pyNastranGUI and mixed plate properties

53 views
Skip to first unread message

clarkbr...@gmail.com

unread,
Dec 12, 2024, 6:31:36 PM12/12/24
to pyNastran Discuss
Steve,
Here is a new one and well posed I think. The original complaint arises with one of our big airplane models where pyNastranGUI fails to draw composite plate stresses. I think I have it duplicated in the attached smaller plate model.
I am using pyNastranGUI_1.5.0-dev-687863558_2024-11-23.exe. You made this in response to my submission 23 Nov 2024-ish. It still seems to be the offered latest version.
I am solving with MSC Nastran 2021.1.
This is related to the number of subcases when there are both isotropic material plates and composite laminate plates. With one subcase, the results navigator will offer both Plate Stress and Composite Plate Stress. With two subcases, the results navigator fails to offer the Composite Plate Stress.
The console offers this complaint about the laminate results, but I don't see what it wants.

[2024-12-12 16:05:24] ERROR : stress.py:959 eids = [1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014
1015 1016 1017 1018 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010
1011 1012 1013 1014 1015 1016 1017 1018]
type=RealCompositePlateStressArray nelements=18 ntotal=72
data: [1, ntotal, 9] where 9=[o11, o22, t12, t1z, t2z, angle, major, minor, max_shear]
element_layer.shape = (72, 2)
data.shape = (1, 72, 9)
element type: QUAD4LC-composite-95
sort1
lsdvmns = [1]; dtype=int64
eids = [1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014
1015 1016 1017 1018]
element_layer:
[[1001 1]
[1001 2]
[1001 3]
[1001 4]
[1002 1]
[1002 2]
[1002 3]
[1002 4]
[1003 1]
[1003 2]
[1003 3]
[1003 4]
[1004 1]
blah blah blah
[1017 4]
[1018 1]
[1018 2]
[1018 3]
[1018 4]]

With one subcase, the console doesn't gripe.

Sincerely,

Clark

FlatPlatePCOMPtest.zip

Steven Doyle

unread,
Dec 13, 2024, 11:11:16 AM12/13/24
to pynastra...@googlegroups.com
Yeah that’s weird…

I fixed the initial problem, but created another bug, so I’ll need to figure out what that one is messed up.

I guess, yeah the error was the same element id was used in two different element types (CTRIA3/CQUAD4) and that happens because the quads from subcase 1 and 2 are both grabbed.

--
You received this message because you are subscribed to the Google Groups "pyNastran Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pynastran-disc...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/pynastran-discuss/5b7d0286-86e7-4d14-b18c-7724ec4ffe52n%40googlegroups.com.

clarkbr...@gmail.com

unread,
Dec 18, 2024, 11:49:12 AM12/18/24
to pyNastran Discuss
Steve,
It looks like your initial investigation was fruitful, but your message of Dec 13 doesn't sound like a complete fix has been implemented. I will want to post my update at school by the end of January to support next semester's classes and it needs to include a fixed/working pyNastranGUI. Like before, I would be very happy to use the "latest" based on an exe from the dev branch.
Sincerely,
Clark

Steven Doyle

unread,
Dec 18, 2024, 11:23:38 PM12/18/24
to pynastra...@googlegroups.com
Clark,

The fix is not implemented. A number of results were dropped. I need to diff the previous version to find out why. I’m busy right now, but will have some time soon.

Steven Doyle

unread,
Jan 13, 2025, 9:35:02 AMJan 13
to pynastra...@googlegroups.com

Best,
Steve

clarkbr...@gmail.com

unread,
Jan 13, 2025, 11:04:32 AMJan 13
to pyNastran Discuss

Steve,

Thanks so much. Just in time. I’ll fetch it, test it out, and include it in this semester’s distribution soonest.

Sincerely,

Clark

clarkbr...@gmail.com

unread,
Jan 13, 2025, 4:02:47 PMJan 13
to pyNastran Discuss
Steve,
If I launch it, it crashes before opening the gui. If I launch it in a cmd window, I can catch the following complaint:
C:\Users\clark\clark\pyNastranGUI_1.5.0-dev2>pyNastranGUI_1.5.0-dev2-2b5d5771_2025-1-12.exe
Traceback (most recent call last):
  File "gui.py", line 19, in <module>
  File "C:\NASA\dev\pyinstaller\PyInstaller\loader\pyimod02_importers.py", line 384, in exec_module
  File "main_window.py", line 31, in <module>
  File "vtkmodules\__init__.py", line 58, in exec_module
  File "vtkmodules\__init__.py", line 318, in on_vtk_module_init_completed
  File "importlib\__init__.py", line 90, in import_module
ModuleNotFoundError: No module named 'vtkmodules.util.execution_model'
[PYI-12772:ERROR] Failed to execute script 'gui' due to unhandled exception!

I figure you left out a vtk module when you bundled it all up in an exe. I don't think of anything I can do on my end.
Please advise.
Clark

Steven Doyle

unread,
Jan 14, 2025, 12:11:57 AMJan 14
to pynastra...@googlegroups.com
Clark,

It should be fixed. Exe building on python is a bit of a whack-a-mole process. I downgraded the vtk version to 9.3 instead of 9.4, which I think I used last ~April. There's a weird OpenGL issue on Windows Sandbox, but I think that's just the test environment.


Best,
Steve

clarkbr...@gmail.com

unread,
Jan 17, 2025, 12:00:12 PMJan 17
to pyNastran Discuss
Steve,
When I try to load geometry from either of the .dat files in the zip attached to the first message here, pyNastranGUI dev3 fails. In the log window it says

[2025-01-17 09:36:40] INFO : load_actions.py:116 reading None file 'H:/FEA4JET/pyNastranGUItests/FlatPlatePCOMPtest/Flat_plate_tip_loads_mixed_1case.dat'
[2025-01-17 09:36:40] INFO : nastran_io.py:590 nnodes=50 nelements=36
[2025-01-17 09:36:40] ERROR : load_actions.py:153 Traceback (most recent call last):
File "qt_files\load_actions.py", line 139, in on_load_geometry
File "converters\nastran\gui\nastran_io.py", line 491, in load_nastran_geometry
File "converters\nastran\gui\nastran_io.py", line 591, in load_nastran_geometry_unvectorized
File "bdf.py", line 3624, in get_bdf_stats
File "bdf_interface\stats.py", line 200, in get_bdf_stats
AssertionError: ['allow_overwrites_set']

[2025-01-17 09:36:40] COMMAND : load_actions.py:183 self.on_load_geometry(infile_name='H:/FEA4JET/pyNastranGUItests/FlatPlatePCOMPtest/Flat_plate_tip_loads_mixed_1case.dat', geometry_format='nastran')


In my Spyder with Python 3.10 environment, I can read_bdf  either of those dat files and then get_bdf_stats ok. I seem to be using pyNastran-1.4.0%2Bdev.0f4554d20-py3-none-any.whl in Spyder. 

Clark

steve

unread,
Jan 20, 2025, 6:13:19 PMJan 20
to pyNastran Discuss
Clark,

I pushed a new build pyNastranGUI_1.5.0-dev4-644cd28a0_2025-1-20 that fixes that allow_overwrites_set bug. It's a new feature, but I didn't think it would break things.

The  1.4.0%2Bdev.0f4554d20  version is a year out of date, but there's been some nice enhancements in the PSHELL/PCOMP post-processing since then. You can look at upper/lower surface results and see how converged you are by looking at max/min/diff/etc.
../_images/results_plate_stress_nodal.png

Steve
Reply all
Reply to author
Forward
0 new messages