A question on restarting file

[MartianDuan]

Hello users,

I need to use k-omega-sst to do some simulation. However, the simulating case keeps diverging. To make the simulation converge, I tried to simulate it is as liaminar flow. This gives me a very rough predictions, I turned on the turbelence model afterward.

However, as there is no fiel for K and omega in the restart file created by the liaminar flow simualting case. The turbuelnce cased can not be run.

I am wondering whether we can add extra field in the restarting file. Or is it possibel to turn off some functions of the turbulence model at the very beginning to give the case a rough prediction and turn them on afterwards?

The listing log is attached.

Thanks very much.

Kindest regards,

Yu.

[Yvan Fournier]

Hello,

Most quantities are computed, but turbulent values are not (it is easier to restart from k-epsilon to k-omega, as turbulence variables are "converted"). You may try to initialize k and omega values in a better manner using the usiniv subroutine (in cs_user_initialization.f90 file), as it is called after the restart file is read.

Regards,

Yvan

[MartianDuan]

Hi Yvan,

Thanks for your suggestions. But I did not find the 'usiniv' on the cs_user_initialization.f90. I am not sure whether it is because the version of code_saturne is different from yours. I am using code_saturne 4.0.2.

Can you please give me hints?

Best regards,

Yu.

Hello,

Most quantities are computed, but turbulent values are not (it is easier to restart from k-epsilon to k-omega, as turbulence variables are "converted"). You may try to initialize k and omega values in a better manner using the usiniv subroutine (in cs_user_initialization.f90 file), as it is called after the restart file is read.

Regards,

Yvan

[MartianDuan]

Hi Yvan,

I am sorry that I did not include the subroutine in the previous post.

It is attache now.

Best regards,

Yu.

Hello,

Most quantities are computed, but turbulent values are not (it is easier to restart from k-epsilon to k-omega, as turbulence variables are "converted"). You may try to initialize k and omega values in a better manner using the usiniv subroutine (in cs_user_initialization.f90 file), as it is called after the restart file is read.

Regards,

Yvan

[Yvan Fournier]

Hello,

Sorry, usiniv was the old name. cs_user_initialization is the correct routine.

So defining turbulence values based on a reference velocity and the litterature is probably a good start.

Regards,

Yvan

[MartianDuan]

Hello Yvan,

Actually I tried the method you suggested. however, I still have the same problem.

I am wondering whether it is due to the subfuntion such as 'call field_get_val_s(ivarfl(ik), cvar_k)'.

I guess it is to get the scaler or velocity field. I am wondering whether it is necessary to do so in the initialization?

Thanks for all your help.

Regards,

Yu.

Hello,

Sorry, usiniv was the old name. cs_user_initialization is the correct routine.

So defining turbulence values based on a reference velocity and the litterature is probably a good start.

Regards,

Yvan

[Yvan Fournier]

Hello,

Tee field_get_val_s_function allows you to access the field values, but it does nothing if you do not modifiy or otherwise use those values.

Regards,

Yvan