dfmdmx

Tejido Parametrico Para Nidos

2022-12-01T00:00:00+00:00

Exploración formal de nido para pájaros a través de la deformación espacial de un filamento orgánico utilizando algoritmos genéticos y vectores biológicos en el proyecto de Av. Carrilet por h3o

Referencia h3o

Proyecto arquitectónico por h3o

Posibles tablas de información para alimentar el algoritmo

Cada especie junto con su ubicación tendrá solucions distintas

Ejemplo de variación extrema entre pajaros y nidos

Concepto de posible evolución formal del algoritmo

La idea es generar un filamento para extrusión en 3D a base de resinas orgánicas y agregados terrosos que vaya adecuando sus propiedades según el requerimiento de la pieza a imprimir, en este caso el nido. Este filamento digitalmente puede servir como el único elemento a deformar por el algoritmo para dar distintas propiedades emergentes al nido como la temperatura, permeabilidad o dimensiones, entre otras.

fig.0 nidos

fig.1

fig.2

fig.3

fig.4

fig.5

fig.6

fig.7

Filamentos extrusión 3D

Proyecto arquitectónico por h3o

Referencias de projectos academicos relacionados

Design principles of biologically fabricated avian nests

https://www.nature.com/articles/s41598-019-41245-7

fig.x

A modeling algorithm for exploring the architecture and construction of bird nests

https://www.nature.com/articles/s41598-019-51478-1

fig.x

Referencias nidos impresos 3D

fig.x

Station Earth Sketch

2020-09-03T00:00:00+00:00

Sketches 1

The future is already here — it’s just not very evenly distributed.
William Gibson

Lost And Found Stranded Beest Simulation

2020-08-03T00:00:00+00:00

Simulation of a Stranded Beest by Theo Jansen made in Rhino3D with Grasshopper and Kangaroo physics

I made this exercise a while ago when I was learning Kangaroo but all I could found was the video. I think it would be amazing to parametrice some values of the beast geometry and take them into account in a genetic algorithm. The organism will try to reach its most optimal form to travel the furthest with the less amount of energy in a fixed amount of time. The genes that come to my mind are the length of the rods and the intersection angles between them but I don’t a have clear strategy at the moment.

I’m goin to try and recreate this simulation using Kangaroo 2.0. Judging by the date, 2014, I probably used version 1.0 anyway. Lately I have been using Flexhopper instead of Kangaroo but I have not been able to combine it with any GA problem solving Grasshopper plugin. I think the zombie solver in Kangaroo pauses in some way that it resets the simulation by itself.

Stranded Beest Webpage

Stranded Beest

Tessellation script in Python

Screen captures from videos published in the Stranded Beest Webpage

Stranded Beest 2D simulation in Kangaroo

Software

For the modeling I plan to use Rhino3D and Grasshopper, same as before. Octopus for the GA solver instead of Galapagos, and Kangaroo2.0 for the physics engine.

Octopus outputs some good looking graphics.

Octopus genetic algorithm solver graphic example

Downloads:

TODO

Recreate 2D rigid model with Kangaroo2
Define parameters to use as genes

Rhino Python Branch Growth Visualization

2020-01-02T00:00:00+00:00

Growth 2D pattern driven by random direction choices and a simple collision detector that includes other branches and a perimeter.

The tree shown here represents one population member in gene pool that will try to set the correct amount of light needed in a room. Each black dot will be drilled on a composite aluminum panel through which light will travel.

The programing was made using Rhino and Python.

The frames were taken with this Python script.

The video was merged using Resolve

Music by Verified Picasso.

Rhino Python Path Tracing Animation And Script

2020-01-01T00:00:00+00:00

Frame by frame screenshot capture for Rhino3D Python script

This is a concept test for creating a frame by frame animation using Rhino and Python. It’s definitely not the fastest way to create an animation but it may be a useful tool to visualize evolving data through time and have a better understanding of some optimization algorithms strategies within a script.

I was pleased with the time it took to create an individual frame. I will try to further implement it on a rigid body optimization software using a genetic algorithm that I have been working for a while now.

Here are the videos I made to test the concept as well as the Python script.

Videos

The videos were merged using Resolve.

