cheng leon 595491 algorithmsketches

46
ALGORITHMIC SKETCHBOOK 2014 AIR LEON CHENG ARCHITECTURAL DESIGN STUDIO

Upload: leon-cheng

Post on 12-Mar-2016

213 views

Category:

Documents


0 download

DESCRIPTION

 

TRANSCRIPT

Page 1: Cheng leon 595491 AlgorithmSketches

ALGORITHMIC SKETCHBOOK 2014

AIR

LEON CHENG

ARCHITECTURAL DESIGN STUDIO

Page 2: Cheng leon 595491 AlgorithmSketches

2

LEON CHENG

STUDIO 13, Semester 1, 2014University of MelbourneTutors: Brad Elias and Philip Belesky

2014 Architectural Design Studio: AIR Journal2014 Architectural Design Studio: AIR Algorithmic Sketchbook

Printed in Melbourne

I would like to take this opportunity to thank Brad Elias and Philip Belesky for their guidance and tutelage for which without it, the publication of this book would not have been possible.

Page 3: Cheng leon 595491 AlgorithmSketches

3

SUMMARY

Week 1 Basic Curves and Loft 4Week 2 - 6Week 3 Week 4Week 5Week 6Week 7Week 8Week 9Week 10Week 11Week 12

Page 4: Cheng leon 595491 AlgorithmSketches

4

Week 1 - Basic Curves and Loft

LOOSE TIGHT

NONE

REBUILD

REFIT

A series of 3 curves are lofted and tested with various loft options available from Grasshopper.

Visible differences can be seen in as the loft options varied in the shaded mode.

SHADED MODE

NORMAL

Page 5: Cheng leon 595491 AlgorithmSketches

5

LOOSE TIGHTNORMAL

NONE

REBUILD

REFIT

RENDERED MODE

No significant differences seen between the various loft options in rendered mode.

Page 6: Cheng leon 595491 AlgorithmSketches

6

STRAIGHT DEVELOPABLE

NONE

REBUILD

REFIT

UNIFORM

Major differences can be seen between the 'None', 'Rebuild' and 'Refit' modes.The curves that form the loft varied in numbers.

SHADED MODE

Page 7: Cheng leon 595491 AlgorithmSketches

7

STRAIGHT DEVELOPABLEUNIFORM

NONE

REBUILD

REFIT

RENDERED MODE

The significant difference between the 6 loft options is the 'developable' option.Here we can see a drastic change in the continunity and came out with rather sharp edges as compared to the normal loft option.

Page 8: Cheng leon 595491 AlgorithmSketches

8

In this exercise, I started out with 3 curves lofted to form a complex shape and slowly adjusted the control points in Rhino to view the possibiliies that can be generated by the algorithmic computation.

Starting with the normal curves, I then went on to alter the varios control points of the individual curves in the 'Z' axis respectively. Following which, exaggerations were made to the curves in the 'Y' axis. The final 2 iteraterations were subjected to all 3 axis of alterations and thus, arriving at a rather complex shape, which is almost impossible to be illustrated.

This activity shows me how complex and vast algorithmic computation can become. And the number of design possibilities that can be discovered.

Page 9: Cheng leon 595491 AlgorithmSketches

9

Page 10: Cheng leon 595491 AlgorithmSketches

10

Week 2 - Understanding Geometry, Transformations and Intersections

Page 11: Cheng leon 595491 AlgorithmSketches

11

Page 12: Cheng leon 595491 AlgorithmSketches

12

Page 13: Cheng leon 595491 AlgorithmSketches

13

Page 14: Cheng leon 595491 AlgorithmSketches

14

Page 15: Cheng leon 595491 AlgorithmSketches

15

Page 16: Cheng leon 595491 AlgorithmSketches

16

Page 17: Cheng leon 595491 AlgorithmSketches

17

Page 18: Cheng leon 595491 AlgorithmSketches

18

Page 19: Cheng leon 595491 AlgorithmSketches

19

Page 20: Cheng leon 595491 AlgorithmSketches

20

In this week's task, we are supposed to generate a form based points generated from three different data input. I've decided to spice things up a little by choosing the following data set related to the Barclay's Premier League (Soccer):1.) Seasons 12/13 and 13/142.) Week 1-203.) Total number of Goals

The resultant form was then lofted.The loft curve was divided to form points, which were then put through the Delaunay Edges function. I found this very interesting as the lofted form was segregated into various lines, enabling a whole new range of possibile functions that I could use. Similiarly, the mesh function was applied to the loft, enabling me to change the number of surfaces that makes up the form.

