Kangaroo Testing — Material form

Montanna Green
May 7, 2017 · 4 min read

Some students went into university on Thursday and Friday to contribute to the pavilion design. Nissie and Dominic worked on the computational design and I explored kangaroo and the material form which would be attached to the design. Both Emily and I researched the effect of gravity on the material if it was draped over the structure and how it would push through. I explored the structure that Nariddh created.

There is a section where the fabric has nothing to hang onto and therefore drops quite low. This is shown in the image on the left side of the dome. This problem is due to the angle of the divisions. If this division was selected I’m sure this issue could be solved.

From this, I explored further possibilities of the material and generated a script that represents a stretchy material and the effect gravity would have if we used a pin joint to attach the material to every intersection of the beams. I was able to create a number of interesting solutions before reaching my goal.

These interesting solutions where generated through minor errors in the script, most of which I have no idea how they happened. The below images are what results I received when building the kangaroo script.

To reach my goal of having fabric looking like it is being draped from the intersecting beams, I needed to split the surface, with the inputs being the surface I’m splitting and the curves that need to split the surface.

If I had not split the surface based on the curves, the fabric would not drape correctly. As shown in the image, the mesh is now split into each panel.

The list of both edges (naked and interior) were then split , and the second list (B) was used the create the springs.

The interior split list was also used to generate which lines would be effected by the force. The amount it was effected was also determined by the strength of the material (Uforce component — input force).

From here, we needed to find the intersecting points that would act as the anchors, this was done with the ‘multiple curves’ component. This was then feed into the closest mesh point component, which was then feed into a mesh evaluation. This then outputted the required points for the anchors based off the mesh.

Everything was then feed into the kangaroo simulation and the setting component and a sequence was added to not over run the computer.

Below is the result.

I actually really like this solution, and a few colleagues also said they like this solution. It’s very interesting and quite possible. After generating it, I ask Ali to put it into UNREAL to see what I would be like in space.

His feedback that the opening was too small, and the interior is a bit confined. This is all stuff that can be easily fixed. The way the surface exported was quite rough, so this issue is something that needs to be solved.

Once the pavilion was finalized, I applied my script to the design.

There is a few glitched in the script that will need to be fixed and I’m not 100% whether it is accurate, but this gives the tutors and idea of what we are aiming for.


CoDe 3100 Digital Collaboration Studio