Grasshopper

algorithmic modeling for Rhino

Hi All, 

My apologies for any incorrect use of terms i am new to the program and desperately trying to resolve an issue. 

I have attached my model and grasshopper file as well as three images. In short I am trying to apply the pattern in picture 1, to the surface in picture 2 and picture three is the issue I am having. 

The error on the TriGrid says " 1. Data conversion failed from Surface to Plane " 

Help would be greatly appreciated as I am still very new to things and feel bamboozled.

Thanks

Views: 1670

Attachments:

Replies to This Discussion

Hi.

"TryGrid" component can't  take your referenced surface(Untrimed or Polysurface one ,I guess) as "P" input.

 And you didn't intrenalise yours, so I just can recommend you would try as follows.

1. Untrim  your "Trimed" Surface.(You can do it in GH or in Rhino "Untrim" command)

2, "create UV Curves" in Rhino.

3. Generate Tri-Grid pattern on the UV Curve generated in Rhino by using your GH definition.

4. Apply tri-grid pattern to "Untrimed Surface" in GH using "Map to Surface" or "Apply UV Curves" in Rhino.

5. Finally trim again like previous trimed surface and clear up the applied tri-grid pattern.

Well David (Still waiting for the goats or some ZZTop decent stuff, mind)

I have bad, bad and bad news (as usual)

Bad news: if you want this type of stuff ... you must WAIT (Rhino is NOT a solid modeler thus cutting a zillion holes ... takes time: nothing serious for a CRAY mind).

Other bad news: you'll discovered them if you want the definition that does the above (I would strongly suggest to change plans, he he).

WARNING: If no goat(s)/ZZTop > no definition > adios Amigos > no mercy!.

no chance of getting this script ?? I have become very desperate... hahaha

OK, OK but this is the LAST time (until goats arrive - in good shape - in Greece).

BTW: Desperate you'll be AFTER exploiting the def, he he.

Attachments:

Thank you very much, I am just working it out now.

So I hook the surface into the X & Y? 

The surface goes where the ... er ... the surface goes, he he (spot the "internalized" comment).

This IS NOT the way to do it: perforating a zillion holes on a curvy thingy ... well it doesn't have any meaning (this would be some "shading" semi transparent layer outside the curtain wall, right?). So we are talking flat panels (poly-carbonate for instance) that are attached to the flat (planar quads) glazing primary support system 

The only real-life approach is to have, say, some variants of these "panels" already defined as blocks (instance definitions, that is) and then place them via the appropriate transformation in suitable planes (defined by the curtain wall panels). This is NOT doable without code mind ... but if the goats finally arrive MAY I post a demo def here (depending on the quality of the goats).

For instance this "panelization" yields 1600 non planar quads (there's ways to "attach" planar drilled stuff to them as discussed above): Imagine perforating all these by, say, different combinations using 100 "mini" triangles > call Samuel ASAP (and have the millions ready).

So the solution is to "attach" ONE (or a limited amount of random "variants") flat shading thingy as an instance definition: that is something that GH can do in real-time (well ... almost).

That is exactly what it is, it sits outside a structural membrane and acts as a shading element. 

I will do my best to keep working through it.

Thanks for making that file, it looks really good so far

No it's not: jokes apart ... is out of question (for good reasons: speed),  and is also out of question (for even better reasons: BAD engineering approach).

See this? It's a thingy (not for your level of expertise - yet - not to mention that it uses solely code) that divides a trimmed brep (with any holes inside):

But it also can yield planar quads (in 4 user defined ways):

Now ... assuming that this surface shown is actually the offseted surface that we need for placing the shading panels:

If these exterior loop polylines are offseted inwards (a bit) - in order to masquerade this stepped transition from the one to the next - they clearly outline the way to go: use planes out of them and place the flat drilled panel (always as instance definition) transformed accordingly.

this is the only way to do it efficiently and not waiting for an eternity to happen things that are BAD engineering thinking to the max.

Here's how to do it (ONE STEP before placing a myriad of instance definitions ... properly transformed). Instance definitions refer to these drilled shading FLAT things.

See this?

The dark thing is a collection of NON PLANAR quads out of the test surface.

The white thing is a NON PLANAR quad grid of an offseted mesh (out of the surface). Offseted because the shading thingies are at a certain distance OUTSIDE the curtain wall.

The pink things are polylines that have all their nodes co-planar > thus they serve as a template for placing the drilled panels.

I hear you: where's the WOW @%#$@% panels?

Well ... where's my goat(s)???.

I hear you: why we can't use this approach for making planar quad glazing panels? (for the main facade that is) Answers to: The Lord, District 9, North Pole.

more soon

Attachments:

so does this mean I can use this and apply any geometry to it for cut outs?

ps, you have saved my life here with these!

Yikes! (I'll never respond to any future thread of yours, he he).

But due to a prolonged good Samaritan crisis ... I'll prepare some stuff (maybe omitting the instance definition thingy - since this requires code and I predict Armageddon^N, unless we can have a LONG Skype session: do this, don't do that, no ... don't touch this ...etc etc).

Here's the salvation:

1. You define some panel with random triangles: This is our shading thingy. It hangs OUTSIDE the curtain wall by means of mysterious fasteners that are complex to explain right now (wait a minute: this is a master thesis or what?). We assume that these are cleaned by flying dwarfs (or smurfs) that do usually that kind of stuff. Stay in one variant for the moment (but if triangles are aplenty and they are created on this FLAT surface via the "same" approach used above [not exactly the same, but anyway]... well ...who's gonna notice it?). This is our seed:

2. Then we create PLANAR quads (polylines) all around our surfaces, as explained above, acting as the "guide frames"  and their corresponding Planes (actually: coordinate systems for transforming the "seed" panel into these "frames").

3. Then we do the transformation required and we get this:

Why I'm not posting the V3 that does this? Because is written solely with code and I must translate it (or some parts of it anyway) into native components.

Moral: this is another amoral case, 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