In summary, I've managed to generate a set of points into curves and then into a lofted surface. Afterwhich, curves and points were then created from the lofted surface.

This exercise helped me to understand the fundamentals of vector, points and how the various components work.

Week 3

Page 21: Cheng leon 595491 AlgorithmSketches

21

Page 22: Cheng leon 595491 AlgorithmSketches

22

Page 23: Cheng leon 595491 AlgorithmSketches

23

In this exercise, I tested with forming a surface with 3 curves, which was then subjected to a series of geodisic function to form a triangular shape pattern on the surface.

Here I can start to see how complex shapes in buildings come to reali-sation and construction. From this, it is possible, through extrusions and alteration of data inputs to formulate the detail con-nections for the triangula-tion framing.

I'll be looking forward to how these shapes can evolve into different patterns instead of using the triangle.

Page 24: Cheng leon 595491 AlgorithmSketches

24

Page 25: Cheng leon 595491 AlgorithmSketches

25

Here, I played around with the forms by trying to come up with my own version of the Birdnest stadium.

However, I find myself facing difficulties as my technical skills and understanding of grasshopper is insufficient.

There is still much progress and understanding of the algorithms to be made.

Page 26: Cheng leon 595491 AlgorithmSketches

26

Week 4

This week's task, is to generate a series of scaled orthorgonal shapes and multiplying them to form a pattern.

Playing around with the algorthim, I took the product of the patterning algorithm and created a seires of extruded curves throughout various divided surfaces to form a stacking design.

I realise that it is very important to understand the kind of data that the components output. Only then will I be able to fully understand the component.

Page 27: Cheng leon 595491 AlgorithmSketches

27

After forming the various stacked curves, I thought about how it is to be constructed and gave it triangular meshes to see the outcome. It became very complex with many connections but I was able to play around with the component to determine the optimal design.

Page 28: Cheng leon 595491 AlgorithmSketches

28

B.02 CASE STUDY - 1.0Project: Biothing - Seroussi Pavilion (2007)Principal Designer: Alisa Andrasek Location: Paris

The formulation of this project is divided into two main phases. The grasshopper algorithm below shows clear distinction between the two phases.

The algorithm begins with an input of four basic curves which is then extrapolated into the final form.

1.) The initial phase of the algorithm generates the ‘legs’ like structure as seen in the picture on the right. It is formulated by introducing point charges which flows outwards and diverging away from the various charges, deriving the form of the structure.

Page 29: Cheng leon 595491 AlgorithmSketches

29

2.) The resultant curves are then extrapolated to form the height of the structure through manipulation of the various ‘legs’. This is achieved by a series of multiplication and moving of the points on individual curves.

Page 30: Cheng leon 595491 AlgorithmSketches

30

B.02 CASE STUDY - 1.0

Kangaroo Physics plugin is used in this algorithm. It is interesting how the phyiscs component is able to calculate and determine an eventual form of the design. Here, I’ve tweaked the parameters to understand how the component works.

Page 31: Cheng leon 595491 AlgorithmSketches

31

From this basic exercise and understanding of the material system and its algorithm, I can see the vast potential in digital design as structures and materials are able to be tested virtually. This would save cost and time.

Furthermore, design ideas and forms can be derived from various force inputs particular to the project site and used as an abstract form.

The algorithm can be diagnosed into three major parts.

First, a series lofts are generated from random points through the voronoi component.

Second, the lofts deconstructed to form breps and into meshes, which are welded together.

Finally, the mesh is then analysed through anchor points and points through out and put through the kangaroo physics compoent which inputs forces to give an output.

