[Nek5000-users] PN/PN and usrdiv

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

In addition to the issue that Stefan identified, it turns
out that the set_outflow_dist() routine in the .usr file
was dependent on zm2 (z coords on mesh2), which defined
only for Pn-Pn-2.

I've updated coef.f so that xm2 ym2 and zm2 are now defined
for both formulations and it seems that the outflow conditions
are now working.

Paul


On Fri, 9 Apr 2010, nek5000-users at lists.mcs.anl.gov wrote:

>
> I'll check into it this afternoon...
>
> Paul
>
>
> On Fri, 9 Apr 2010, nek5000-users at lists.mcs.anl.gov wrote:
>
>> Paul,
>> 
>> I think know why. We should not multiply usrdiv with the mass matrix
>> 
>> so just replace
>> 
>> call col3 (h1,usrdiv,bm1,ntot1)
>> call add2 (qtl,h1,ntot1)
>> 
>> by
>> 
>> call add2(qtl,usrdiv,ntot1)
>> 
>> at the beginning of plan4().
>> 
>> Unfortunately I don't have the time to test this fix.
>> 
>> Let me know if this works for you.
>> Stefan
>> 
>> 
>> 
>> On Apr 9, 2010, at 4:21 PM, nek5000-users at lists.mcs.anl.gov wrote:
>> 
>>> 
>>> Yes - I just checked and it's indeed not doing anything for
>>> Pn-Pn.
>>> 
>>> I'll get to this by the end of the afternoon... should take
>>> minimal effort to repair I think.
>>> 
>>> Sorry about this Markus.
>>> 
>>> Paul
>>> 
>>> 
>>> On Fri, 9 Apr 2010, nek5000-users at lists.mcs.anl.gov wrote:
>>> 
>>>> I have not used this feature for a while but usrdiv is used in plan4().
>>>> 
>>>> --- snip plan4.f ---
>>>> 40       ! add user defined divergence to qtl
>>>> 41       call col3 (h1,usrdiv,bm1,ntot1)
>>>> 42       call add2 (qtl,h1,ntot1)
>>>> --- snip plan4.f
>>>> 
>>>> 
>>>> Stefan
>>>> 
>>>> 
>>>> On Apr 9, 2010, at 1:49 PM, nek5000-users at lists.mcs.anl.gov wrote:
>>>> 
>>>>> Hi,
>>>>> 
>>>>> since I switched the pressure scheme from P_N/P_N-2 to PN/PN, as 
>>>>> recommended by an earlier email, I have divergence trouble (simulation 
>>>>> crashes) at the outflow, were a stabilization technique as in examples 
>>>>> "peris" and "jet" is in place (adding divergence via usrdiv).
>>>>> I spent a long time trying to fix this (several test runs with increased 
>>>>> outflow normal element length, increased divergence magnitude, mesh 
>>>>> deformation to nozzle shape, which avoids crashing but has an adverse 
>>>>> effect on the upstream flow,...) but I am becoming suspicious that 
>>>>> usrdiv is just not used in PN/PN.
>>>>> Is that true? If so, how to fix it?
>>>>> 
>>>>> Thanks,
>>>>> Markus
>>>>> 
>>>>> _______________________________________________
>>>>> 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
>>>> 
>>> _______________________________________________
>>> 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
>> 
> _______________________________________________
> Nek5000-users mailing list
> Nek5000-users at lists.mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
>