[Nek5000-users] Rayleigh Benard with symmetry BCs

Mon Feb 18 13:57:19 CST 2013

David-

Here is the .box file I used to generate the mesh.  I noticed the CFL target number issue, it looks like nek resets it to 0.5 anyway.

                                      -Jared

CCS2/CNLS Postdoctoral Research Associate
Los Alamos National Laboratory
________________________________
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, February 18, 2013 11:24 AM
To: nek5000-users at lists.mcs.anl.gov
Subject: Re: [Nek5000-users] Rayleigh Benard with symmetry BCs

Hi Jared,

If you send me your .box file, I will check that just in case.

Also, I see you are using variable time steps, so try reducing the target Courant number to 1/2.

Another thing to be aware of with variable time stepping: if the Rayleigh number is large and the simulation starts from a static state, the exponentially growing velocities might still cause the CFL condition to be violated. The solution to this is to decrease the maximum allowed time step, DT. This is unlikely to be your problem at the current Rayleigh number, however.

-David

On Mon, Feb 18, 2013 at 11:53 AM, <nek5000-users at lists.mcs.anl.gov<mailto:nek5000-users at lists.mcs.anl.gov>> wrote:
My apologies upfront for any simple misconceptions as I'm new to nek.

I have successfully run the included 2d RB example and thought things looked reasonable.  I then changed the mesh to an aspect ratio 2, 4x8 element mesh and modified the boundary conditions to be 'SYM' sidewalls and 'SYM' (free-slip) for the top and bottom velocity (see attached rea and map file).

Without changing anything else from the example (I did not modify SIZE) the velocities appear to blow up, and I get a series of errors like the following in the logfile:

     1   100 **ERROR**: Failed in HMHOLTZ: TEMP   1.4999E-02   3.7064E+01   2.4836E-03
          1   2.0000E-02  6.0370E-03 Heat done

     1   100 **ERROR**: Failed in HMHOLTZ: VELX   2.8743E-21   1.7086E-16   2.3416E-23
  0.000000000000000E+000  p22            1           1
 New CG1-tolerance (RINIT*epsm) =   1.010379228335448E-010
  2.341589846098494E-023

     1   100 **ERROR**: Failed in HMHOLTZ: VELY   1.0837E-01   1.0104E+03   1.0104E-10
 New CG2-tolerance (RINIT*10-5/10-10) =   2.795810823528297E-003
  3.472834773690102E-025

Do I need to adjust the tolerances because the mesh/BCs are different?

I don't think it matters, but I ran this on 16 procs (there are 16 per node on the machine I'm using).

                                               -Jared

CCS2/CNLS Postdoctoral Research Associate
Los Alamos National Laboratory

_______________________________________________
Nek5000-users mailing list
Nek5000-users at lists.mcs.anl.gov<mailto:Nek5000-users at lists.mcs.anl.gov>
https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users

On Mon, Feb 18, 2013 at 11:53 AM, <nek5000-users at lists.mcs.anl.gov<mailto:nek5000-users at lists.mcs.anl.gov>> wrote:
My apologies upfront for any simple misconceptions as I'm new to nek.

I have successfully run the included 2d RB example and thought things looked reasonable.  I then changed the mesh to an aspect ratio 2, 4x8 element mesh and modified the boundary conditions to be 'SYM' sidewalls and 'SYM' (free-slip) for the top and bottom velocity (see attached rea and map file).

Without changing anything else from the example (I did not modify SIZE) the velocities appear to blow up, and I get a series of errors like the following in the logfile:

     1   100 **ERROR**: Failed in HMHOLTZ: TEMP   1.4999E-02   3.7064E+01   2.4836E-03
          1   2.0000E-02  6.0370E-03 Heat done

     1   100 **ERROR**: Failed in HMHOLTZ: VELX   2.8743E-21   1.7086E-16   2.3416E-23
  0.000000000000000E+000  p22            1           1
 New CG1-tolerance (RINIT*epsm) =   1.010379228335448E-010
  2.341589846098494E-023

     1   100 **ERROR**: Failed in HMHOLTZ: VELY   1.0837E-01   1.0104E+03   1.0104E-10
 New CG2-tolerance (RINIT*10-5/10-10) =   2.795810823528297E-003
  3.472834773690102E-025

Do I need to adjust the tolerances because the mesh/BCs are different?

I don't think it matters, but I ran this on 16 procs (there are 16 per node on the machine I'm using).

                                               -Jared

CCS2/CNLS Postdoctoral Research Associate
Los Alamos National Laboratory

_______________________________________________
Nek5000-users mailing list
Nek5000-users at lists.mcs.anl.gov<mailto:Nek5000-users at lists.mcs.anl.gov>
https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20130218/9b234130/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: my_ray.box
Type: application/octet-stream
Size: 638 bytes
Desc: my_ray.box
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20130218/9b234130/attachment.obj>