[Nek5000-users] Restarting with PNPN using PNPN-2 results

nek5000-users at lists.mcs.anl.gov nek5000-users at lists.mcs.anl.gov
Tue Jun 14 17:44:38 CDT 2016


Philipp and Aleks,

Thanks for the replies.  The Uzawa method sounds exactly like what I am
looking for.  I would like to pay a penalty upfront to converge to a good
pressure solution without taking too many (or really any) timesteps and
then move forward with a resolved pressure solution.  I am hoping there is
a way to do this since a pressure projection method is being used.

My solution is already well resolved temporally and I don't want to
perturbe it with a long time series of unresolved pressure fields/slightly
divergent velocity fields. Can anyone comment on the use of Uzawa with
PNPN?  Also, Philipp is there a specific routine you call to perform just
one of these iterations?  I haven't used this method before.

Kind regards,

Phil

On Tue, Jun 14, 2016 at 11:00 AM, <nek5000-users-request at lists.mcs.anl.gov>
wrote:

> Send Nek5000-users mailing list submissions to
>         nek5000-users at lists.mcs.anl.gov
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
> or, via email, send a message with subject or body 'help' to
>         nek5000-users-request at lists.mcs.anl.gov
>
> You can reach the person managing the list at
>         nek5000-users-owner at lists.mcs.anl.gov
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Nek5000-users digest..."
>
>
> Today's Topics:
>
>    1. Restarting with PNPN using PNPN-2 results
>       (nek5000-users at lists.mcs.anl.gov)
>    2. Re: Restarting with PNPN using PNPN-2 results
>       (nek5000-users at lists.mcs.anl.gov)
>    3. Re: Restarting with PNPN using PNPN-2 results
>       (nek5000-users at lists.mcs.anl.gov)
>    4. Re: lambda2 in turbChannel (nek5000-users at lists.mcs.anl.gov)
>    5. g2gi() no geometry found (nek5000-users at lists.mcs.anl.gov)
>    6. Re: g2gi() no geometry found (nek5000-users at lists.mcs.anl.gov)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Mon, 13 Jun 2016 13:09:05 -0600
> From: nek5000-users at lists.mcs.anl.gov
> To: nek5000-users at lists.mcs.anl.gov
> Subject: [Nek5000-users] Restarting with PNPN using PNPN-2 results
> Message-ID:
>         <mailman.23731.1465844987.24727.nek5000-users at lists.mcs.anl.gov>
> Content-Type: text/plain; charset="utf-8"
>
> Hello neks,
>
> I have a DNS simulation that I was running as PNPN-2 and I would like to
> run it as PNPN without starting back at the initial conditions.
> I am curious if anyone has experience running a PNPN case with initial
> conditions from a PNPN-2 simulation?  I have tried this and the simulation
> runs okay, except the pressure term never seems to converge when running
> the PNPN simulation.  This dramatically increases runtime and makes the
> results questionable.  It seems to me that restarting with PNPN-2 results
> should be doable, and that it should only take a few iterations for the
> pressure term to start converging.  However, my simulation is still not
> converging after 100+ iterations, but it was converging fine when I was
> running PNPN-2.
>
> Any suggestions for things I should look for that might cause the problem,
> or potential fixes?
>
> Thanks
> --
> Phil Sakievich
>
> PhD Candidate - Mechanical Engineering
> Arizona State University - Ira A. Fulton School for Engineering of Matter
> Transport and Energy
> Tempe, Arizona
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <
> http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20160613/790d807b/attachment-0001.html
> >
>
> ------------------------------
>
> Message: 2
> Date: Mon, 13 Jun 2016 19:15:32 +0000
> From: nek5000-users at lists.mcs.anl.gov
> To: "nek5000-users at lists.mcs.anl.gov"
>         <nek5000-users at lists.mcs.anl.gov>
> Subject: Re: [Nek5000-users] Restarting with PNPN using PNPN-2 results
> Message-ID:
>         <mailman.23735.1465845499.24727.nek5000-users at lists.mcs.anl.gov>
> Content-Type: text/plain; charset="iso-8859-1"
>
> Hi Phil,
>
> The regularity of the pressure field is much larger for Pn-Pn as compared
> to Pn-Pn-2 so it should take some time to readjust to the larger resolution.
>
> Aleks
> ________________________________
> From: nek5000-users-bounces at lists.mcs.anl.gov [
