Inconsistent Buckling Results in Karamba for Arched Roof Model

Grasshopper Karamba3D
1

Hi Karamba Team,
I’m currently comparing the buckling results of an arched roof model in Karamba and RFEM. The results mostly match well. However, I’ve noticed that the same model occasionally produces different results in Karamba. Interestingly, when I reconnect the model inputs, the correct results appear again.

Could this be due to a bug or some issue with the computation process?

Thanks in advance for your help!

image
image1503×617 171 KB

2

Hi, are you able to share the script? and which version of Karamba3D are you using?

3

Hi Matthew,
It’s quite a large script, so I’m just providing the inputs by internalizing them.
I’m currently using Karamba3D version 3.1.40531.0.
Please let me know additional detail if you need.

250708-KarambaArchedRoofTrial.gh (712.8 KB)

4

Hi, when I open the script unfortunately, the supports are not working properly. but i am taking the points that you saved as the input. I am however not sure if the results match yours.

Also, the results seem to align each time. Do you experience this problem with the reduced script too?

image
image2674×1064 340 KB

5

Hi Matthew,
Thanks for your quick response.

  1. I open Grasshopper, then upload the script, and I get a buckling factor of 0.42 (It is 0.35 in Picture 2). After that, I copy and paste the components (Analysis, Buckling Modes, and List Item). Most of the time, I get correct results but not always (see Picture 1). The correct buckling factor should be 2.37.
  2. Actually, my support definition was working. You defined fixed supports for all nodes are defined with fixed supports, whereas in my setup, they should be sliding in the longitudinal direction, except for one fixed support on each side. Therefore, I reupload the algorithm by showing actual definition of supports to show what I define. We can see if the problem caused by supports. (please see gh file)
  3. Also, I noticed that our support component are different. Is there any update for the support components to define support with translational or rotational spring coefficient?(See Picture 2)
    Murat Ümitcan Acar

250709-KarambaArchedRoofTrial.gh (730.2 KB)

image
image787×691 78.9 KB

image
image1110×750 99.2 KB

6

thanks, I am able to reproduce the error now. We will check this internally and hopefully find a resolution. Thank you for reporting this.

7

Hi @Murat_Ümitcan_Acar just checking your script, you used the Analyse component and not the Analyse Th2 component. This means that your model probably does not have second order normal forces NII defined, unless you defined it in the element itself - however when I checked your model I could not see any defined. When you use the Analyse Th2 component before the BucklingModes component you should see consistent results.

8

Hi Matthew,
Thanks for response and solution. I was previously using the Analyze component and then applying a magnification factor to scale the internal forces. This was because using the AnalyzeThII component significantly increased the computation time ( 5 to 30 minutes, depending on the selected load cases).
I’ll try to implement your suggestion and let you know the results.

9

Hi Murat,

if you know the second theory forces you want to use for the buckling analysis, you can apply this as a property using the ModifyElement component which means that you do not need to run the TH2 analysis.

10

Hi Matthew,
Thanks for additional information,
Just a confirmation, I will take the maximum compressive force for each element.

image
image929×274 36.3 KB

image
image1307×388 7.17 KB

11

Yes, for example, also there are now opportunities to control the settings of the second theory calculation: 3.2.5.3 Load-Case-Combination Settings | Karamba3D v3