Grasshopper

algorithmic modeling for Rhino

Hi all,

I have 5 circles, which i want to loft. But grasshopper does not loft it correctly. It makes two separate lofts? Where is the mistake?

Thank you for your help.

Views: 1542

Attachments:

Replies to This Discussion

Too many cables for me, he he.

Let's start from the beginning (and I'll translate this into GH native stuff later on): do you want to create "rods" by rotating the curves (from split at intersection point) around the corresponding axis yes? And what would you like to do then? Maybe altering the shape of the curve segments in order to control more the "rod" shape? or what? (make some sketch).

Attachments:

Yes so many cables! Haha Didn't know how to make it with using less, as I never scripted before. Yes, I wanted to alter the shape a little bit, like 'Hyungsoo Kim' has drawn it. Thank you.

Well ... I used that C# because is 100 times faster (for me) to think that way. Anyway a "translation" is promised ... so no harm done, he he.

Other than that ... your issue is Not that simple (explanation is complex). In fact ... play with the options in the 2nd test provided and see the "twisted" results (if you flip the curves). If you want this working in all situations for any curve "shape" (non planar, that is) for any number of profiles (even different ones) ... well ... requires a few lines of code more in order to properly align the planes for placing the profiles (and that ... er ... hmm ... translation, he he).

Use the R file provided above.

Attachments:

Thanks a lot. Now I have another issue, maybe you know the answer :)

I commented it below. Thanks in advance.

... you know the answer...

Er ... hmm ...  maybe ... who knows? he he.

Introductory steps to heaven (actually to hell, but no pain no gain):

1. I would suggest (at first) to forget the balls the cones and other truss type of paraphernalia (for instance the roofing/curtain wall supporting system) and focus to truss matters. 

2. See attached: is this what are you trying to achieve? (ether "flat" or with W depth). C# is used because in order to do what this does with components ... well .. I wouldn't recommend it (but the real reason is not [yet] apparent - have some patience).

3. Then we can start exploiting the totally wrong way to put some "real" things (and what means this at baking time and response time) ... and the other way, he he.

Attachments:

Hi.

On the sidelines of Peter's advanced solution, regarding  your specific matter,

Branch structure should be collapsed before merging.

Hi, thank you! that's what I meant.

Now I have another issue.

I want to apply the system that I used to create that single module, on a grid. Like u see, I already used a lot of 'cables' to produce that single module. With that strategy it is impossible to do the whole grid. How can I do it smarter/faster?

Thank you.

attached files

Attachments:

Hmm ... you are trying to create a MERO type of truss? (either geodetic or with W "depth").

If yes ... and the N of nodes is big and you like the maximum efficiency ... this is only doable via code. Why? because instead of doing N*2 "cones", N balls and N rodes ... you should use instance definitions (blocks in plain English): ONE cone, ONE ball ... and unfortunately N rods (Rhino is not a feature driven CAD app, sorry). Complexity (and file size) increases "exponentially" if you want to mimic a real MERO system.

Recently a friend of mine send me (for inspection) a "big" canopy type of W MERO truss with 2300 nodes that was 500Mb (baked). After the "magic" treatment it become 1.2Mb (when baked).

Notify if you need such a C# based solution: (a) for solving any truss on any collection of surface Lists AND (b) putting "real" stuff (exact MERO members) on that  (but is a "bit" complex).

Yeah its kinda like the MERO type. But, the parts between the sphere have to be shaped like the single module shows. So, what would you suggest for the whole truss then? How do I find all the vectors between the center points of the sphere? How do I use these instance definition? (I am sorry, I'm new in the grasshopper field)

Well ... if the thing that you have in mind (the truss) is a toy (or a decorative/artistic thingy) and you have plans to make it with 3d printing ... then you can shape the rods as you like (but sill you should use Blocks if the nodes are many).

If however you are an Engineer (even a graduate/student) and you have some real-life AEC type of stuff in mind I would strongly suggest to follow these Germans (or the Italians if their current attempts come into a happy ending).

See this? It's a rather freaky situation: you get only points ... and you must deliver a W truss.

See that? It's what happens if you zoom in: an exact MERO type system consisting from the adjustment sleeves (in order to allow tightening that ... $%%$ bolt), cones, balls (not shown) and real tubes. The thing is that there's clash potential issues here - obviously - not every topology is doable (for obvious reasons) so the code does it for you. But the point is that we have 16560 "objects": in order to get the gist of my meaning ... make a cone like this (in Rhino) multiply it by 5000++ and then save the file ... what happens? 

The bad news are that the efficient way to do it (via blocks/instance definitions) is out of question with components no matter what your experience is.

So it's a turning point for you (and the fact that you are novice ... makes things worst). It's just one of these cases where half-measures yield full Armageddon.

I can provide you any solution imaginable on that matter ... just tell me what you finally decide (the green pill  or the red one? he he).

RSS

About

Translate

Search

Photos

  • Add Photos
  • View All

Videos

  • Add Videos
  • View All

© 2024   Created by Scott Davidson.   Powered by

Badges  |  Report an Issue  |  Terms of Service