CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   CFX and FORTRAN (https://www.cfd-online.com/Forums/cfx/24176-cfx-fortran.html)

Bruno June 25, 2007 08:51

CFX and FORTRAN
 
hi

i'm simulating a dispersion of pollutants on atmosphere... and i need to use a special boundary condition on inlet of my domain...

how can i introduce an equation using fortran language in cfx????

thanks...

bruno

opaque June 25, 2007 09:55

Re: CFX and FORTRAN
 
Dear Bruno,

Would you mind giving a few more details, so those in the forum can give you a hand?

There is a lot that can be done via CEL expression before tacking the User Fortran route.. When you said "an equation", are you referring to solve a new equation, or writing an equation for your boundary condition?

Opaque


Bruno June 25, 2007 10:11

Re: CFX and FORTRAN
 
hey

what I want to do cannot be done by cel expression... I'm wanting to apply an equation for my boundary condition, where I'll have to read values from a file...

i didn't understand the steps to use a subroutine in CFX... it's confuse...

thanks by your attention

bruno

opaque June 25, 2007 14:50

Re: CFX and FORTRAN
 
Dear Bruno,

Still that does not explain why you need User Fortran..

Assuming you want to interpolate values from the file, and evaluate an alternate expression that produces values for the boundary condition, you still do not need any User Fortran in CFX.

Create an interpolation function in CFX, evaluate as many equations as you want via CEL, and input the final results into the boundary condition panel.

However, if you boundary condition is not one of the 3 typical: Value (Dirichlet), Flux (Neumann), or Flux Transfer Coefficient (Robins), then I understand why you might need use of the User Fortran route..

How complex is your boundary condition expression? Even with the profile file.

Regards,

Opaque.


Bruno June 25, 2007 21:03

Re: CFX and FORTRAN
 
One more time, thanks by your attention!!

I have an expression like this:

U(z) = Uaverage(z)+lambda*ustandarddeviation(z)+lambda*us tandarddeviation(z)*A

where A is a random function i.e. at each time step it is a number between 0 and 1.

There are two problems to use CEL expressions. First I have to read uaverage and usdv from a file and after this there isn't random function on CEL expressions

sorry if I'm not clear... it's difficult to explain on e-mail sorry by my english too!!

thanks!!

bruno


Joey2007 June 27, 2007 16:48

Re: CFX and FORTRAN
 
Sounds like LES?

You can do this by programming a User CEL function which calls for example the fortran random function. This User CEL could be combined with the interpolated functions. This would keep your development effort on very low level.

However the other question is: does such boundary condition fit into the modeling strategy followed by the CFX solver (and by most commercial solvers)?

HekLer June 27, 2007 16:59

Re: CFX and FORTRAN
 
I think you only need to create a simple user CEL function which calls a fortran version of the random function.

The other two functions of z can probably be done in CEL.

Jenny June 28, 2007 05:58

Re: CFX and FORTRAN
 
Bruno, you can compile one Fortran routine and define the boundary condtions in CFX-pre via user fortran.


All times are GMT -4. The time now is 07:37.