minimal surfaces from curve network??

Hi Jon,

I'm working on a tensile net installation that will hang down from a set of handrails in a public space at my university; I'm using kangaroo for the physics of it and geogym for the curve networks and surface generators. I was wondering if there is a possible way to make these surfaces (see image below) minimal surfaces (as if each triangle were being stretched from each corner).

So far I've been able to adapt Daniel Piker's wooly path script as a post-process to simulate the minimization of the surfaces. It works just fine, but I'd like to make the process a bit less painful by not having to set up different timers (one for kangaroo and one for wooly paths) and more integrated (minimizing the surfaces as they are generated from the tensioned net with the sdrCNF component). Please refer to this link (and the image below it) to see what I intend to do with the minimal surfaces:

https://www.youtube.com/watch?feature=player_detailpage&v=cH10w...

Any help will be greatly appreciated, let me know if you need the gh files!

thanks

Replies to This Discussion

Permalink Reply by RENE on November 2, 2012 at 4:27pm

this is a half-scale physical study

Permalink Reply by Jon Mirtschin on November 3, 2012 at 12:54am

Hi Rene,

Thanks for the post, and if I've understood correctly, I've added an improvement for the next build of StructDrawRhino (I need to implement another improvement, but should be shortly or later today) that will help.

I think the issue to date has been the mesh was always restrained at the perimeter vertex. I've added an option to the parameters for leaving them loose, so by nominating an edge strength you can control the extent of surface retraction.

As a comment, I don't believe many membranes are minimal surfaces (although this generally gives the correct appearance). The actual surface area is dependent on unstressed pattern and the forces that then act on it.

Hope it helps, simple test case attached.

Cheers,

Jon

Attachments:

121103 mesh relaxation.gh, 6 KB

Permalink Reply by RENE on November 3, 2012 at 1:21am

Thank you Jon, I think this is it; however when I open the .gh file I get this message:

does it mean I must download your plug-in again for this improvement?

Permalink Reply by Jon Mirtschin on November 3, 2012 at 1:50am

Yes you need to update. I will try and upload ASAP.

Permalink Reply by RENE on November 3, 2012 at 1:51am

thank you, I'll pay attention to that

Permalink Reply by RENE on November 3, 2012 at 9:01pm

it worked perfectly, thanks again Jon!

Permalink Reply by Jon Mirtschin on November 4, 2012 at 1:55am

Looks great!

Permalink Reply by Daniel Piker on November 2, 2012 at 5:29pm

Hi Rene - It's not clear to me from that video how you are using the wooly paths definition (which one by the way ?).

You could maybe simplify things a bit and get rid of the multiple timers by doing the surface relaxation in Kangaroo if you are already generating the curves in there.

But I'm not really following your process, perhaps you can describe it in clear stages ?

Permalink Reply by RENE on November 2, 2012 at 7:46pm

Hi Daniel,

Thank you for replying! I apologize for the lack of clarity. I should have explained the sequence a bit more as the video was used for explaining the project during a review and not so much for this purpose. So, without paying attention to the video, this time, this is the synthesis of the process:

1. generate 3d network from points and applying tensegrity physics via kangaroo:

2. option 1: generate surfaces from network using network fillet/patch tools from GeometryGym:

3. option 2: same as option 1, but this time the network is affected by the wooly paths script in order to create the 'tensioned' curves via attraction and then creating the surfaces with GeometryGym:

4. close-up:

The goal is to achieve "option 2" without using wooly paths so that the minimal surfaces are generated as the curve network is affected by kangaroo-driven tensegrity. I want the model to be a more 'real-time' simulation instead of generating the tensioned network and doing the surfaces as a post-process. I hope that explains my intention a bit more!

the wooly paths script I'm using is the "woolier paths" one or #2, I believe.

Thanks again! I hope you can help me, and sorry for any confusions!

Permalink Reply by RENE on November 2, 2012 at 10:50pm

here's an image of the thing viewed from above, still using the wooly paths to deform the curve network and make the surfaces look like minimal ones; I hope everything looks clear!

Permalink Reply by shehzad irani on November 2, 2012 at 11:39pm

Why would you not use "rhino membrane" for developing the form? Instead of going through all this trouble?

just curious thats all.

Permalink Reply by RENE on November 3, 2012 at 12:09am

Thanks for the suggestion, I will look into that soon! However isn't Rhino membrane a plug-in for Rhino only? I am interested in the real-time simulation of the network as it's tensioned and hangs down.