Error in previously saved files after patching and starting calculation

Hi all,

I am working on a 2D transient model with energy and species transport. I first converged flow and turbulence in a steady state model. After that, I calculated a solution in a transient model for the energy equation and now I want to calculate the transient solution for species transport. However, I entered all of my settings for species transport, than saved the case and data file before patching and solving, because I know that things can go wrong in those steps.  

Afterwards, I saved the case and data file under a new name and I patched the values I needed in this new file. During patching I got error messages saying the turbulent viscosity ratio was limited to 1e5 and when I start the solution I got an MPI error:

Thinking that I could start over from the case and data file that I saved before patching and solving, I opened this previous file. However, opening this file (which previously gave no problems) resulted in an error and I could not use this file anymore and had to start over.

 I have had this problem multiple times and it is very time consuming to have to start over and over again, always entering the same settings, hoping it won’t result in an error that corrupts all of my files.

 

Can anyone help me deal with this problem? What is the reason for these errors and for the error in the original file before patching and can I prevent having to start over with entering the settings each time such an error occurs? I am pretty new to CFD, so making some mistakes is almost inevitable for me.

Thanks in advance for the help.

Kind regards,

Anne-Sophie

Best Answer

Answers

  • Let me try to understand your issue better. When you solve just the flow, turbulence, and energy equations, you are able to solve for a steady state solution. You then save your files (case and data). You now switch on the transient solver, set up your species transport, and attempt to run the solution. You are however getting a error message.

    Is this the issue? Please let me know if I did not understand your question correctly.

    Thanks.

    Karthik

  • Hi Karthik,

    No, sorry I probably didn't explain it well:

    1) I solve flow and turbulence for a steady state solution.

    2) I switch to transient and run the energy equation. I stop the solution after 0.00001s.

    3) I activate and set up the settings for my species transport in this same file. Before patching or solving, I save this file =file 1. No problems occur.

    4) I take a copy of file 1, saving it as file 2. In file 2 I patch the values (which gives warnings about turbulent viscosity ratio) and run the solution, which results in the MPI-error. Due to the MPI-error Fluent freezes and I cannot do anything else than closing file 2 via task manager. After this, file 2 is unusable: the MPI-error and freezing of Fluent appears again if I reopen file 2.

    5) Because file 2 is unusable, I try to reopen file 1 to alter some settings, hoping that I can find out which setting was wrong and caused the MPI-error. However, also opening file 1 results (just by opening the file) in the MPI-error and in a total freeze of Fluent.

    As a result I need to start over from step 2, setting up my species transport all over again. Since I don't know the reason for the MPI-error and since I am quite new to CFD, chances are that I make the same mistake that resulted in this error and that I will have to start over (losing file 1 and 2, going back to step 2) again and again.

    My questions are:

    • What is most likely the reason for such an MPI-error? Has this something to do with the turbulent viscosity warning and how can I fix this?
    • Why do I get the MPI-error and freezing of Fluent in file 1, whereas I didn't have any problems with this file before?
    • Is there a way to recover file 1 somehow?

    I hope this makes it more clear?

    Kind regards,

    Anne-Sophie

  • RobRob UKForum Coordinator

    If you're not using parameters copy the case & data files out of the Workbench folder structure and see if you can open them using Fluent in standalone mode. If you're lucky it's the workflow part of the system that's broken and not the actual Fluent files.

    The mpi error is just a node failing, it's the SIGSEV that's the problem. I suspect it's linked to the turning on/off of the equations, do you have any UDFs in the model?

  • Hi Rob,

    Thanks for the advice. Unfortunately, I am already working in Fluent standalone, so that doesn't solve the problem.

    I am using UDFs, could they be the problem?

    Kind regards,

    Anne-Sophie

  • Indeed, I checked my UDF and saw that there was a UDS that I had not yet defined that I needed for the calculation of a source term for one of the species.

    Thank you for your help!

Sign In or Register to comment.