Shell dome from "wood" panels - inaccurate results?

Hi,

I'm doing a structural analysis comparaison for a wood panels dome between Karamba and RFEM. My ultimate purpose is to demonstrate the advantage of working with Karamba in a iterative design process for non-standard wood construction in CLT by using structural analysis as input to optimize the form.

I first encounter very different values between the Karamba and RFEM models (at this point, I suspect the material proprieties - isotropic and calculation method to explain the difference). Even if annoying it was not discarting the benefit of Karamba since the deformation and force distribution were similar to the RFEM model. As conclude, I still can use the ratio of the values to optimize my dome panel pattern.

Unfortunatly, after doing a cycle of optimization, the new dome show, in RFEM, great amelioration, but I obtain the exact opposite result in Karamba. This one showing greater deformation than the original basic pattern. I'm really can't explain this difference, and it's was certainly not the kind of result that I want to show. I could only wonder if there is a problem with my script or if there is a bug somewhere in Karamba programming.

I join my grasshopper file. Use the boolean toogle to pass from the basic pattern analysis to the optimize one. I also join the RFEM models in JPEG.

Thank you very much in advance for the help

Attachments:

Replies to This Discussion

Permalink Reply by Karamba3D on March 22, 2017 at 2:29pm

Hi Zoe,

when I activate the optimized shape and set the orientation of the 'Live Load' to global I get a maximum displacement of 7.4mm when using second order theory (see attached definition).

The difference of results for the basic geometry is strange though. Could you send me the corresponding RFEM-file?

Best,

Clemens

Attachments:

wooddome_cp.gh, 392 KB

Permalink Reply by Zoe Tolszczuk-Leclerc on March 22, 2017 at 3:42pm

Thank you very much for looking at it! It's indeed the basic geometry which seem to have a problem.

There the RFEM files: https://drive.google.com/open?id=0B0PSO3X5UkfMXzhldVNZdG1vTFE

Unfortunatly I was not able to join it to the message

Permalink Reply by Karamba3D on April 7, 2017 at 6:25am

Dear Zoe,

in the rfem model 'Optimize dome' there are no loads in LC2 - LiveLoad.

When adding the live load for 'Optimize dome' in RFEM I get the following displacements:

LC1 Dead weight: umax(BasicDome) = 0.9mm; umax(OptimizeDome) = 1.7mm

LC2 LiveLoad: umax(BasicDome) = 8.0mm; umax(OptimizeDome) = 15.1mm

Best,

Clemens

Permalink Reply by Zoe Tolszczuk-Leclerc on April 12, 2017 at 4:19pm

Hi,

Thank I indeed totaly miss this part. I work again on the model, still trying to optimize my panel pattern. But again I found some inconsistency in the results.

All for the case load 1.25D+1.5L+1S

I'll join the recap of my results in a comparative table (jpg.). If it "fits" for the M1, which was the previous model, it didn't for the M2. This time Karamba give me less deformation, but RFEM more. I was more carefull I didn't forget load this time.

Moreover, any idea what could explain such difference in the values between Karamba and RFEM? Find hard to believe it's only the material. Am I wrong in the setting of my load case in Karamba?

For the RFEM files: https://drive.google.com/open?id=0B0PSO3X5UkfMXzhldVNZdG1vTFE

Thank again for the help! It's really appreciated!

Zoé

Attachments: