Rotating-drum with Mfix 2.3.1 doesn't run without any error

[Silence-017]

Dear teachers, hello, I want to run a case of a rotating drum in my windows 22.3.1. I can build the solver successfully and start running without any error, but it can not run.It always stopped by saying"Previous MFiX run is resumable. Reset job to edit model. MFiX process has stopped"
Can you help me to find out what might be the problem? Thank you very much.
rotating_drum.zip (27.5 MB)

[cgw]

Hello Silence

This problem turns out to be an interesting combination of problems.

First off, I noticed that you are building for SMP. As a general principle, if you are having issues, first turn off SMP/DMP. However doing this did not help.

Secondly - for troubleshooting, try disabling UDFs. But for this case, that does not help either.

On Linux there is a build option for Debug build, for some reason this is disabled on Windows - I enabled it for the upcoming 22.4 release. However, running your model with that option enabled also did not yield any clues.

Finally - I remember to examine the .LOG files and I see:

 From : PARSE_RESID_STRING
 Error: Phase index 1 in RESID_STRING P1   out of bounds.
 **********************************************************************

 From : PARSE_RESID_STRING
 Error: Phase index 1 in RESID_STRING U1   out of bounds.
 **********************************************************************

 From : PARSE_RESID_STRING
 Error: Phase index 1 in RESID_STRING V1   out of bounds.

There are a few problems here - one, these should have turned into popup error messages, and incremented the “Errors” counter in the GUI.

Second - there is a solid phase defined so I don’t see why phase index 1 should be out-of-bounds. I’ll look into these problems. But to move forward I disable the troublesome residual output

before:

after:

Running once again - no errors in ROTATING_DRUM.LOG but the simulation terminates immediately.

Comparing your file with the rotating_drum.mfx in the MFiX distribution, I found the problem. You have set time=2.0, so the simulation reaches the end time as soon as it starts. I should have noticed this in the “Dashboard”

but this is easy to miss - perhaps we should automatically switch to the Dashboard when a job starts.

This value is the starting time of the simulation and generally it does not make sense to set this to anything but 0. This is why there is no control for it in the GUI. (Stopping time, tstop, is set in the `Run panel). You can access this keyword, however, through the ‘Advanced’ pane.

Trying to reset time to 0.0 revealed another small bug, the GUI says that it must be >0. I will also fix this for the upcoming release.

However you can simply delete the keyword from Advanced and you will find that the job runs.

I’m not sure why you set time=2.0 but it looks like you did it in the changes taggerd “Version 5”

Thanks for the bug report! I will follow up separately on the Phase index 1 in RESID_STRING P1 out of bounds error.

– Charles

[Silence-017]

Thank you very much for your attention to my problem.

For time=2.0, I mistakenly assumed that simulating 2s required changing this value to 2.

I will continue to modify the model and look forward to the release of the new version.Thank you again.
-Silence

[Silence-017]

Dear Charles,
I found that once I uncheck this “pure granular flow”, the solver will report an error like this. Can you tell me how to solve this problem?
I would like to make one bottom side of the rotating drum enter the particle and the other bottom side expel the particle, and perform gas-solid reaction in it at the same time. Do you think this is feasible?
Thank you very much!

[cgw]

1. The program is exiting with a zero-division error. If you have any custom Fortran code (UDFs) the problem may be in your code. Check all denominators before dividing.

2. On Linux we are able to produce better error messages showing where the zerod-division happened. On Windows, MFiX currently is not able to pinpoint the file and line (due to available software for debugging). So unfortunately on Windows it’s a little harder to track down these types of errors.

3. Please always include all input files when seeking help - go to the main menu and click on ‘Submit bug report’ - then upload the resulting .zip file.

4. Due to the winter holidays, responses on the MFiX forum may be delayed until next week.

Cheers,

– Charles