Design Tooling - Metric-based Design

 

Return to front page

 
 

Information

Introduction

The notion of a metric for design becomes important as soon as a generative system is based on choice and selection and no human is directly in charge of making the decision as for instance in fitness functions in evolutionary design. Design metrics are often used in optimization of designs for structural or environmental goals. Essentially the choice of what to measure and how to evaluate and weigh can be seen as part of the design approach and not merely as a technical detail and is therefore an important design driver.
 
 
 
 

Ecotec developed by Andrew Marsh university of Cardiff- Performance driven design evaluation

A software developed by Andrew Marsh that allows the fast and interactive evaluation of building structures in context of light, sound and temperature.

Some scripting capabilities.
www.squ1.com

Form generation based on performance criteria

window shade generated from the light path of the sun in relation to a window opening in a wall. The device shades the opening for a certain time of day/time of year.

(c) Andrew Marsh

     

sunlightstudy in ecotec

shadeing louver generated from sunpath

 

 

 
   
   
 

Conditional louvers on Nurbs surfaces

An experiment done in Rhino scripting demonstrrates selective placement and orientation of louvers along a double curved NURB surface. This technique allows the generation of a set of louvers along a free form surface taking into account a light direction and shading of all openings from direct sunlight. The approach relies on the principle of using a design surface.

To use conditional statements in generative design is a simple way to introduce rules that apply locally or globally to adapt the design moves to a geometric context.

 

axel kilian 2003

     

louvers applied to a NURBS surface

- surface pattern

- source code in .rvb

 

-
   
   
 

Topological optimizer - by Ole Sigmund, Denmark

Optimzation of the topology of a structure given a maximum material envelope and a loadcase.

TOPOPT is an acronym for TOPology OPTimization. The TOPOPT-homepage is the homepage for the research group headed by Ole Sigmund. Currently we are sponsored by the Danish Technical Research Council (The "Phonon Project") and by the Poul Due Jensen Foundation.
developed and maintained by Dmitri Tcherniak, Ole Sigmund, Thomas A. Poulsen and Thomas Buhl

(c) Ole Sigmund

 

TOPOPT example images

TOPOPT working in 3d

 

 

  
       

TOPOPT example images

TOPOPT working in 3d

 

 

  
 
 
 

Fluid dynamic simulation using Cellular Automata

A hexagonal and aquare grid based CA simulates fluid dynamics through a set of collision rules
The CA model makes it possible to simulate a fluid which in nature is made up of billions of atoms with a very small finite set of elements that interact on a simple set of collision rules. The rules and the controlled hexagon enviroment allows for fluid like behavior at fraction of the computational power necessary to simulate a model with even closely the same amount of particles.
System is modeled based on the description in "the nature of mathematical modeling" by Neil Gershenfield.
Other grid patterns are possible, a grid pattern for instance which reduces the complexity of the necessary rules.

axel kilian 2003

 

- Fluid dynamics with cellular automata

- Rule set for a hexagon grid

- source code in .java

 

 
 
 
 

Light Catchers
In this example, VBScript is used to write generative scripts for execution within the Rhinoceros modeling environment. The script produces a digital geometric model which, in turn, can be exported to a Z-Corp 3-D printer. The merits of this methodology are demonstrated here through the model of an architectural surface composed of light-modulating conical components. The design intent in this example is a surface of responsive components which traverse a complex curved structure and steer toward a light. The written script is an explicit representation of this intention. Methods in the script use external parameters to generate a digital geometric model. The form of the subsequent printed model is calculated as a function of the initial parameters; two boundary splines and a vector indicating the direction of lighting. By varying these parameters, a set of design options can be generated and 3-D printed for comparison. The combination of scripting and 3-D printing allows complex design intentions to be managed in a concise, sharable format and modeled iteratively without manual intervention.

 

Yanni Loukissas 2003

- 3D Model of Light Catchers

- 3D Printed Light Catching Surfaces

 

 
 
 
 

Chicken Cooker
In this example, VBScript is used to write generative scripts for execution within the Rhinoceros modeling environment. The script generates an array of components which follow a ruled surface defined by the user. The components in this script are small reflectors which are individually aligned by the script to channel perpendicular rays of light to a single point in the digital environment. This task is complicated by the placement of reflectors along the original ruled surface. This example simply illustrates how a metric rule can be used to guide the individual geometries of a number of components in a complex composition.

The complexity of the the example is a function of the environment. Patterns might arise from such a strategy, however these will be informed contextual patterns and not based on abstract geometry. Here, the generated form is quite complex and would be difficult to model manually on a computer let alone physically. This example produces a matrix of light responsive apertures, each with a unique geometry. This is not accomplished by isolating each local condition and resolving it. It is done by specifying a general method which is applied to many different conditions. As long as each condition has the appropriate defining parameters, the general method can be applied. This results in a different final geometry for each distinct condition. In this way scripting can elevate the design process to the level of relationships as opposed to individual components.
 

Stylianos Dritsas +
Yanni Loukissas 2003

"When you watch an ant follow a tortuous path across a beach, you might say, "How complicated!" Well, the ant is just trying to go home, and it's got to climb over little sand dunes and around twigs. Its path is generally pointed toward its goal, and its maneuvers are simple, local responses to its environment. To simulate an ant, you don't have to simulate that wiggly path, just the way it responds to obstacles."

(Herbert Simon. From an interview with Doug Stewart, June, 1994.http://www.omnimag.com/archives/interviews/simon.html)