Grasshopper

algorithmic modeling for Rhino

Hello everyone! 
For my thesis I'm studying glass facades supported by tensile structures, so I'm trying to replicate the behavior of cables in a vertical facade in Grasshopper, in particular with Karamba. 
By now I've done a beam component for the cables and a mesh that simulates the glass.
The forces I want in my cables are the pretension, the weight of glass applied to attacks and the wind as a point load on the attacks. On the glass I want to simulate the wind as a mesh load. So my question is: Karamba is the best plug-in to do that structural analysis? If yes, can I transform the "mesh load" of the wind in "point load" just for the cables and maintain both?  

Views: 2153

Attachments:

Replies to This Discussion

Hello Sofia,

I think you forgot to flatten the input of the 'Assemble'-component which results in multiple models.

When you give truss elements an initial tensile prestress and calculate them using 'Analyze Th II' they have a positive transverse stiffness and can be used instead of beams.

You could apply the wind load via a MeshLoad-component on the glass and model the connection between glass and cables via short springs or beams. In order to avoid singularities at these point-wise supports you could embed short beam elements in the glass which roughly model the connection detail. The mesh you currently use for the glass is too corse. Try the 'Mesh Breps'-component of Karamba for integrating the connection points in the mesh.

Best,

Clemens

Hello Clemens, thank you so much for the replay!
I'm trying to modify the algorithm separating the cable-beams from the glass-mesh and creating connections-beams between them. I've found the Joint-component to connect both but it doesn't work, I can't understand the mistake. 
Also the error in the Assemble-component is obscure to me! 

Attachments:

Hello Sofia,

you did not flatten the input of the Assemble component, which resulted in multiple models. There were some other minor issues (see annotation of groups in attached definition).

The cables lie in the same plane as the glass. I think they should be offset to the end of the connection elements and converted to trusses. The hinges of the connection elements should be a the cables so that they form small cantilevers which are fixed in the glass.

Best,

Clemens

Attachments:

Hello Clemens! I modified the file that you have created and I think now it works, so thank you!
But now I have to do the same with more complex shapes, like cable trusses.
In this case I can't understand the strange distribution of the glass deformation (and why the external cable trusses have a rotation towards the outside), and why some supports seem not to work. 
And how can I verify that the orizontal elements are struts? because in the axial stress there is no negative value.

thanks,
Sofia

Attachments:

Hello Sofia,

there were duplicate points when you assembled the structure. This partly disconnected the substructure form the glass plane.
In order to get a realistic response for the glass plane the mesh needs to be fine enough (see attached definition).

Best,

Clemens

Attachments:

Hi Clemens,

I have a doubt about loads: till now I didn't use Gravity-load, but it's necessary to simulate the weight of the structure? Because when I try to use it there are strange deformations, on cables expecially! 

And a second question: the maximum displacement that I read in output of Analyze, what is referred to? Is the maximum deformation of the element that is maximally deformed in the whole model? 

thanks,
Sofia

Hi Sofia,

when you use beam elements for the cables and do not prestress them they show large deflections due to their small bending stgiffness. Is this what you mean by 'strange deformations'`? You can remedy this by switching off bending and turning them into truss-elements which do not bend.

The maximum displacement is the largest absolute value of all deformation vectors of points in a model (see manual for details).

Best,

Clemens

Hi Clemens, 

when I try to switch "false" the bending toggle, the Analyze-component says me that structural system buckles in 75 modes. And if I stay "true" there are the deformation because of the gravity, as you can see in the attachment "Cable truss". (And why do I have only 3 colors for the beam view and the shell view?) 

In the meanwhile I'm trying to develop a curve glass facade with cables (Curve Facade), but the deformations are isolated only in a part of the glass, even if the mesh load is uniform, and it happens in any casual curve I set, I can't understand why.  

Kind regards,
Sofia

Attachments:

Hi Sofia,

try to make the model as small as possible for testing. Consider e.g. only one vertical cable truss.

With the 'Buckling Modes'-component one can see, that the cable trusses buckle sideways. In case the connectors are modeled as trusses there is a full hinge where they cross the first tendon. This makes them buckle.

The problem with the coloring is due to the fact that in the ModelView-component the result thresholds in submenu 'Render Settings' are set very tight.

Try to solve the issue with the second model by reducing its size. Apply the divide and conquer strategy for finding the problem.

Best,

Clemens

Attachments:

Hi Clemens, 

Thank you for the help with calbe truss, it was also a problem of visualization of the model, in which I'm not so able as you can see! 

Anyway in the curve model I've tried to change the base-curve, and when I create a symmetrical facade with flat and concave parts, the flat ones deform normally but the concave ones have a really small deformation in comparison. I would like to know if it's the algorithm wrong (but is the same of the other models), or it represent the reality considering the influence of the form in relation with loads. 

Best,
Sofia

Attachments:

Hi Sofia,

the reason for the small deflections is the geometry of the glass plane in combination with the support and loading conditions you chose.

Make sure that the support conditions are realistic; also include a wind load case with wind only on one half of the curved part.

Your mesh is rather course. You should refine it.

Best,

Clemens

Hi Clemens, 

you were right because it was a problem of combination of loads and geometry, thank you.

I have a question about how Karamba treats the element rope/cable: using the "Beam Resultant Forces" component I see that in my model with wind load, glass weight as point load, and pretension, the wind load doesn't take part at the Normal Force. 
I have to do a comparison with another software and I did it with Strand7, that gives me back a different result, so I was thinking there was a base problem with linear/not linear analysis or anyway with the rope/cable special case.

Best,
Sofia

Attachments:

RSS

About

Translate

Search

Photos

  • Add Photos
  • View All

Videos

  • Add Videos
  • View All

© 2024   Created by Scott Davidson.   Powered by

Badges  |  Report an Issue  |  Terms of Service