surface controlled by outline

Hi,

I would like to control surface deformation by its outline.

Physical model represent surface which i would like to achieve.

Dividing outline curve into multiple 4curve surface is not a good idea...

Making multiplay sections and lofting them does not work well...

If you have any idea how to approach it is problem please give me a hint

best

Attachments:

14_08_17_powierchnia.gh, 8 KB
14_08_18_surface.3dm, 187 KB

Replies to This Discussion

Permalink Reply by SA on August 18, 2014 at 5:50am

Just a clarification, patch could help but what is it you really wanna achieve, solution depends on that.

Attachments:

1.3dm, 459 KB
1.gh, 16 KB

Permalink Reply by Karol Wawrzyniak on August 18, 2014 at 6:28am

Thank you for fast replay. It looks like something worth testing.

best

Permalink Reply by ng5 Alex on August 18, 2014 at 6:59am

You can also try kangaroo, if you don't mind working with meshes.

also check this comment

http://www.grasshopper3d.com/forum/topics/kangaroo-matters-relaxing...

to see a definition for similar problem but with surface as output

cheers

alex

Attachments:

crv_deform.gh, 13 KB

Permalink Reply by Karol Wawrzyniak on August 19, 2014 at 12:08pm

I was playing with patch.

It is working ok so far although surface outline is not exactly passing the initial curve

Is there a method to achieve similar results not with one closed curve but with a hole in the surface? like example on the right.

solid difference enables desired results but ... it is not nice.

Permalink Reply by ng5 Alex on August 19, 2014 at 12:16pm

please internalize the curves or upload the rhino file too

cheers

alex

Permalink Reply by Karol Wawrzyniak on August 19, 2014 at 12:23pm

Sorry, i am attachning the file with internalize curves.

best

Attachments:

3.gh, 10 KB

Permalink Reply by ng5 Alex on August 19, 2014 at 12:59pm

By not nice you mean that the surfaces created from patch do not have as boundary the initial curves right? since they are not planar patch does the best to keep it close. try giving a greater value in spans and it gets closer

i believe what you did is the best you can do, since lofting the two curves results in a very different surface.

if that is not enough for your case then i think you should define you first surface, not by just only one curve, and then the cutting curve.

cheers

alex

Permalink Reply by Karol Wawrzyniak on August 19, 2014 at 1:28pm

Alex,

"patch do not have as boundary the initial curves" This is true but for me at this point it is acceptable, because it is just a early concept sketch.

Using boolean operations is not "nice" in the sens that adding extrusion nad solid diference makes definintion hard to work with- i have experance it many times. so i try to avoid boolean. but now i dont have any idea how to do it without it. I am looking at your proposition with meshes but solving it without kangaroo would be better form me.

thank you for your help

Permalink Reply by ng5 Alex on August 19, 2014 at 1:40pm

Ok then

since it is a speed issue and you would like to avoid the booleans you can try the following

first you can avoid the second patch component, which is heavy too, and project the curve to xy plane since you already are extruding in z direction.

then instead of difference component you can use trim solid component which in my PC from 8 seconds of the difference component goes down to 400ms.

cheers

alex

Attachments:

4.gh, 11 KB

Permalink Reply by Karol Wawrzyniak on August 19, 2014 at 1:49pm

defenetly more elegant than my slution:)

Thank you for your imput.

best

Permalink Reply by SA on August 18, 2014 at 7:06am

Hey Alex,

Cud you tell why my definition is not working, I was working on the similar lines with a slight diff approach. The connection fail if I connect the lines formed by flip matrix but i need a full grid. Any suggestion.

Best,

Aakriti

Attachments: