Cladding Panels "load has not been distributed correctly"

[raried]

Hi,

I have searched the forum for help on this but cannot find anything so creating new post:

I have a steel frame roof model with lots of cladding panels to apply the loading. This has worked fine. I then ran a P-Delta analysis and got the warning:

"The load has not been distributed correctly - there are no objects or elements carrying the load overlap."

Clicking on this to highlight the panels on the roof shows that about 50% are affected. The roof is a free-form doubly curved shape.

Any suggestions?

Regards, Rob

[raried]

I have checked the reactions and the load is still in the model. Odd.

My control model (that works well) has a reaction Fz for a particular load case that uses the cladding panels of 

82,632 kN

For the P-delta model with the cladding panel errors it is

82,638 kN

So slightly higher and not 50% lost as expected.

So, what to make of the error message?

[raried]

My mistake. My control model has incorrectly distributed loadings from the clading panels too.

[StefanoPasquini6790]

Hi

Probably your claddings are not flat or are not properly modeled in plane with the substructure. Try to divide in triangles the claddings (there is an API that make this automatically, search in the API box in this forum).

You can also transform your claddings in unmeshed panels with 0 thickness.

Good luck

[raried]

Agreed that the panels are not in-plane with the structure. We have created these within the Rhino software on the same model of the bar centrelines that define the structure in Robot (we use our own tools that use the Robot API to bring the model into Robot including the cladding panels). Our models to date have worked with this method. All the panels are triangular, except on the flat facade where we have more control (The tool you describe sounds good though).

I first thought it was just due to using the P-D analysis, but my base model this time also has the problem.

If only cladding panels were defined by nodes... and not their own coordinates.

I am curious about the idea of using FE meshes 0mm thk instead of cladding panels, however I would have expected Robot linear analysis to not cope with zero stiffness etc, and even if it did then our modal and eigenvalue anlysis would have problems.

~R

[Rafacascudo]

can you send the model?

[raried]

Here is a picture of the plan view:

Roof is about 100m by 200m.

I don't think I would be allowed to send the model out. Perhaps to Robot Support Staff only. I will find out. However I am looking into why we have had problems with this one and not the previous ones. Will report back if I find out.

[Rafacascudo]

each little triangle is a cladding??

[raried]

yes. the highlighted ones relate to the error message in the dialogue box. There are also cladding panels on the 5m deep edges of the roof and the inclined facades below. For the time being we have not put cladding panels on the external soffit of the roof.

[raried]

here is another view:

[StefanoPasquini6790]

Send us your model, we try to investigate for a solution...

[Rafacascudo]

Just do a test and try to create the claddings inside Robot.

You can do it in 2 steps

Select all the top roof bars(those that form the triangles) and use  geometry/panel

It will create panels within each closed triangle. Convert them to cladding using this addin.

http://forums.autodesk.com/t5/robot-structural-analysis/api-macro-for-converting-panels-to-claddings...

by the way, Nice model!!!!