Reducing complexity in surfaces - More Questions

[johndellostritto214]

With Form tool ( Smooth function)  and Cylindrify, Can a surface with mild inflections and bends be recognized by the computer as a non smooth surface - NOT  Part of a cylinder, cone and other known geometric or conic . Could such a surface(s) be flagged in the software and be candidate for automated smoothing per AI or other semi automated method? I'm just wondering what makes a surface with inflections/Protrusions unique and can data entities be identified without human visual judgement and mouse click selection? IF so, can surfaces be "smoothed" to match tangency and contour of consistent GD part geometry?

Please comment with ideas and thoughts.

Thanks,

John

[mango.freund]

hello, you set a 3-point offset plane on the surface and cut off a piece of the surface and then lift it again. You also have enough visual methods in your toolbox. greetings mango

[TrippyLighting · Accepted solution]

@mango.freund I have a hard time imagining how that would work on highly organic quad-mesh models created by generative design 😉

[mango.freund]

Hallo @TrippyLighting  --- peter natuerlich hast du auch da recht, aber der googeltranslator liefert nicht immer genaue informationen. so hab ich von der generativen oberflaeche anfangs nichts mitbekommen.

wie fit bist du in dieser umgebung?

der fragesteller hat aber einen denkanstoss geliefert, den ich durchaus nachvollziehen kann.

wenn das generative design fertig ist kann es dann nicht aufgebrochen und aus leichter modelierbaren volumenkoerpern wieder zusammengesetzt werden? ich faende das ziemlich gut zumal ich glaube, das nicht jede unregelmaessige oberflaeche sinnvoll ist. sieht zwar schick aus, aber.............notwaendig ???

lg mango

ps: machte absichtlich keine private nachricht daraus weil es ein ewiges seiten hin- und hergeschippere ist.

[johndellostritto214]

I remember a complex sin cosine equation that drew a side profile of a woman's face from mathbook circa 1955 20 years ago. I think if you have inflection and inversions then a flag to the software would be checking for a sine cosine curve fit  and a notice banner to smooth curves and surfaces . 2nd degree polynomial curves are already smooth transition non inflection  curves. so those structures made with 2nd degree polynoms do not need smoothing just adjusting size, flow and taper.

I am entertaining this thought experiment because the next generation of generative design software should smooth the surfaces without additional multistep processes and this should be a part of the Artificial Intelligence toolbox to aid the designer in optimizing his design.   

Welcome comments

[johndellostritto214]

Mango  Danke, My german at best is sketchy. I think You are saying that looking at the attributes of volume data may provide clues to answers on reducing surface complexity. It depends on how the surfaces have evolved from the volume. In CAD class (1990s)  was told that 2nd degree polynomials and some third degree create conic shells that are mathematically pure, tangent  and watertight. I think there is a lot of waste of resources tolerating unneeded complexity and keeping things simple works. Generative Design should lean to simplification if possible.

[mango.freund]

Hi, @johndellostritto214  When I look at the parts generated by generative design, they have a very high "natural" share, but this excludes many industrial companies from manufacturing processes that they could model with less complex "natural" shapes. i'm also thinking of mold makers in smaller industries. Who can buy these sinfully expensive technologies and bring them back profitably again and again? Taking the incentive of "natural" regenerative work and then bringing it into less highly technical work processes is the thought that you encourage me to do. greetings mango