code_saturne User's Forum

Hello all,

I am using the GUI and V 4.0. I am modeling a pipe flow where I have different zones (volumes) in the model.

I want to model the flow in one of the zones to be laminar. While modeling all other zones turbulent. Is there anyway to do this for that specific zone ? when doing zone initialization the options are initialization, momentum, porosity, thermal source. I couldn't find anything to specify flow regime in one zone.

Thanks,

Hello,

no, the turbulent model is global.

One solution would be to couple 2 instances of the code, one for each zone (but this is complex and requires extra precautions).

I'm not a turbulence specialist, but using LES for example, you should naturally have the correct type of behavior in different zones, but computation requires a fine, high quality mesh, and time integration, so it can be very expensive.

I'll check with a colleague, but I believe the "elliptic blending" turbulence models (Rij-EBRSM for example) could be an option here.

Regards,

Yvan

Yvan Fournier wrote:Hello,

no, the turbulent model is global.

One solution would be to couple 2 instances of the code, one for each zone (but this is complex and requires extra precautions).

I'm not a turbulence specialist, but using LES for example, you should naturally have the correct type of behavior in different zones, but computation requires a fine, high quality mesh, and time integration, so it can be very expensive.

I'll check with a colleague, but I believe the "elliptic blending" turbulence models (Rij-EBRSM for example) could be an option here.

Regards,

Yvan

Thanks Yvan, I want to invoke a laminar zone within a zone that I am defining as a porous media. I know that the flow is laminar there because of the geometry and the choice of modeling this zone as porous is to actually avoid a fine high quality mesh because in my case it it impossible to mesh.

I would like to know more about this elleptic binding model option, and instance code coupling once you have more info.

Your help is much appreciated. Thanks very much.

Hello,

If you have a porous zone, the elliptic blending model might not be the best solution after all.

What you probably want is to use turbulence source/sink terms to reduce / adapt the turbulence (at least the k or Rij terms) in the porous zone.

We usually leave turbulence "as is" in porous zones, but are careful when interpreting results. Also, in porous zones, assuming you define also head losses, I prefer using symmetry rather than wall conditions, to avoid counting friction twice (as is is already counted in the head losses).

Regards,

Yvan