Ivan Kiryakov

Male

Stara Zagora

Bulgaria

Profile Information:

Web or Blog
http://kiryakov.tumblr.com/

Comment Wall:

  • brendon_carlin

    hello, thanks ivan, ill post an image of the file.
  • Philipp

    http://grasshopper3d.ru/
  • Leah Bell

    Thanks for the tip, Ivan - I'm so clueless with Grasshopper though, I wouldn't even know what definition to use! In Rhino I would just contour it and extrude the result, but there's no contour command in Grasshopper :( Any tips?
  • digitaltoolbox

    Thank you very much for the compliment.
  • voronoiii

    Hey thanks for replying,.. I just downloaded the file,..But i really didnt understand how to use it,...do we have to copy paste in our defination and plug it or something,..??
  • voronoiii

    Hey Ivan,.. Could you please let me know how to use that Renderanimation.ghx file,.. it would really great if you can give me a small explaination about it, If possible give me some example file where you plugged in some defination into that renderanimation file,.. I will try to figure out how to plugin my file by looking at it,....
    I would really appreciate your time,..
    Thanks a lot,...
  • Steve Lewis

    Hey Ivan
    The work I did was not from using GH. Everything was done in Fortran on a 486 pc that had 256 MB of RAM way back when. I looked at linking packages together to pre/post-process data for the geometry of form-found surfaces. The novel part was linking with a program called Formian. Formian was a command line/script driven program that was able to generate complex forms and geometrical information. Similar to GC and Catia in that generative components were defined, and then instantiated on to any geometry.

    I'm hoping to try and replicate some of the surfaces in GH at some point. The algorithm I used was Dynamic Relaxation with Kinetric Damping. The great thing about DR and kinetic damping is that it is able to find numerical solutions where ordinary solvers would fail where the displacements, change in geometry during relaxation, are simply too high. This applies to traditional numerical solvers which use a "Stiffness matrix" approach. You can see from the images in my gallery that the surfaces are as far away as possible from a minimal surface, even so, DR is rigorous enough to find the form of these minimal surfaces as they deform massively. There are already a lot of minimal surface solvers out there in Rhino/GH which I am sure can do as good a job. It would be cool if someone could do a verification test on some of the surfaces produced to check the accuracy. Simple verifications would be a catenoid - increasing the height should cause the soap film (minimal surface) to become unstable and break at a mathematically known height. Another example is to use the box example - see my gallery. The location of the midpoint of the minimal surface should be exactly at mid height and the radii of curvature in the opposing directions should be exactly equal to one another. The only other way to prove a minimal surface exists is to do soap film tests. If you are interested at looking at benchmarks let me know and I can send you soem more info.