Biomorphic

Chair, 2021

Biomorphic

Chair
AAVS Toronto Workshop, Summer 2021

Tools

SideFX Houdini
Autodesk Maya
Adobe Suite
Vray 5

Project Description

In this workshop taught by Nic Lee from the mediated matter group at MIT, I studied growth, agency, and adaptation in nature. Tasked with growing a chair using environmental forces we used Craig Reynolds’s boid algorithm in order to run flocking simulations which would form the chair based on various environmental and synthetically introduced forces. A finite element analysis was used for structural simulations which would inform and optimize the flocking simulation, followed by a Frei Otto inspired kernel density-based edge bundling algorithm which is used to optimize the form and printability of the chair.

Targets

Agents are instructed to target areas that it naturally misses. This synthetic force ensures the chair is usable..

Alignment

In proximity with other agents, they are instructed to align with eachother creating grain and flow within the chair

Movement

Agents are told to start at the bottom of the chair, and move upwards, as if they were being 3D printed.

Surface

Areas of contact between the chair and  a human are simulated as predatory areas creating surface area due to scattering

Structure

A finite element analysis of the chair dictates where the agents will cohese in order to create strength.

An Environmental Creation

Rather than directly designing the chair, an environment is created whereby a number of agents can form the chair based off of natural decisions such as....

Simulation Parameters
‍

01

Cohesion

Agents will gradually flock together as if a flock of birds were incressing their saftey through numbers.

This parameter is usefull for areas of the chair needing strength due to the increased density of agents.
‍
This parameter is controlled by an structural simulation (FEA) of the chair aswell as areas of the chair that may be under a higher stress such as the edges.

Agents Following Cohesion Map
C=1, S=0, T=0, A=1
‍

(FEA) Cohesion Map on Base Mesh
‍

Full Cohesion

No Behaviour

02

Separation

Agents will separate from each other as if the flock was trying to avoid colliding with other agents

This parameter is useful for areas of the chair needing surface area due to the decreased density of agents.

This parameter is controlled by a human contact map of the chair which ensures seating areas will have surfaces

Agents Following Seperation Map
C=0, S=1, T=0, A=1
‍

Human Contact Map on Base Mesh
‍

Full Separation

No Behaviour

03

Alignment

Agents will align directions with eachother as if the flock was following a leader

This parameter is useful for controlling the directional and visual flow of the chair.  

Given the intended 3D printing of the chair, agents start at the bottom and are pushed upwards as if they were being printed.

Agents Following Alignment Map
C=0, S=0, T=0, A=1
‍

Upwards Facing Vectors on Base Mesh

Vector Direction

No Behaviour

04

Targets

Agents will move towards a designated area as if the flock were moving towards a destination

This parameter is useful for guiding the agents towards areas where it may naturally miss

This is an artificial parameter which gives the designer more control over the movement of the flock. In this result it’s being used to guide agents up the arm rests and the center of the chair

Agents Following Target Map
C=1, S=0, T=1, A=1
‍

Target Locations Map on Base Mesh
‍

Target Locations

No Behaviour

Simulation Process

01

Agent Based Point Simulation

First the agent based simulation is run on points based on the various behavioural inputs given to the model
Parameters: C=0.156, S=1.385, A=0.6, T=0.3,NAgents=2000

02

Pathtracing the Agents

The path of each agent is then traced over time. This tells us how each agent behaved in response to it’s environment

03

Bundling the Paths

For an optimized 3D printed fabrication, kernel density based edge bundling is used in order to bind floating elements together, as well as emphasise form

04

Density Mapping

In preperation for the thickening of the paths each path is given a density value based on it’s proximity to another agents path

05

Mesh Creation

The paths with their assigned density values are then converted to a set of voxels which is then converted to a high quality mesh suitable for fabrication