Exoskeleton on 3D loft surface

Hi everyone,

Thank you for the attention, here is my question:

As seen in the first picture, a Hex grid was used to initiate the pattern. (please see attachment for the .gh for a clearer idea of the script) So far so good...

However, problem comes when I try to apply the some logic to a loft surface. You can see the Exo-skeleton command is not reading the three curves as "a set of three" therefore not doing the boolean to the three pieces of geometry as seen below in the second image. I think it has something to do with the data converting as I can see the data tree is acting a little bit different between the two panelling command (2D Hex Grid VS LunchBox Hex panel).

I am still trying to learn how to use grasshopper properly, please go easy on me =P

Huge thanks in advance!!!

All the best to you all

Attachments:

woven with gird initiation.gh, 29 KB

Replies to This Discussion

Permalink Reply by martyn hogg on October 23, 2015 at 10:23am

your GH file does not include the rhino geometry that it references. If you right click any component that references rhino geometry and select "Internalise Data" then it will be included inthe GH document.

The problem is possibly to do with the data structure going into the Exoskeleton component. If you want lots of individual meshes, each with 3 lines driving them, then the input will probably be lots of branches each with 3 items and hovering your mouse over the Exo input should show something like...

{0} n=3

{1} n=3

...

etc

Anyway, rather than speculating, if you internalise your data and repost then someone can take a look at it!

Permalink Reply by Vincent yeung on October 24, 2015 at 1:08pm

First of all, Thanks for all the teaching and sharing Martyn.

As I have improved the .gh file, I now have a new question.

As seen in the script, the tessellation can now happen on any loft surfaces no problem. However, as it is not on a flat XY plane any more, moving the centre point of each hexagons to regenerate the new form of the geometries will eventually cause distortion (as the points travel on the Z direction but the geo could be on a curving side of the surface).

My thinking is to replace the Z vector with an offset of the loft surface, that should be able to make the centre points travel with the offset? I tried something but not sure why it didn't work.

Please help me out when you are free.

As always, big thanks in advance.

Attachments:

wovenwithgirdinitiation improved.gh, 27 KB

Permalink Reply by martyn hogg on October 24, 2015 at 3:44pm

I did that I think...

Use Surface Closest Point and evaluate the surface at that point which gives you the surface normals. Use these to move the points instead of z vector.

Attachments:

wovenwithgirdinitiation.gh, 34 KB

Permalink Reply by Vincent yeung on October 23, 2015 at 10:46am

Thanks for the advice Martyn ! I have internalised the curves now.

Actually, speaking of that loft surface, I do have another question about it.

As I uses LunchBox panelling, the pattern is always set inside a boundary (which is my initial surface). Is there anyway to cull out the naked edges of my surface so I can get full Hex around the edges as well just like what the 2D Hex command will give me?

Sorry for spamming loads of elementary questions.

Thanks again

Permalink Reply by martyn hogg on October 23, 2015 at 11:22am

First up... there is some problems with the lines going to the Exo...

I've flattened that input and then run it through topologiser which cleans up the lines in terms of them being a good input for Exo (There's more info on Topologiser on this forum somewhere!)

I've reduced the input rads for the struts in Exo. If you increase them or the node depth then the nodes can engulf the struts and you get an error.

I also changed the Z move of the centre points to be in the direction of the normal of the surface at that point. I think this helps avoid short struts that can cause problems.

I'll have a look at the part hexagons at the edge of the surface...

Attachments: