code_saturne User's Forum

Hello every everybody.

I have mesh in which the refinement is very high (2 cm dx). To avoid the problem of Y+ I learn in user manual that it's possible to set a scalable function for the velocity boundary conditions (ideuch=). I use the --nogui option and I wonder how to implement it, which routine to modify ?

thank you for your help

Hello,

I think that you have to copy and paste the iniini.f90 file in your SRC folder.
you can find ideuch in iniini.f90. the default value is -999 (chosen by the turbulence model).

I never try to modify this value, but I think that with ideuch = 2, you can use scalable wall function

if you want to confirm the fact that the modification is well taken into account, you can print in the listing file the value thanks to :

write(nfecra,*) 'ideuch = ', ideuch


let me know if it works.

good luck

Sylvain

Hello again Sylvain.

Thank for your help, I did what you say and it work well now. The calculation goes to the last iteration without the warning message and with coherent results!

I am effectuating some experiences to validate the modules of combustion and radiative. I try to work with both now, but I'm facing problems (there is no subject in this forum which deals with combustion + radiative activated modules experience, I will create one).

Thanks again

Just so you know you can change the wall-function in usini1. Type the command in usini1 in the usipsu part of the subroutine.

Thanks,

James

Dear James, I have validated your method, it works as the same as sylvain method. Thanks you for your help

Hello,

In any case, the recommendation is to use usini1.f90, which contains user subroutines, and not iniini.f90, wich defines code defaults, and is not a user file (though reading it is the best place to find examples for the variables you can set, as it is oftrn more practical than searching in the user documentation).

Granted, choosing the correct place to set this in usini1 (usipgl, usipsu, ...) is not always easy, though in most cases, usipsu should be used.

Regards,

Yvan