.

[jeffrey]

hi,Acto de Lima Cunha,
I just meet a similar problem as yours, have you solved your problem? I want to combine k-e model with LES model in two connected domains and the same error comes to me. Can anyone help me? Thanks a lot in advance.

[ghorrocks]

Have a look at the SAS and DES models in CFX. Joining a k-e model to a LES model is not easy as you have to generate the large scale turbulent vorticies in the LES side of the interface from the k value on the k-e side. This is not easy to do.

[TrII4d]

hi Acto de Lima Cunha,

it is exactly like ghorrocks said ... if u define the parameters in ur porous domain u will get darcy's law automatically by CFX ... there is no need to define it by yourself... run the simulation with the turbulence model only and check the streamlines ...

[jeffrey]

hi,ghorrocks,
thanks for your reply.I have tried SAS and DES, it still doesn't work.How should I do to join k-e with LES.Changing k value or something else?details of error is below.
jeffrey
Details of error:-
----------------
Error detected by routine PEEKI
CDANAM = ../../ZN2 /ISUB
CRESLT = NONE

Current Directory :
/FLOW/ALGORITHM/GLOBAL/SUBSYS/WALLSCALE_SG1/WALLSCALE_CM1/ZIF1

ERROR #001100279 has occurred in subroutine ErrAction.
Message:
Stopped in routine MEMERR

[ghorrocks]

People are writing journal articles on how you can couple k-e turbulence models to LES style models. I am certainly not going to write a thesis on it on a forum. If you are considering this then you will need to do a lot of research and background reading as it is a very difficult task.

[jeffrey]

thanks a lot, ghorrocks.