Hello,

I want to use a new turbulence model, that is not currently available in CFX. I am thinking which approach is better (or possible) to include the new model: using CEL ou Junction Box Fortran routines?

Any advice will be apreciated!

flga

Hi,

CEL functions are easier to implement, providing the functionality you want can be defined in terms of a CEL function.

Glenn Horrocks

Dear flga,

Your question is a bit vague.. What model do you intend to implement? CFX has a few hidden turbulence models, perhaps you should ask tech support before embarking into some that has already been done..

Is this model a two-equation model, or a full fledge second moment model? Does it require special numerics, and consistency with CFX way of dealing with turbulent wall functions? Can it be done using Additional variables? or is just a modification of a well known model that can be approched via a source term, or coefficient model change?

The better you explain the details of the model, the easier for members of the forum to help you..

Good luck

Opaque.

Dear Opaque,

I will explain my project in more detail. I work with a experimental team which provides me data from a flow configuration, and my job is to make the simulation of this flow.

The flow is a challenging one, and I have benchmarked all the RANS turbulence models from CFX (even the hidden ones, connected via ccl, like EARMS). What I want to do now is to implement some other turbulence models that I found in the literature. I want to begin with two-equation turbulence models, but my intention is to try some second moment closure models in the future (if the two-equation models fail).

I am not sure I need special numerics to solve the new models, and only changing the coefficients of the CFX turbulence models is not enough in my case.

I hope it helped to you better understand my problem.

flga