Freeform surface panel edge length optimization

I'm new in GH and I'm trying to reduce the differentiation of panels of my freeform surface. I want to equalize length of panels edges to make the manufacturing cheaper.

I tried to obtain this effect using Kangaroo's spring, I achieved equalized panels, but the whole surface has been distorted. --->

Now I try to reduce the quantity of different panel shapes to minimum (could be 5 up to 20), whereas the original smooth form of my structure and her borders should remain fixed.

I found few methods in the following articles:

http://www.gasathj.com/tiki-read_article.php?articleId=31
http://sechel.de/publications/AAG2014Periodic.pdf

Unfortunately I can't obtain similar effect like in articles above,

can anybody make out with it?

Thank You for your help!

Attachments:

panel edge length optimization.gh, 23 KB

Replies to This Discussion

Permalink Reply by Damon on October 27, 2014 at 12:51pm

I'm sure there are those here who could get GH to do what you want. However, there is another tool you might check out: EvoluteTools for Rhino.

However, looking at your surface, I'm not sure you can get equal size panels if you want the divide the surface the way you have. Your surface converges at the top, so if you are using the same number of divisions at the bottom and top, you can't have the same size panel. How did you create the design surface? A lot of panelization problems can be improved or solved by changing the geometry rather than dealing just with the panelization.

One alternative would be to create a panelization that falls over the edge, then trim the panels along the edges.

Permalink Reply by Maciej Kukiełka on October 27, 2014 at 1:30pm

I agree with your suggestion that in my constrictive surface I can't use the same number of divisions at each size, when I want to achieve panel equalization. I did it at first because UV division of surface is the only one method that I know. So in conclusion I'm looking for another method of dividing, analogous like described in following article: http://www.gasathj.com/tiki-read_article.php?articleId=31, or http://sechel.de/publications/AAG2014Periodic.pdf but unfortunately I can't read out the GH script scheme because of low resolution. There are some directions how to do it but my GH crafts are too poor to reproduce described reasoning. In passing not each panel must be equal but the aim to achieve surface consiting of few different panels. Could be approximately from 5 up to 20 different panel shapes.

You can find the scheme of my surface in " panel edge length optimization.gh" file enclosed in the message above. This surface is a part of whole building shape which looks like this:

Thanks for your advice!

Permalink Reply by Damon on October 27, 2014 at 2:15pm

I didn't realize you are using ruled surfaces. Here's an idea that gives you equal sized panels. Granted, it changes the aesthetics.

What I did was took the bottom curve, copied it up the twisting form without scaling (I could have gone back to your copied triangles before you scaled). By creating a loft from this non-scaled form, I can panelize it knowing the triangles are equal. Then I trimmed it to the boundary of the form you want. All the internal panels are equal. All the edge panels are unique, although because the way you created form, the are mirrored.

172 equal panels in the middle, 39 pairs of unique panels along the edges. Not in the 5-20 range, yet, but another idea.

Permalink Reply by Maciej Kukiełka on October 27, 2014 at 3:12pm

It seems to be good idea but I have a problem with rebuilding your GH script. When I try to use my origin triangles sth doesn't work. What's more when my triangles aren't scaled, their edges are in different place than in scaled one. And one question- is that surface, which you made, rulled or is it planar?

I did sth like in file enclosed under but it doesn't work. Maybe I don't understand your idea in 100%.