Need help about using Kangaroo for form finding

Replies to This Discussion

Permalink Reply by Victor Golovin on November 20, 2015 at 7:40am

 Hello

 You need to flatten force inputs and to define your anchor points.

Permalink Reply by Zhengyu Xue on November 20, 2015 at 8:10am

Dear Victor,

Ja, it works!!Thank you!

Sincerely,

Zhengyu

Permalink Reply by Zhengyu Xue on November 20, 2015 at 8:46am

Dear Victor,

There comes another question. 

My GeometryOut is empty..

Can you please check it?

Best,

Zhengyu 

Got it..

I should define curve as line.

Best.,

Attachments:

• 1st workable model.gh, 21 KB
• 1st workable model.3dm, 154 KB

Permalink Reply by peter fotiadis on November 21, 2015 at 4:54am

Well ... see some Kangaroo examples posted for you in the C# thread of yours (Tensegrity etc).

Get this attached as well (a simplified version of the Bend_a_... already posted: only 4 corners as anchors; if you need any combo use the Bend_a_3dgrid_V1A as it is). Since you have C# aspirations ... there's something there for you as well, he he.

NOTE: Divide et all AFTER Kangaroo NOT prior (pointless and CPU taxing for no reason).

NOTE: This is NOT the way to define a X Frame tensegrity truss (see images posted in the C# thread).

Attachments:

• China-error.gh, 28 KB

Permalink Reply by peter fotiadis on November 21, 2015 at 5:19am

BTW: Back to your initial goal of yours (tensegrity): added an ability to create a NURBS out for the "relaxed" points. Like this for example (4 anchors used and "negative" gravity):

Then a classic  XFrame tensegrity truss is made using this surface and a user defined division points (U/V) tree ... plus a variety of variables controlling W depth, X vertical position, strut "shape" and some other stuff.

Surface is required in order to use vectors that ... er ... define the vertical X Frames.

Attachments:

• China-error_V1.gh, 32 KB

Permalink Reply by Zhengyu Xue on November 23, 2015 at 2:57am

Dear Guardian of the Light of Justice,

I have read all your replies carefully and then started to follow your instructions and apply some components in my own design.

1)Yep,it takes some time. Here I have a question about C#. In the China-error V1.gh file, you made the four corner points as anchor points. What if I want to make all the points on edges as anchor points(since in my mind of the square tensegrity the edges are I-beams as compression hoops)

2)Secondly, in the discussion below, I tried to open the 3dpdf file but it displays an empty file(blank). If this does not work, can you please send it to my email: xt92phoenix@gmail.com?

http://www.grasshopper3d.com/forum/topics/i-want-to-make-this-form-...

3)In the reply you said we could arrange a skype discussion ASAP or I can go Sceppuku.@#$%&* You are really the most humorous man I have ever talked with. Indeed I really appreciate your generous help. But I do not know what to ask so far.. I will try to make some progress in this week. And is it possible to skype on weekends?

Best,

Zhengyu

Permalink Reply by peter fotiadis on November 23, 2015 at 5:05am

Dear King of Despair, Prince of Oblivion

1. The China_error is a "subset" or the K def already attached (re-attached here again) in the main thread of yours (Code subForum). This allows you to have any combo of anchor points (although using all the peripheral points ... well ... that's a good starting point for your adventures in C# > get a tree: say x: branches and y: items and gimme all the points in the perimeter, he he).

2. There's nothing wrong with the 3dPDF (but I'll mail to you anyway) > maybe you use some archaic Adobe Reader build? Have you configured the Reader properly? Have you ever attempted to exploit a 3dPDF thingy in the past?

3. Most humorous: you mean in the Monty Python way? (WHAT a bunch of genius was that "group" ...).

4. Skype weekends: well , since the MotoGP season is over why not? (ALAS: due to an enormous conspiracy [by Honda] > Hero-Valentino > got the short end of the stick > all that because he kicked someone [is this a big deal?] > why should I live ? > Seppuku ASAP).