> 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: Monday, June 13, 2016 2:09 PM
> To: nek5000-users at lists.mcs.anl.gov
> Subject: [Nek5000-users] Restarting with PNPN using PNPN-2 results
>
> Hello neks,
>
> I have a DNS simulation that I was running as PNPN-2 and I would like to
> run it as PNPN without starting back at the initial conditions.
> I am curious if anyone has experience running a PNPN case with initial
> conditions from a PNPN-2 simulation?  I have tried this and the simulation
> runs okay, except the pressure term never seems to converge when running
> the PNPN simulation.  This dramatically increases runtime and makes the
> results questionable.  It seems to me that restarting with PNPN-2 results
> should be doable, and that it should only take a few iterations for the
> pressure term to start converging.  However, my simulation is still not
> converging after 100+ iterations, but it was converging fine when I was
> running PNPN-2.
>
> Any suggestions for things I should look for that might cause the problem,
> or potential fixes?
>
> Thanks
> --
> Phil Sakievich
>
> PhD Candidate - Mechanical Engineering
> Arizona State University - Ira A. Fulton School for Engineering of Matter
> Transport and Energy
> Tempe, Arizona
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <
> http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20160613/09bbb78b/attachment-0001.html
> >
>
> ------------------------------
>
> Message: 3
> Date: Mon, 13 Jun 2016 21:31:08 +0200
> From: nek5000-users at lists.mcs.anl.gov
> To: nek5000-users at lists.mcs.anl.gov
> Subject: Re: [Nek5000-users] Restarting with PNPN using PNPN-2 results
> Message-ID:
>         <mailman.23743.1465847039.24727.nek5000-users at lists.mcs.anl.gov>
> Content-Type: text/plain; charset=windows-1252; format=flowed
>
> Hi,
>
> In principle, you can do a Navier-Stokes restart without pressure field.
> For PnPn-2 you can for instance do one step with Uzawa (nested
> iterations) to have a good pressure. So usually, when we do a restart of
> a solution wi
> thout pressure (e.g. when doing Arnoldi or Newton) then one Uzawa step
> fixes everything. I am not too familiar with PnPn, but I guess that
> there is a flavour that is not dependent on a previous pressure
> solution. Perhaps a PnPn-expert has some opinion on a Uzawa-like method?
>
> Best regards,
> Philipp
>
> On 2016-06-13 21:15, nek5000-users at lists.mcs.anl.gov wrote:
> > Hi Phil,
> >
> > The regularity of the pressure field is much larger for Pn-Pn as
> > compared to Pn-Pn-2 so it should take some time to readjust to the
> > larger resolution.
> >
> > Aleks
> > ------------------------------------------------------------------------
> > *From:* nek5000-users-bounces at lists.mcs.anl.gov
> > [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:* Monday, June 13, 2016 2:09 PM
> > *To:* nek5000-users at lists.mcs.anl.gov
> > *Subject:* [Nek5000-users] Restarting with PNPN using PNPN-2 results
> >
> > Hello neks,
> >
> > I have a DNS simulation that I was running as PNPN-2 and I would like to
> > run it as PNPN without starting back at the initial conditions.
> > I am curious if anyone has experience running a PNPN case with initial
> > conditions from a PNPN-2 simulation?  I have tried this and the
> > simulation runs okay, except the pressure term never seems to converge
> > when running the PNPN simulation.  This dramatically increases runtime
> > and makes the results questionable.  It seems to me that restarting with
> > PNPN-2 results should be doable, and that it should only take a few
> > iterations for the pressure term to start converging.  However, my
> > simulation is still not converging after 100+ iterations, but it was
> > converging fine when I was running PNPN-2.
> >
> > Any suggestions for things I should look for that might cause the
> > problem, or potential fixes?
> >
> > Thanks
> > --
> > Phil Sakievich
> >
> > PhD Candidate - Mechanical Engineering
> > Arizona State University - Ira A. Fulton School for Engineering of
> > Matter Transport and Energy
> > Tempe, Arizona
> >
> >
> > _______________________________________________
> > Nek5000-users mailing list
> > Nek5000-users at lists.mcs.anl.gov
> > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
> >
>
>
> ------------------------------
>
> Message: 4
> Date: Mon, 13 Jun 2016 16:11:14 -0700
> From: nek5000-users at lists.mcs.anl.gov
> To: nek5000-users at lists.mcs.anl.gov
> Subject: Re: [Nek5000-users] lambda2 in turbChannel
> Message-ID:
>         <mailman.23757.1465860244.24727.nek5000-users at lists.mcs.anl.gov>
> Content-Type: text/plain; charset="utf-8"
>
> Hi Wei,
>
> I do not think so. But you can definitely, modify the lambda2() routine to
> generate the Q criterion. It is quite straightforward.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <
> http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20160613/40eedaf9/attachment-0001.html
> >
>
> ------------------------------
>
> Message: 5
> Date: Tue, 14 Jun 2016 11:21:16 +0200
> From: nek5000-users at lists.mcs.anl.gov
> To: nek5000-users at lists.mcs.anl.gov
> Subject: [Nek5000-users] g2gi() no geometry found
> Message-ID:
>         <mailman.23773.1465903728.24727.nek5000-users at lists.mcs.anl.gov>
> Content-Type: text/plain; charset=ISO-8859-1; DelSp="Yes";
>         format="flowed"
>
> Hello Neks,
>
> I am trying to use the g2gi subroutine, but keep getting the error
> massage ABORT: No geometry found in:
>
>   call userchk
>   grid-to-grid interpolation
>   140   8   8   8 2.0010000E+03 4002 U P                             4
> NELT,NX,N ?a?@
>
>                                   NONSTD HDR, parse_hdr, abort.
>   ABORT: No geometry found in newgeom.fld
>
> call exitt: dying ...
>
> Anybody familiar with this error? Do I give him the wrong files? I
> assumed Nek wants a .fld. Any help or working case as an example would
> be highly appreciated.
>
> Br
> Michael
> e1429308 at student.tuwien.ac.at
>
>
>
>
>
>
> ------------------------------
>
> Message: 6
> Date: Tue, 14 Jun 2016 11:40:05 +0000
> From: nek5000-users at lists.mcs.anl.gov
> To: "nek5000-users at lists.mcs.anl.gov"
>         <nek5000-users at lists.mcs.anl.gov>
> Subject: Re: [Nek5000-users] g2gi() no geometry found
> Message-ID:
>         <mailman.23774.1465904407.24727.nek5000-users at lists.mcs.anl.gov>
> Content-Type: text/plain; charset="us-ascii"
>
>
> Hi Michael,
>
> You need to compile with the ifmpiio flag set in makenek:
>
>
> # (set PPLIST=? to get a list of available symbols)
> PPLIST="MPIIO"
>
> All files should be .f00000 files, which is the current default.  (Make
> certain that your
> current .usr file does not reset param(66) to 4.)   [ You can convert a
> .fld file to a .f0000
> file by running a 0 step, 0 FINTIME job and setting p66=6, p67=4 in the
> .usr file, say, in
> subroutine usrdat(). ]
>
> When running g2gi, make certain that the new file has the geometry in it.
>  The old file does not need the
> geometry, only the data to be interpolated.
>
> hth
>
> Paul
>
>
> ________________________________________
> From: nek5000-users-bounces at lists.mcs.anl.gov [
> 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: Tuesday, June 14, 2016 4:21 AM
> To: nek5000-users at lists.mcs.anl.gov
> Subject: [Nek5000-users] g2gi() no geometry found
>
> Hello Neks,
>
> I am trying to use the g2gi subroutine, but keep getting the error
> massage ABORT: No geometry found in:
>
>   call userchk
>   grid-to-grid interpolation
>   140   8   8   8 2.0010000E+03 4002 U P                             4
> NELT,NX,N ?a?@
>
>                                   NONSTD HDR, parse_hdr, abort.
>   ABORT: No geometry found in newgeom.fld
>
> call exitt: dying ...
>
> Anybody familiar with this error? Do I give him the wrong files? I
> assumed Nek wants a .fld. Any help or working case as an example would
> be highly appreciated.
>
> Br
> Michael
> e1429308 at student.tuwien.ac.at
>
>
>
>
> _______________________________________________
> 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
>
>
> End of Nek5000-users Digest, Vol 88, Issue 13
> *********************************************
>



-- 
Phil Sakievich

PhD Candidate - Mechanical Engineering
Arizona State University - Ira A. Fulton School for Engineering of Matter
Transport and Energy
Tempe, Arizona
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20160614/220906cd/attachment-0001.html>


More information about the Nek5000-users mailing list