Modelling simple elastic shells

brisidi
Posts: 9
Joined: Fri Oct 02, 2020 4:04 pm

Modelling simple elastic shells

Post by brisidi » Fri Oct 02, 2020 4:16 pm

Hello! I am trying to model a reinforced concrete slab in bending. As a first step I thought it would be a good idea to create a linear model. I did the following:

1. imported geometry from an iges file I created using Autocad.
2. I used "planar" to create the surfaces
3. I created elasticMembranePlateSection and assigned it to all the shells
4. I created MITC4 shell and assigned it to all elements.
5. I did the rest: conditions, steps etc.
6. I used structured, Quad, Linear mesh (although I tested also other combinations).

Boundary conditions seem to be adequate, and the analysis runs with no error, but I don't see the deformed shape when I open the database. I see that the values of deflections are enormous (to the order E10). So I think there is a problem with the model. Do I need to "sew faces" in this case? I sewed them, but the problem persists.

Here is what the model looks like:

Image

Note that there is a shell element in every four nodes. Any help would be appreciated! THanks!

STKO Team
Posts: 3066
Joined: Tue Oct 29, 2019 8:45 am

Re: Modelling simple elastic shells

Post by STKO Team » Mon Oct 05, 2020 2:52 pm

Dear User, could you please post your *.SCD file? It's difficult to tell what the problem is, just by looking at the image.
Anyway, you should make sure of the following:
  • geometries should be joined
  • make sure all material properties are well- defined (E, nu, thickness)
  • make sure all boundary conditions are defined and assigned into a constraint pattern in the analysis steps

brisidi
Posts: 9
Joined: Fri Oct 02, 2020 4:04 pm

Re: Modelling simple elastic shells

Post by brisidi » Tue Oct 06, 2020 1:04 pm

Thank you for your reply! The requested file is here: https://www.dropbox.com/s/c2i1rfdqb0j0u ... t.scd?dl=0
Thank you in advance!

STKO Team
Posts: 3066
Joined: Tue Oct 29, 2019 8:45 am

Re: Modelling simple elastic shells

Post by STKO Team » Wed Oct 07, 2020 8:46 am

Dear user,
I just tested your file and it runs fine, and the results are meaningful.. the maximum deflection at the center is about 0.0187.
Could you post a picture of the output you obtain?
What opensees version are you using?

brisidi
Posts: 9
Joined: Fri Oct 02, 2020 4:04 pm

Re: Modelling simple elastic shells

Post by brisidi » Thu Oct 08, 2020 8:45 am

Thank you for your time! That's good to hear! I am using OpenSees Version 3.2.2 64-Bit. In the post-processor I don't see the deformed shape but the legend shows a very large displacement magnitude. I am not sure if I am doing something wrong in the post-processor, here is what I get (maybe I will have to post-process the results manually):
Image

STKO Team
Posts: 3066
Joined: Tue Oct 29, 2019 8:45 am

Re: Modelling simple elastic shells

Post by STKO Team » Fri Oct 09, 2020 7:56 am

Ok, so if the SCD file we used is the same, it means that there is a difference in the opensees version we are using. Please, could you tell me how did you obtained your opensees exe? From the official website? GitHub? Did you compile it from source code?

brisidi
Posts: 9
Joined: Fri Oct 02, 2020 4:04 pm

Re: Modelling simple elastic shells

Post by brisidi » Fri Oct 09, 2020 10:24 am

I obtained the Opensees executable from the official site (i.e., pre-compiled): https://opensees.berkeley.edu/OpenSees/ ... wnload.php.
Should I perhaps use an older version of OpenSees?
*Edit: Also, please notice that I am using an Evaluation License. Have I perhaps exceeded the number of nodes allowed for this license?

STKO Team
Posts: 3066
Joined: Tue Oct 29, 2019 8:45 am

Re: Modelling simple elastic shells

Post by STKO Team » Mon Oct 12, 2020 9:02 am

*Edit: Also, please notice that I am using an Evaluation License. Have I perhaps exceeded the number of nodes allowed for this license?
No, the evaluation (trial) license has no limitation.
I obtained the Opensees executable from the official site (i.e., pre-compiled): https://opensees.berkeley.edu/OpenSees/ ... wnload.php.
I just donwloaded it today, and copied the hdf5.dll szip.dll and zlib.dll inside the same folder where I downloaded OpenSees.exe (c:\opensees\OpenSees3.2.2-x64.exe\bin\), and the output it still fine:
shell.png
shell.png (209.46 KiB) Viewed 4931 times
One more question: how did you obtain the HDF5 libraries for the MPCO recorder? Are you using the ones provided in the STKO installation directory? Or, are you using another (maybe incompatible) version?

brisidi
Posts: 9
Joined: Fri Oct 02, 2020 4:04 pm

Re: Modelling simple elastic shells

Post by brisidi » Tue Oct 13, 2020 10:11 am

Thanks a lot for your help! I am using the dll files from the original installation directory. I don't know what the problem is, but I was able to make it work by running "source main.tcl" directly in OpenSees. This produces a database file which can than be loaded into STKO and the result is the same as the one you showed in the picture. I think my issue is in the last step before clicking "run analysis". It seems that STKO is not linking correctly to my OpenSees.exe file. In "Kits", I think I have loaded the correct OpenSees.exe under "Manual", but apparently I have done something wrong or it is not working correctly.

After I click run analysis, a terminal window pops up and closes soon after (I don't see what it says). The message in STKO's terminal is:

Code: Select all

Done.Input file correctly written!
Running ""F:/OpenSees/bin/OpenSees.exe" ""  & pause" in "F:/OpenSees/bin"
End Command: RunAnalysis
It seems that it is not running the analysis successfully, and maybe the database I was loading earlier was an older one with an incorrect model. Anyway, my problem is solved (the workaround I found is OK for me) and thank you for your time! Also, keep up the great work! STKO is a superb product!

P.S:
In case you want to investigate further to make sure that this is indeed not an issue of STKO but just an issue with my computer or my use, I am ready to provide more info. I was thinking that possible culprits could be the fact that I have several OpenSees.exe files in my machine or the fact that my harddisk is divided into several partitions. I could also be just a mistake I am making in the "kits" form or somewhere else.

STKO Team
Posts: 3066
Joined: Tue Oct 29, 2019 8:45 am

Re: Modelling simple elastic shells

Post by STKO Team » Tue Oct 13, 2020 10:34 am

After I click run analysis, a terminal window pops up and closes soon after (I don't see what it says).
STKO tries to run your OpenSees.exe using the "& pause" post-command.
Running ""F:/OpenSees/bin/OpenSees.exe" "" & pause" in "F:/OpenSees/bin"
It basically should keep the cmd window open so that the user can see the ouptut. However if the first command (in this case OpenSees.exe mai.tcl) crashes, the "& pause" command has no effect.

If it closes, it means that you have an error in launching OpenSees.exe. The best way to understand what is going on is:
  • Create a "run_opensees.bat" file where you have OpenSees.exe (in F:/OpenSees/bin)
  • Edit it in a text editor and write the following 2 lines:

    Code: Select all

    OpenSees.exe %1
    pause
  • Now go the STKO Kits command, and load the a new manual kit that will now point to the "run_opensees.bat" file instead of the OpenSees.exe executable.
In this way it should keep the window open even if you have an error. SO that you can read what is says. Then, post the output here

Post Reply