Stress Result in a cylinder Shell

About the development of the FEM module/workbench.

Moderator: bernd

User avatar
HarryvL
Posts: 1043
Joined: Sat Jan 06, 2018 7:38 pm

Re: Stress Result in a cylinder Shell

Postby HarryvL » Thu Jan 17, 2019 12:24 pm

jake77 wrote:
Thu Jan 17, 2019 7:30 am
Running the same model on calculix directly and displaying the results using cgx the results are as expected. I guess this is just some typo in the way results are imported from the datasets. Could you please confirm this?
Jake you would have to share the FC file for us to judge the issue. Deleting the result object and clearing the mesh will get it down to reasonable size.

I ran your INP file with CCX and plotted the results both in CGX and FC and see full consistency. The file names explain which is which ...


jake77_FC_disp.png
jake77_FC_disp.png (324.04 KiB) Viewed 193 times

jake77_CGX_disp.png
jake77_CGX_disp.png (20.15 KiB) Viewed 193 times

jake77_FC_PS1.png
jake77_FC_PS1.png (372.76 KiB) Viewed 193 times

jake77_CGX_PS1.png
jake77_CGX_PS1.png (31.85 KiB) Viewed 193 times
User avatar
HarryvL
Posts: 1043
Joined: Sat Jan 06, 2018 7:38 pm

Re: Stress Result in a cylinder Shell

Postby HarryvL » Thu Jan 17, 2019 12:30 pm

PS: your picture shows external pressure, but results in CGX and FC show outward bulging. Is the INP file you uploaded fully consistent with the picture you posted?
User avatar
HarryvL
Posts: 1043
Joined: Sat Jan 06, 2018 7:38 pm

Re: Stress Result in a cylinder Shell

Postby HarryvL » Thu Jan 17, 2019 12:33 pm

PS2: I got a shock when I scaled the displacement result to a factor much bigger than 100. The result requires parental guidance and can only be viewed in the privacy of your PC :shock:
jake77
Posts: 19
Joined: Thu Jan 17, 2019 6:23 am

Re: Stress Result in a cylinder Shell

Postby jake77 » Thu Jan 17, 2019 1:01 pm

HarryvL wrote:
Thu Jan 17, 2019 12:24 pm
jake77 wrote:
Thu Jan 17, 2019 7:30 am
Running the same model on calculix directly and displaying the results using cgx the results are as expected. I guess this is just some typo in the way results are imported from the datasets. Could you please confirm this?
Jake you would have to share the FC file for us to judge the issue. Deleting the result object and clearing the mesh will get it down to reasonable size.

I ran your INP file with CCX and plotted the results both in CGX and FC and see full consistency. The file names explain which is which ...



jake77_FC_disp.png



jake77_CGX_disp.png



jake77_FC_PS1.png



jake77_CGX_PS1.png
Harry,

thanks for looking at this. I used the 0.17 version which I then abandonned as proposed by @berndt. I never saw any suspicious using the newer version. Its also possible I made somewhere an error the rookie I am. :roll:
jake77
Posts: 19
Joined: Thu Jan 17, 2019 6:23 am

Re: Stress Result in a cylinder Shell

Postby jake77 » Thu Jan 17, 2019 1:03 pm

bernd wrote:
Thu Jan 17, 2019 10:42 am
jake77 wrote:
Thu Jan 17, 2019 10:01 am
OK I see you have it regularly updated. I'll fire it up then.
you can get it here for windows, just unpack and run it https://github.com/FreeCAD/FreeCAD/releases

btw. compiling on windows int not that difficault, since we gone provide precompiled library pack.
Thanks, I already switched.
jake77
Posts: 19
Joined: Thu Jan 17, 2019 6:23 am

Re: Stress Result in a cylinder Shell

Postby jake77 » Thu Jan 17, 2019 5:32 pm

So I spent the day with trying to solve my problem. I attached the model too look on (without mesh as suggested) if anyone is interested.
The model has a simplified geometry, as the half-pipe shaped stiffeners should be a helix. I could create that but the geometry was warped for some reason and meshing was unsuccesful. The attached model is a good aproximation of the real geometry (at least for my aims).

