Kangaroo Form Finding & Planarization

Hello!

I have used kangaroo 0099 for form finding a semi-dome composed of voronoi patterns. The components I used were: Springs, Pressure, UForce and CurvePull. I was wondering if I can do the same with Kangaroo 2? I read the release notes but i couldn't figure out a way of using components in Kangaroo 2 to do such a thing.

Another thing I was trying was to planarize the form I have generated from kangaroo using the PlanarHex.gh example from the Kangaroo page. It seems to work out in terms of planarizing but the cells are too greatly deformed. What parameters are good to adjust if I wanted to create a smooth dome. Does the number of cells in the form finding process affect this?

Thanks very much in advance!

Attachments:

Semi Dome 3.3dm, 80 KB
Semi Dome 3.gh, 39 KB

Replies to This Discussion

Permalink Reply by William Carroll on May 27, 2015 at 10:29am

Hi Andy:

I am not sure if this will help too much, but I think I can comment on the PlanarHex dilemma. Planarized shells are a topic that I have been recently interested in, ever since Achim Menges lectured at my university; one of his projects is a free-form shell pavilion, based off of the finger-locking joinery of sand dollars. He has a great time-lapse video of the form finding process on his website, which is worth a watch if you haven't seen it already. It appears as if he uses Galapagos or some other evolutionary approach for his solution. When I saw that Daniel Piker replicated this using only Kangaroo, I was thrilled.

I believe that in order to get smoother, less clumsy results from the PlanarHex, it is almost necessary to ensure that your setup is conducive for the definition to work. I think this includes beginning the definition with the hexagons having the ability to be EITHER concave or convex (Lunchbox). In Daniel's script, I think he set the default value at .5, which was in the middle of the two options and resembled a square. I think also that the number of hexagon cells in the U and V direction should be setup in a way that they can lend themselves to a proportion that is used later in the definition.

Here is a quote from Piker that better explains the importance of the proportions and edge lengths in the setup:

The ratio of U/V divisions need to be chosen so that the initial edge lengths are close to equal.

When planarizing, particularly with hexagons, we need to ensure that the polygons do not become self intersecting. Here this is achieved by constraining the maximum and minimum edge lengths to some ratio of the overall average edge length. For this to work though, it helps if the initial state has them within or close to this range. [...]

Judging from the photo you provided of your results from the PlanarHex, I think some of the issues may be a result of you having a mixture of trapeziums and hexagons. This could potentially be causing the mangled appearance of the resultant shell.

Again, I do not know if this will be too much help for you as I am somewhat of a beginner with Kangaroo. But if nothing else, I hope that this can serve as a catalyst for what could be an interesting discussion about free-form vaults and shell structures... Also, check out the work of Philippe Block if you haven't. Here's a picture of a hexagonal vault model that he designed that is quite lovely.