Simulation of Tensile Textiles and 3d printing


I was wondering if is there any chance to simulate(even unnacurately) the behaviour of a textile of any fabric like lycra to see how it folds after you 3D print over it.

I attached a youtube link for the behaviour of the textile I want to simuate. I am relatively new with kangaroo I read the manual and made several examples to make tensile structures but I dont know if there is any way to simulate that behaviour with kangaroo. The idea that I had was to simulate the fabric and see if can get one of these results before physically doing it.


Yes I think that this can be simulated with Kangaroo. But it is a free newtwork system. There are no anchor points.
I suggest you start with the circle example first.
The 3D printed boundary must have the same length both in 3D and 2D. The textile is just pretensioned. The final shape is an equilibrium one.

The only difficult thing I see is to do this without any anchor points.

Here’s a starter. It needs a nudge to start the buckling out of plane
For more accurate simulation of the boundary you might also get better results with the beam elements, since they also model torsion. (8.1 KB)

For more complex boundary shapes, you could model these as flat shells with bending resistance and then minimise the membrane. I used this approach back here (though using Nucleus, but same principles would apply with Kangaroo):


Thank you very much for the answer. I have a small problem with two components that are not working because I think thery are from a newer version I tried to install them but it did not work. I connected directly from the mesh to the springs and the rod and it seems to be working also. I used a different component for making diagonals from what I read gives shear stiffness.

Anyway thank you very much I will try to use the beams to model the torsion on the boundary10 .

Which version of rhino are you using?

Rhino 6

If you update to R6 Service Release 16 these unrecognized object errors should go away.

As you found - it will work without the remeshing step, but I added that because if the the quads are all diagonalized in the same direction and treated as springs, it can introduce a slight directional bias to the fabric tension, which might not be what you want.


Thank you very much for the answer

1 Like

Thank you for the information I will check that also

1 Like

Hello,I dont know if I got it correctly but the direction of the diagonals could introduce a bias in the behavior of the textile? But what should be the correct way to create the mesh for different shapes(like a star shape, rectangle)

Also, I was trying to understand the influence of the mesh subdvision in the behaviour of the fabric under tension. When I change the subdivisions the behaviour change, but then if I modify the lenght factor of the springs, I can get the same result(more or less) than before but with a different number of mesh subdivisions.

Hi Daniel@DanielPiker , thank you for the starter. This is very similar to my current research. May I ask you for some hints on the more complex pattern simulation?

#A. The pattern to simulate

#B. The crease I feed to the rod component
#B.1#exterior pattern

#B.2#internal pattern (/per zone, one color stands for one crease)

#C. Reason of my choice
From my understanding, the core of the simulation is the angle constraint. I assume the internal pattern may affect the exterior pattern as well. One proof is that I feed the crease as following.
C.1#internal pattern

C.2#exterior pattern

Then there is no constraint between interior and exterior. It will be like this:

#. Question want to ask
Since my strategy is to make sure the crease is in one system, I also make sure this is the same as my 3d printing path.
I was wondering is there a better way to make the crease affect the system as a “whole”? Or do you think there is something I can implement? My simulation result is the following and you might notice the edge is not curvy as the 3d printing file(which is 3d scanned after printing).

Thank you in advance! (31.8 KB)

1 Like

Could you explain the role of the grab component and if there’s a replacement for it? I’m trying to use the Zombie Solver although it seems that it’s incompatible with the grab component