Page 32: Cheng leon 595491 AlgorithmSketches

32

Week 5

Page 33: Cheng leon 595491 AlgorithmSketches

33

Page 34: Cheng leon 595491 AlgorithmSketches

34

This exercise is an attempt to recreate the Biothing Pavilion.

I played around with the graph styles as well as the fields to generate a rather curvy like leg structures.

The field component was rather intriguing as it reminds me of the flow of energy from a point and out into space. Resemblence of whirlpools can see seen. I like how it is evolving as I increased the strength of the charges.

Page 35: Cheng leon 595491 AlgorithmSketches

35

A patterning exercise with the use of a voronoi component together with various graph styles and culling pattern

Page 36: Cheng leon 595491 AlgorithmSketches

36

Week 6This week focuses on the management of lists and methods to organise the data from the definition to be extracted so as to facilitate specific design inputs.

Tree Dimensions

Tree Structure

Tree Menu

Tree Dimension exercise introduced the management of points and allowed me to see how every point has a unique address. Specific points can now be called out to further expand the algorithm.

Tree Structure here is similar to the dimension as it takes on a 3D sphere, allowing me to see clearer how this could be fitted into my design.

Tree Menu was a bit more interesting as I looked into generating patterns across a surface based on booleans.

Page 37: Cheng leon 595491 AlgorithmSketches

37

Path Mapper

Python Script

The python plug in is where I get to see the bones of Rhino. How boxes, lines, points, etc are drawn. Here, I am able to create my own component based scripting language. I find it difficult to fathom coding language.

The Path Mapper function is a summary of management of tree data. Here, I found it a little bit more difficult to comprehend the functions. However, this is where I really got to learn the roots of how lists are managed.

Page 38: Cheng leon 595491 AlgorithmSketches

38

Reverse Engineering

The main concept of the contemPLAY pavilion was the moire effect, which curved strips of cladding formed. We took two base curves drawn in Rhino, representing the shape of the

pavilion, then divided them to extract points out. Vertical lines were then connected from the top curve to the bottom curve and points on

the vertical lines were shifted in order to generate a curved strip.

The inner cladding was drawing using the same definition as the outer cladding. Except that an offset component was

used to project the claddings on to the inside.

Page 39: Cheng leon 595491 AlgorithmSketches

39

With two curves referenced, we then designed for two additional bracings to be in between. This would allow for the strips to be fixed onto the structure.

This was achieved by projecting a vertical line from the top to the bottom curve and using the “List Item” componenet to acquire our desired row of points.

Finally, to provide support inbetween the bracings and the entire structure, a truss pattern was designed and projected onto the definition.

Page 40: Cheng leon 595491 AlgorithmSketches

40

Despite replicating the moire strips and the overall form of the pavilion, I feel that the reverse engineering process was not very successful. In our case, we managed to generate the form but they were generated seperately. The main components of the definition were on their own and was not able to integrate with other (i.e. strip cladding and trusses)

Page 41: Cheng leon 595491 AlgorithmSketches

41

Main part of this project was dealing with ‘Lists’ and how we can better extract the points which we want to be further worked on and our group faced many problems with that.More research and playing around with the program is required. We need a better understanding of the program to fully utilise Grasshopper’s abilities.

Page 42: Cheng leon 595491 AlgorithmSketches

42

Week 7From the original definition produced from the reverse engineer-ing exercise, we then took the definition further by adding random components and seeing what forms we might produce.

This species was a product of further extrapolating points from the curves and plotting an arc, which produced an interesting wave of strips

Here, cull pattern was used to reduce the density of strips as well as the integration of the trusses to create this form.

Page 43: Cheng leon 595491 AlgorithmSketches

43

The strips from the original definition were extended and 'Railed' with the original base curves to produce a shell-like stripped form with varying sizes.

In this definition, two rail sweep produced an 'open' form compared to the previ-ous species.

Page 44: Cheng leon 595491 AlgorithmSketches

44

Similar to the previous two definitions, this species looked at the rail revolution to produce a whirlpool effect with the strips and trusses.

Another set of revolving components applied to the frame of the strips.

Page 45: Cheng leon 595491 AlgorithmSketches

45

Page 46: Cheng leon 595491 AlgorithmSketches