Zollinger timber structure A tennis roof Paris 1999 storm in karamba 3d

Hello,
I’m trying to find an alternative to the reconstruction of a Zollinger roof structure located on top of a Parisian building failed during the 1999 storm. I model the structure in karamba 3d with beaver plugin

ancient structure failed with the storm :

The new diamond structure is reinforced by glued laminated arches:

the Zollinger framework is relegated to a bracing role. I tried to tighten the mesh and increase the sections a little more but I don’t understand why there is a weak point at the intersection between the edge and the ridge beam :

I have some problem also to apply forces to nodes with beaver plugin

202604garage la motte picquet grenelle- ancient structure.gh (140.5 KB)

202604garage la motte picquet grenelle- other proposition.gh (148.4 KB)

202604garage la motte picquet grenelle- new structure.gh (137.3 KB)

Hi, as your question relates to beaver, it would be good to tag it with Beaver Latest Plug Ins/Beaver topics - McNeel Forum

It looks like the the Zollinger Structure is modeled as a diagrid. In a Zollinger structure, at each node one lamella is continuous, while the other two are connected by freely rotatable joints. Because of that, the system does not behave like a stiff diagrid and, without the cladding, it does not provide bracing by itself. The Zollinger connections themselves introduce other undesireable effects as well. (While beeing very elegant from a manufacturing and geometric stand point.)

I guess the weak point is due to the discontinuity and lack of a counterpart for the two semi arches.

You can find alot on the topic in the follwoing masters thesis, but it is in german: https://nbn-resolving.org/urn:nbn:de:bsz:l189-qucosa2-999150

The original structure is a simple 2D arch with diagrid joists between. That was efficient to build. To build without the arches will require shoring and complex joinery. Possible, but not as cost effective as the original structure.

What is most fascinating about these old structures is the craftsmanship. To build this at that time without all our modern equipment is impressive. To build now just takes $$ and fancy equipment.

Thanks for the link @cristoph.

This work reminded me of the following doctoral thesis, which might also be of interest to some of you:

Weitgespannte Lamellendächer der frühen Moderne - Konstruktionsgeschichte, Geometrie und Tragverhalten

title in english:
Wide-span louvre roofs of the early modern period - History of construction, geometry and structural behaviour

In that reference the image below is the closest to the shape of the OP. They are using a concrete structure to form a ring beam. Many others are a far more efficient arch shape.

This thread made me think of this thread with similar detail construction:

Digital Fabrication - How do I attain CNC precision with this scheme? - Grasshopper - McNeel Forum

At flex.htwk-leipzig.de we have done a lot of work on Zollinger-Structures over the years.:slight_smile:

(Image: FLEX@HTWK-Leipzig, MSZ)

This structure is a demonstration of the developments in the TimberPlanPlus Project, see the following link for in depth informations and a lot of images: document (although german again)

Hi Christoph, thanks for all this information. I have already tried to model a Zollinger node with one continuous beam and two beams with rotational joints, but I couldn’t find a way to do this in Karamba + Beaver. It seems that Karamba requires a node whenever there is an intersection, so I assigned a certain stiffness to the nodes. Is there a way to model this in Karamba?

:+1: Great work! Did they use Grasshopper for the geometry and RFEM for the structural analysis?

Hi @Henry_Boutet_de_Monv you need to work with element ids as well as the BeamJoint Agent to correctly define where you want to release the joint on a specific element.

Here is a simple script to explain the functionality and you can see that the My joint is only released when it connects to another element - at the end of the beam for example

apply joints.gh (27.8 KB)

Thank you! :slight_smile:

Yes, in that case it was Grasshopper with scripting and RFEM. But I also used Karamba quite a bit when working on Zollinger-like structures.

If you control the start and endpoints of the beams, you can use that information for the joint definition as well.

Thanks for your answers, I’ll give it a try.