There are a bunch of observations from my experience with the model:

FreeCAD is promising, and the FEM WB is certainly something to consider to use for calculation. Please don't take the following as an insult, I may have pushed a little the boundaries here :-)

Creating the mesh with the settings auto shape, max: 100 mm, min: 20 mm takes about 2 minutes on my laptop, setting max to 50 mm yields meshing time well over 20 minutes. When running the static analysis FreeCAD reports a succesful analysis, athough it fails for some reason, as seen in the output window:

"FEM: No nodes found in Frd file.
Problem on frd file import. No nodes found in frd file.
FEM: No result object in active Analysis."

Perhaps due to model size? I could not figure out based on the ccx printout (attached). A less dense mesh (say max 150, min 20) runs, but provides inaccurate results, as the mesh itself is incorrect (negative jacobians etc.).
I used ccx directly (v 2.10) to run the bucking analysis - the experience is roughly the same.

So making meshing better by providing more control (element type? quad mesh? zones? the ability to mesh only certain regions of the model?) and a faster algorithm would certainly improve the overally experience and usability. I'll check meshing the model using macway later tonight, that has some more options if I remember correctly.

I came across the following issues. During the day I several times re-started the work from scratch, and I'm obviously not an experienced user. This is not not an exhaustive list and I may be wrong on some issues.
- meshing, as elaborated above.
- the defined material sometimes gets lost, probably during save/load.
- the ability to select multiple faces for loading, material, thickness etc. would be great.
- I met the infamous "Pressure on shell mesh at the moment only supported for meshes with appropriate group data." error, althogh the checkbox in Preferences/FEM is checked. Checking out-in this checkbox seems to work. As I get the German version of the GUI, reading the English forum is not always helpful.
- Localisation issues. Is it possible to enforce one?
- Having some feedback from the meshing process (e.g. progress) would be great, watching the memory consumption of gmsh in the task manager if it hangs is not really informative.
- It would be nice to have the output window scroll automatically to always show the last line, or at least enable Ctrl-End to jump to the bottom.
- Is it possible to define constraint on a vertex? One can do it in FreeCAD but is it transferred to ccx? I could not make it work.
Attachments
ccx fails.png
ccx fails.png (12.9 KiB) Viewed 171 times
doppelstock_FEM_tosubmit.FCStd
(51.19 KiB) Downloaded 7 times
User avatar
HarryvL
Posts: 1043
Joined: Sat Jan 06, 2018 7:38 pm

Re: Stress Result in a cylinder Shell

Postby HarryvL » Thu Jan 17, 2019 6:15 pm

@jake77, the analysis ran ok, but no load was applied, as per the following error message:

Code: Select all

Constraint pressure: FemConstraintPressure
Pressure on shell mesh at the moment only supported for meshes with appropriate group data.
[('FemConstraintPressure: face load', [])]
Writing time input file: 2.202442
 
and this is the relevant part of the INP file:

Code: Select all

** FemConstraintPressure
*DLOAD
** FemConstraintPressure: face load

***********************************************************
** Outputs --> frd file
** written by write_outputs_types function
*NODE FILE, OUTPUT=2d
U
*EL FILE
S, E
Now what "appropriate group data" are and how to create them I don't know. For that I have to refer to the Gurus.

PS: on my system the mesh took 25 seconds to generate and the CCX analysis tool 100 seconds.
User avatar
bernd
Posts: 8320
Joined: Sun Sep 08, 2013 8:07 pm
Location: Zürich, Switzerland

Re: Stress Result in a cylinder Shell

Postby bernd » Thu Jan 17, 2019 7:05 pm

Code: Select all

v 2.10
Do you really use ccx 2.10 !!! This is years old. You should use 2.15 or at least 2.14

For all the rest i will give some input later on ...
User avatar
bernd
Posts: 8320
Joined: Sun Sep 08, 2013 8:07 pm
Location: Zürich, Switzerland

Re: Stress Result in a cylinder Shell

Postby bernd » Thu Jan 17, 2019 7:23 pm

runs here too apart from the pressure problem (which is a FreeCAD limitation ATM for curved shell geometry and CalculiX DLOAD) ... has quite funny mode shapes ...

Screenshot_20190117_202128.png
Screenshot_20190117_202128.png (1007.63 KiB) Viewed 160 times
User avatar
HarryvL
Posts: 1043
Joined: Sat Jan 06, 2018 7:38 pm

Re: Stress Result in a cylinder Shell

Postby HarryvL » Thu Jan 17, 2019 7:52 pm

Why funny? That’s how I would expect it to vibrate.