Pratt Computer Aided Construction Summer 2014

This discussion is for Grasshopper-related file sharing, questions, and general tomfoolery in CAC Su14.

Replies to This Discussion

Permalink Reply by Ethan Gross on July 23, 2014 at 6:33pm

Hi Brian,

I was playing with your lofted voronoi definition by stopping short of your lofting steps and (grouped in blue) extruding the filleted perimeters then capping each cell. I did the same to the scaled down cells and did an sdiff, getting the 'doughnuts' I hoped. I wanted to smooth them out so I converted to a mesh and then tried wb Catmull-Clark. What I ended up with are 'puckered' doughnuts, which are interesting but not what I expected. Do you have any idea what I am doing wrong?

Thanks,

Ethan

Attachments:

puckered voronoi.gh, 20 KB

Permalink Reply by Brian Ringley on July 24, 2014 at 5:32pm

Hi Ethan

You should avoid "solids" operations in Grasshopper whenever possible - they're super slow. I've highlighted (see attached def) how you could have modeled the same closed polysurfaces using lofts.

Smoothing ops like those found in Wb work best on polys with low face counts. If you hit Ctrl+M in Grasshopper you'll see that you've meshed your breps at too high a resolution for the smoothing algorithm to work properly, resulting in the puckered geometry.

So there are two workflows I would recommend - using the breps you have and using surface filleting to smooth the geometry, or going back to the 1 degree polyline voronoi cells and extruding low poly forms for eventual meshing (if you keep the NURBS faces planar you can make it so that 1 NURBS face = 1 mesh face) in Wb. Check out this Mode Lab Education Course for more info.

Attachments:

puckeredvoronoiBR.gh, 26 KB

Permalink Reply by Ethan Gross on July 24, 2014 at 6:17pm

Thank you Brian for your easily digestible solution! It is certainly a lot smoother than my attempt, though it will be hard to ween myself off of solids. I thought of filleting in Rhino also, but that would have been 'impure'. I appreciate the Modelab link too.

Ethan

Permalink Reply by Brian Ringley on July 24, 2014 at 8:01pm

Some of you have asked for an example of curve attraction used to drive unit variation rather than distance-based point attraction. This allows you to create unit parameter intensity along a seam rather than at a locus.

Please see the attached Gh def. In it I use curve attraction via the Nudibranch add-in (download here) to demonstrate variability in unit aperture as well as variability in unit size and location using a simple Voronoi base grid.

Attachments:

curveAttractor.gh, 32 KB

Permalink Reply by Brian Ringley on July 24, 2014 at 8:45pm

Some of you have been interested in the idea of overlapping. This can be done by creating a series of twisted boxes that go from the subsurfaces of the base surface to the subsurfaces of the offset surface and then shifting their respective lists to create the overlap. I've attached a quick sample (using Olivia's units).

Attachments:

overlappingScales.3dm, 2.2 MB
overlappingScales.gh, 13 KB

Permalink Reply by Qin Shasha on July 25, 2014 at 7:54am

Hi Brain,

Could you help me change the grid height? I try to connect the distance which doesn't work.

Attachments:

7_25_Height.gh, 62 KB

Permalink Reply by Brian Ringley on August 1, 2014 at 12:12pm

Hi Shasha,

Prior to connecting the Distance data you will need to Remap the data to a new range. This range would be from the shortest desired height to the tallest desired height in your system (e.g. a range in inches if defining a height for a move or extrusion, or a percentage range if you're scaling existing geometry).

The length of the Remapped distance list should match the length of your unit list. You will likely need to graft both inputs into whichever transformation you use to modify height so that each transformation value is applied to each geometry unit individually, as opposed to something like moving 50 units 50 times each for a resulting 2500 pieces of geometry.