[Nek5000-users] Initial Condition & Recycling

[nek5000-users at lists.mcs.anl.gov]

To avoid the division by zero (for P_N/P_N), you may need to define

nu_star = param(2)            ! constant implicit viscosity part

in usrdat. Otherwise, nu_star may be undefined in the routine split_vis, 
and the vdiff array will be zero.

On 06/11/2018 03:46 PM, nek5000-users at lists.mcs.anl.gov wrote:
> Dear Paul,
>
>
> Indeed thanks for the input However, the volume average velocity still
> tends to zero and as a consequence the rescaling routine tries to
> compensate making the code unstable until it diverges. The reason why
> the volume average velocity tends to zero could simply be because the
> flow relaminarizes in the absence of fluctuations in the prescribed
> inlet. From my understanding the recycling routine only rescales the
> flow to a target mean velocity without adding fluctuations.
>
> To be sure of that change of dynamic I set up a simulation with a
> (1/7)th velocity profile at the inlet (same code but without recycling)
> convinced that the flow would laminarize and tend to a "blasius-like"
> profile. To my surprise the flow remained steady.
>
> Finally as I want to use a dynamic Smagorinsky model for this project I
> set the filtering to none and variable properties to TRUE however it
> seems that there is a division by zero which I couldn't trace back in
> the source code.
>
>
> Just after starting to compute the eddy viscosity :
>
>
>  Calculating eddy visosity
>  flt amp 0.0000 0.0000 0.0000 0.0000 0.0000 0.0500 0.5000 0.9500
>  flt trn 1.0000 1.0000 1.0000 1.0000 1.0000 0.9500 0.5000 0.0500
> forrtl: severe (71): integer divide by zero
> Image              PC                Routine            Line
> Source
> nek5000            00000000008CF105  Unknown               Unknown  Unknown
> libpthread-2.17.s  00002B44DF7BD370  Unknown               Unknown  Unknown
> nek5000            00000000005D70BE  Unknown               Unknown  Unknown
> nek5000            0000000000408121  Unknown               Unknown  Unknown
> nek5000            0000000000404D3F  Unknown               Unknown  Unknown
> nek5000            000000000040F572  Unknown               Unknown  Unknown
> nek5000            000000000040E0DF  Unknown               Unknown  Unknown
> nek5000            0000000000404A1E  Unknown               Unknown  Unknown
> libc-2.17.so       00002B44DFCEDB35  __libc_start_main     Unknown  Unknown
> nek5000            0000000000404929  Unknown               Unknown  Unknown
>
>
> Sincerely,
> Armand.
>
> Le 07/06/2018 à 15:43, nek5000-users at lists.mcs.anl.gov a écrit :
>>
>>
>> Dear Armand,
>>
>>
>> The attached .usr file seems to work ok... just a few
>>
>> small modifications to the .usr file you provided.
>>
>>
>> I think the main issue was the declaration of some of the
>>
>> variables in useric.
>>
>>
>> Best,
>>
>>
>> Paul
>>
>>
>> ------------------------------------------------------------------------
>> *From:* Nek5000-users <nek5000-users-bounces at lists.mcs.anl.gov> on
>> behalf of nek5000-users at lists.mcs.anl.gov
>> <nek5000-users at lists.mcs.anl.gov>
>> *Sent:* Thursday, June 7, 2018 3:51:09 AM
>> *To:* nek5000-users at lists.mcs.anl.gov
>> *Subject:* [Nek5000-users] Initial Condition & Recycling
>>
>> Dear All,
>>
>>
>> In an attempt to simulate an established turbulent flow over a flat
>> plate with a recycling method for the inlet, it has to my attention that
>> a (1/7)th Power Law would be suitable as an initial condition. However
>> the solver divergences and I think I know why.
>>
>> Here the layout of the recycling procedure :
>>
>> 1. call set_inflow_fpt in userchk
>>
>> 2. If it is the 1st call then call setp_inflow_fpt_setup
>>
>> 3. call rescale_fpt (which rescales the copied field to a user specified
>> mean velocity ubar)
>>
>> 4.In the rescale_fpt routine we call get_flux_and_area which gives the
>> flux and area of a face with cbc == 'v   ' so a user specified velocity
>> boundary condition.
>>
>> 4. a) As the initial condition is wrong once that routine sums the flux
>> on the boundary it returns zero in this case and the variable scale
>> which is divided by the latter returns infinity !
>>
>>
>> However I set the initial conditions as in many other examples, such as
>> the turbInflow. Would have an insight that could lead me to a solution ?
>>
>>
>> Please find attached the related files.
>>
>>
>> Sincerely,
>>
>> Armand, ONERA - The French Aerospace Lab.
>>
>>
>>
>>
>> _______________________________________________
>> Nek5000-users mailing list
>> Nek5000-users at lists.mcs.anl.gov
>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
>
>
>
> _______________________________________________
> Nek5000-users mailing list
> Nek5000-users at lists.mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
>