Attachments:

• Bend_a_3dgrid_V1A.gh, 142 KB

Permalink Reply by Zhengyu Xue on November 23, 2015 at 5:50am

Dear Lord Seppuku,

Before reading your reply, I want to first attach some files. 

In the GH file I wrote some questions and some of my ideas.

Best,

Kind of Hope.

Attachments:

• 11-23 1st workable model.3dm, 135 KB
• 11-23 1st workable model.gh, 38 KB

Permalink Reply by peter fotiadis on November 23, 2015 at 6:07am

BTW: here's what happens if the skip this and skip that is set to 1

or if you want "airy" stuff (gone with the wind ... er ... I mean gravity/anti gravity)

best, Lord of SardineLand

Permalink Reply by peter fotiadis on November 23, 2015 at 1:13pm

OK, let's get serious:

Are you ready for the Dark Side? [you said so if memory serves well]

Are you a real man on a mission or some chicken?

This def attached (added the insane "all-vectors-point-to-a-point" useless option plus the surface as found in the china def).

Now the WOW stuff:

1. If you don't provide a surface the C# does a flat (restricted: read comments) grid and the rest are history.

2. If you  provide a valid surface the C# does a grid (restricted: read comments) by dividing the surface ... and the rest are also history.

Next update: Form finding on a sardine (no kidding): 100++ years ahead of anything known to mankind.

best, Lord of Darkness (the Merciless)

Attachments:

• Bend_a_3dgrid_V1B.gh, 164 KB

Permalink Reply by Zhengyu Xue on November 24, 2015 at 1:38am

Dear Lord Sardine,

1)  I understand a surface is necessary. But I also found that the surface generated by SrfGrid does not go through all the points but the command Delaunay Mesh does.  I've sent you an email with my work about the tensegrity so that you may have a clearer idea how it look like and how I made it. In my graduation work I may start with this shape and then modify and find a better one.

2)I deem "all-vectors-point-to-a-point" a possible way for form finding (form follow force).  Plus, I also plan to use an inverse iteration to do form finding(this part need C# knowledge)..

3) I am ready for the Dark Side, I am learning related knowledge about C, about GH although a bit slow in the beginning. But I'm not sure how deep I will dive into C#.. At least make a plugin in C# so that I can do parametric design in GH, define all the elements,properties,materials etc, press a button in GH then Femap do analysis and give me the result in GH..You can see the file I uploaded, only the calculation will be done in Femap. Also I want some comments from you about the optimisation procedure.. All the softwares I am learning now serves for the optimisation.. In the end I may create a tool for optimisation (how ? Maybe God knows. Maybe Lord Sardine knows at this time!). And final goal is to design a light-weight  tensegrity roof..

I like your GH files! 

Best

Zhengyu Prince of Caribbean.

Attachments:

• General idea about Optimisation.pdf, 114 KB

Permalink Reply by peter fotiadis on November 24, 2015 at 4:41am

Dear Prince of Something

1. First things first: FORGET form finding and all that stuff: where's some attempt for that %@#$ tensegrity thingy? And BTW what tensegrity type have you in mind?

2. Of course I could provide a turn key solution (only one black button > press > be a happy bunny) but we don't want that: because without pain + tears + agony + frustration + despair + Seppuku (optional) > life has no meaning. Is the walk and not the arrival that counts the most.

3. The Dark Side: well ... start "replacing" all the xxx_V1B native components (other than the Kangaroo related ones) with C# > goal should be to "absorb" everything into the left most C# (including Kangaroo stuff [Kangaroo2 to be exact] at some later time).

BTW: Light weight tensegrity roof means: heavy weight cost wise. Think about that (Modesty and not Vanity should guide us all into the unknown).

best, Lord of SardineLand

• ‹ Previous
• 1
• 2
• 3
• …
• 7
• Next ›
• Page