Hyperstudy problem with Flux solver batching
Hi
I have used hyperstudy with flux solver for a while. with the new version, I get into a new problem.
In the DOE, sometimes it will run into an error (attached picture).
When I increase the number of multi execution, the number of failure increase drastically.
I checked the files of the instances that have errors:
1. the flux file was not created at all or 2. the resulted Hstp file is not created.
Before this error, it will give this warning too:
[64612|1642966487198|1|WARNING ][UseCase|UseCaseManager.changeUseCase] : No use case transaction found
[64687|1642966645613|1|WARNING ][ExecuteBatchSpyAction|ExecuteBatchSpyAndExitAction.doAction] : Execution of C:/Program Files/Altair/2021.2/hwdesktop/hst/scripts/python/python3.5/win64/alt/hst/eac/cmd/hstupdate_flux.py failed: Traceback (most recent call last): File "C:/Program Files/Altair/2021.2/hwdesktop/hst/scripts/python/python3.5/win64/alt/hst/eac/cmd/hstupdate_flux.py", line 585, in <module> main() File "C:/Program Files/Altair/2021.2/hwdesktop/hst/scripts/python/python3.5/win64/alt/hst/eac/cmd/hstupdate_flux.py", line 550, in main hstUpdate.updateProject() File "C:/Program Files/Altair/2021.2/hwdesktop/hst/scripts/python/python3.5/win64/alt/hst/eac/cmd/hstupdate_flux.py", line 550, in main hstUpdate.updateProject() File "C:/Program Files/Altair/2021.2/hwdesktop/hst/scripts/python/python3.5/win64/alt/hst/eac/cmd/hstupdate_flux.py", line 453, in updateProject raise ex
Thank you in advance for your kind support.
Answers
-
Hello,
The failed runs can be due to unfeasible designs (e.g. failure to modify the geometry), especially if the solved model is not saved under the run directory. I see on the screenshot that an exception is raised when updating the model.
The warnings are not blocking. You can ignore them.
You said that this behavior occurs with the new version. Does it mean that the same model is solved correctly with a previos version?Or it is another model?
I'd like to test it on my machine. Would it be possible to share the HyperStudy archive (*.hstx) or a zip with the Flux files (*.F2HST.FLU and *.F2HST, and postprocessing python file if any)?
If you prefer not sharing your files within the forum, please feel free to open a support request and upload there the files. I'll take them from there.
Looking forward your feedback,
Best regards,
Diana
0 -
Diana M._20503 said:
Hello,
The failed runs can be due to unfeasible designs (e.g. failure to modify the geometry), especially if the solved model is not saved under the run directory. I see on the screenshot that an exception is raised when updating the model.
The warnings are not blocking. You can ignore them.
You said that this behavior occurs with the new version. Does it mean that the same model is solved correctly with a previos version?Or it is another model?
I'd like to test it on my machine. Would it be possible to share the HyperStudy archive (*.hstx) or a zip with the Flux files (*.F2HST.FLU and *.F2HST, and postprocessing python file if any)?
If you prefer not sharing your files within the forum, please feel free to open a support request and upload there the files. I'll take them from there.
Looking forward your feedback,
Best regards,
Diana
Dear Diana,
No, I haven't checked this model with previous models.
I thought the same thing about model geometry errors. But when I checked the files, I saw that some of them are solved and only the resulting file is not created.If I use only one execution at a time, the failure will be 1 in every 10 cases. If I use multi execution, the failure will be 7 out of 10 cases. I don't think the model file is the problem here. there is something odd with the solver.
I made a case support case, with attached example files.
Regards,
Ehsan0
Altair Employee