search

ucns3d-team / UCNS3D

Unstructured Compressible Navier Stokes 3D code (UCNS3D)
https://ucns3d.com
GNU General Public License v3.0
238 stars 87 forks source link

Problem in restarting the run #34

Open anikm21 opened 1 year ago

anikm21 commented 1 year ago

The restart file is being written according to the WALL CLOCK MAXIMUM TIME or, TOTAL SIMULATION TIME. It cannot be written according to the restart rate. And the more important problem is that when I am restarting my run, it diverges at the first step calculating a very high value of timestep. Please reply me soon. I am adding the history file. history.txt

TakisCFD commented 1 year ago

Make sure that the correct flags are used when restarting from a RESTART file

anikm21 commented 1 year ago

May you suggest me how I can make sure that the correct flags are used? Because when I am restarting my job I only keep the restart file in the directory.

On Fri, 23 Sept 2022 at 20:30, Takis Tsoutsanis @.***> wrote:

Make sure that the correct flags are used when restarting from a RESTART file

— Reply to this email directly, view it on GitHub https://github.com/ucns3d-team/UCNS3D/issues/34#issuecomment-1256327690, or unsubscribe https://github.com/notifications/unsubscribe-auth/AZL6LOSKIUJGSUJJHZDPSRDV7XAX7ANCNFSM57XJUYMQ . You are receiving this because you authored the thread.Message ID: @.***>

anikm21 commented 1 year ago

Thank you. Problem is fixed.

On Tue, 4 Oct 2022 at 09:49, Anik Mandal @.***> wrote:

May you suggest me how I can make sure that the correct flags are used? Because when I am restarting my job I only keep the restart file in the directory.

On Fri, 23 Sept 2022 at 20:30, Takis Tsoutsanis @.***> wrote:

Make sure that the correct flags are used when restarting from a RESTART file

— Reply to this email directly, view it on GitHub https://github.com/ucns3d-team/UCNS3D/issues/34#issuecomment-1256327690, or unsubscribe https://github.com/notifications/unsubscribe-auth/AZL6LOSKIUJGSUJJHZDPSRDV7XAX7ANCNFSM57XJUYMQ . You are receiving this because you authored the thread.Message ID: @.***>

qhma commented 1 year ago

I have the same question when I am restarting my run, it diverges at the first step calculating a very high value of timestep when I run multispeices Euler equation. Please reply me how to fix the problem.

TakisCFD commented 1 year ago

You will need to ensure that :

  1. the MULTISPECIES.DAT file is present in your simulation directory
  2. the RESTART.DAT file is present in your simulation directory
  3. the flags selected for the restart file are correct (such as restarting from steady or unsteady simulation, turbulent or non-turbulent)
  4. the types of equations solved should be -1
  5. the time advancement method should be explicit such as Runge-Kutta3

Hope this helps.

falcaoceg commented 1 year ago

I am sorry to bother you guys again with the same issue, but I'm having the same problem and I would like to know about your solutions for this problem. I see some of you posted that the problem is solved... My simulation is very simple, single component, non-turbulent, unsteady RK2, N-S (all flags are according to it).

I run the simulation, the restart.dat file is generated at the end of the simulation. (not at xxx iterations :/ ).

When I simply run it again in order to restart it, the simulation stops because the simulation "end time" is reached no matter what time limit I impose, of course, the simulation is probably blowing up.

Thanks for your attention.

TakisCFD commented 1 year ago

please send me your ucns3d.dat and multispecies.dat, since i can indicate if something is wrongly setup in the restart.

On Tue, Jan 17, 2023 at 6:13 PM falcaoceg @.***> wrote:

I have the same question when I am restarting my run, it diverges at the first step calculating a very high value of timestep when I run multispeices Euler equation. Please reply me how to fix the problem.

Did you solve the problem?

— Reply to this email directly, view it on GitHub https://github.com/ucns3d-team/UCNS3D/issues/34#issuecomment-1385833049, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADZVYISSWOUEZXVB4J655D3WS3OLRANCNFSM57XJUYMQ . You are receiving this because you commented.Message ID: @.***>

falcaoceg commented 1 year ago

Thank you for your help.

Just as a reminder, I am not the guy performing multispecies simulations, I specified some aspects of my case in the comments.

My simulation is very simple, single component, non-turbulent, unsteady RK2, N-S. At the end, RESIDUAL.dat is generated. At this point, I just re-run it to continue, without success.

best regards -- Carlos

Em ter., 17 de jan. de 2023 às 15:19, Takis Tsoutsanis < @.***> escreveu:

please send me your ucns3d.dat and multispecies.dat, since i can indicate if something is wrongly setup in the restart.

On Tue, Jan 17, 2023 at 6:13 PM falcaoceg @.***> wrote:

I have the same question when I am restarting my run, it diverges at the first step calculating a very high value of timestep when I run multispeices Euler equation. Please reply me how to fix the problem.

Did you solve the problem?

— Reply to this email directly, view it on GitHub <https://github.com/ucns3d-team/UCNS3D/issues/34#issuecomment-1385833049 , or unsubscribe < https://github.com/notifications/unsubscribe-auth/ADZVYISSWOUEZXVB4J655D3WS3OLRANCNFSM57XJUYMQ

. You are receiving this because you commented.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/ucns3d-team/UCNS3D/issues/34#issuecomment-1385840239, or unsubscribe https://github.com/notifications/unsubscribe-auth/A4ALLQDDQPRHUV3ECE6ISVLWS3PB3ANCNFSM57XJUYMQ . You are receiving this because you commented.Message ID: @.***>

TakisCFD commented 1 year ago

Post your ucns3d.dat file and I will indicate what could be wrong

falcaoceg commented 1 year ago

UCNS3D.txt

(I changed the extension to *.txt only to upload here)

TakisCFD commented 1 year ago

CFL should be less than one since it might crash otherwise.

Check Running the same setup for time stepping option 3 until I check that it might be a bug associated with RK2

falcaoceg commented 1 year ago

CFL should be less than one since it might crash otherwise.

Check Running the same setup for time stepping option 3 until I check that it might be a bug associated with RK2

Ok, the RK3 test I already performed but I'll do it again. I am running with this time step for a while now, but i will decrease it in order to verify. Appreciate your advices.

sidkamat commented 1 year ago

Thank you. Problem is fixed. … On Tue, 4 Oct 2022 at 09:49, Anik Mandal @.> wrote: May you suggest me how I can make sure that the correct flags are used? Because when I am restarting my job I only keep the restart file in the directory. On Fri, 23 Sept 2022 at 20:30, Takis Tsoutsanis @.> wrote: > Make sure that the correct flags are used when restarting from a RESTART > file > > — > Reply to this email directly, view it on GitHub > <#34 (comment)>, > or unsubscribe > https://github.com/notifications/unsubscribe-auth/AZL6LOSKIUJGSUJJHZDPSRDV7XAX7ANCNFSM57XJUYMQ > . > You are receiving this because you authored the thread.Message ID: > @.***> >

Hi. Can you tell me how you solved the code diverging on the first time step?