[Nek5000-users] problem with turbCahnnel case

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

PS - note that you can circumvent the P-dependency
      by using a random number generator that is based
      on a P-independent generator.  There is an example
      in the rayleigh/ray0.usr file in useric().




On Tue, 22 Jun 2010, nek5000-users at lists.mcs.anl.gov wrote:

>
> Dear Francesco,
>
> Are you using 4 processors (or perhaps 8?)
>
> My guess is that the periodicity you're seeing is the imprint
> of the rand1 function on the initial conditions, which is not
> really random in this case since it gives the same result for
> each processor.  Hence, you get this periodicity in the ICs.
> After you pass through transition, however, which typically
> occurs around a time of t ~ 150-200, the ICs are forgotten
> and you should end up w/ a nonperiodic problem.
>
> Paul
>
>
>
> On Tue, 22 Jun 2010, nek5000-users at lists.mcs.anl.gov wrote:
>
>> 
>> Hi Neks,
>> 
>> I am a newbie and i am trying to learn Nek5000 from the examples.
>> 
>> I run the TurbChannel case but i am not convinced by the results.
>> If i plot the x velocity the results look fine but if i plot the y-velocity 
>> in an xz plane close to the wall (y=-0.9) the results show clear 
>> periodical patterns both in x and z that are believe are not Physical. 
>> Moreover these strong periodical structures coincide with the element's 
>> boundaries which is suspicious.
>> 
>> I have tried to implement previous suggestions on the turbchannel case 
>> found on the mailing list( dt=0.025 and simple filtering)
>> but i didn't notice improvements.
>> 
>> Do you have any comments/suggestions on this? Do you think i am making a 
>> mistake in the plotting or in the settings ( i did a copy and paste of the 
>> case in the repository).
>> 
>> Thank you for any help
>> francesco
>> 
> _______________________________________________
> Nek5000-users mailing list
> Nek5000-users at lists.mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
>