Python script

Rhino Python standard imports.

import rhinoscriptsyntax as rs
import os

The Timelapse class.

class Timelapse():

    def __init__(self,folder='',size=(1920,1080),prefix='',format='png'):
        self.frame = 0
        self.folder = folder
        self.size = size
        self.prefix = prefix
        self.format = format

    def screenshot(self):
        save_path = os.path.join(self.folder,'%s_%s.%s'%(self.prefix,self.frame,self.format))
        rs.CreatePreviewImage(save_path, view=None, size=self.size, flags=2, wireframe=False)
        self.frame += 1
        return self.frame

The Sphere class.

class Sphere():

    def __init__(self,radius,location=False,color=(255,0,0),path=[]):
        self.location = location if location else path[0]
        self.radius = radius
        self.object = rs.AddSphere(self.location,self.radius)
        self.color = self.set_color(color)
        self.material = None
        self.set_material(color)
        self.path = path
        self.frame = 0

    def move(self,position):
        rs.MoveObject(self.object,tuple([sum(x) for x in zip(position,(-n for n in self.location))]))
        self.location = position
        return position

    def advance(self):
        self.frame += 1
        return self.move(self.path[self.frame])

    def set_color(self,color):
        rs.ObjectColor(self.object,color)
        self.color = color
        return color

    def set_material(self,color):
        if rs.ObjectMaterialIndex(self.object) == -1:
            self.material = rs.AddMaterialToObject(self.object)
        rs.MaterialColor(self.material,color)
        self.color = color
        return color

Tests the animation.

def run_test():
    # Set time lapse length
    max_frames = 330
    # Sets folder to store images and creates a time lapse object
    folder = rs.BrowseForFolder(message='Select folder to store frames')
    timelapse = Timelapse(folder=folder,size=(1920,1080),prefix='test',format='png')
    # Creates a point array from base curves to specify a motion path for the spheres to follow
    curves = rs.GetObjects("Select curves for path vector", rs.filter.curve,True,True)
    point_paths = [rs.DivideCurve(curve,max_frames+1) for curve in curves]
    # Creates the sphere objects
    spheres = [Sphere(20,None,(50,50,50),path) for path in point_paths]
    # Run until frame rate hits max frames
    while timelapse.frame < max_frames:
        for sphere in spheres:
            sphere.advance()
        timelapse.screenshot()

Run the script.

if __name__=="__main__":
    run_test()

Hexagonal Cloth Chair

2019-08-01T00:00:00+00:00

Hexagonal semi-rigid cloth

HEX 2.0 is a concept chair design I’ve been working for a few years now. I first thought about it with Marina Peniche at 2008, a few years later I built it with some modifications. It looked good but it wasn’t comfortable. This is my third attempt into finding the right amount of hexagons with the exact amount of tension.

My start approach will be to use to use a mesh relaxation algorithm in Grasshopper’s Kangaroo physic’s engine and try to design in a more realistic environment. I have a slight intuition of what is wrong with the cloth based on siting in the prototype.

This is the current state of the chair. Sorry for the bad photos, had a backup problem. Unpardonable… I think NASA did the same things with the original film of the lunar landing.

Bridge Form Finding With Genetic Algorithms

2019-07-15T00:00:00+00:00

Basic strength and simple fitness evolutionary algorithm

The goal is to create a bridge like model that suffices these parameters in the most optimal way:

Create a link between two separate planes.
Reduce the amount of material used.
Increase the strength of the link.

A Rhino Python script will be used to decode the gene sequence into a 3D structure. Then it will be process by Kangaroo (or FlexHopper) to evaluate its deformation. This and the total amount of material used its taken into account in the fitness function that defines the direction of the genetic algorithm run by Galapagos.

Concept drawings - Expected geometry

Grasshopper definition (Was created using an old plugin version, not tested in GH for Rhino6)

Unevolved (random) organism v.s. Evolved and more structurally fit organism

| file: bridge_hex_2.gh

Retro Mosaic Script

2017-02-02T00:00:00+00:00

Script de reconocimiento entre similitudes de dos imágenes

Esta script lo escribí como parte de un trabajo que requería del reconocimiento de areas pobladas en imágenes satelitales para la recopilación de datos estadísticos. Decidí hacer el experimento con fotos normales y mosaicos retro en vez de los patrones urbanos a distinguir en el proyecto principal. Seguiré experimentando con este script a manera de filtro pues me pareció interesante el resultado.

Este es el mismo principio aplicado a un video.

Los videos se pueden reproducir al mismo tiempo.

Additive 3d Machine Concept

2016-01-24T00:00:00+00:00

Concept machine derived from a startup creative session

This was a design exercise at the beginning of the incubation period of SPK. The project was abandoned because it didn’t offer many advantages over a conventional 3 axis printer, never the less, I’d love to see it working sometime. In the meantime I think its worth sharing.

Relaxing Meshes Concept Roof

2015-06-08T00:00:00+00:00

Roofing concept utilizing relaxing meshes to force reverse catenary curves

This is a concept of a lightweight roof to explain how complex structural designs can be achieved with simple compression and tension structural rules.

Mesh relaxation example

Case study proposal