Algorithmic Sketchbook

ARCHITECTURE DESIGN STUDIO

AIRALGORITHMICSKETCHBOOK

ALICE KHOURY587451

TUTORS: HASLETT AND PHILIP

2

WEEK ONEPARAMETR IC AND

ALGOR ITHMIC DES IGN

WEEK ONE

VORONOI 3D TRIANGULAITON ALGORITHM

Parameter, in the most sense, is a factor that helps to define the overall limits and performance of a system

PARAMETRIC DESIGN:- working parametrically- understand how data flows- divide a model into manageable parts- think Abstractly- think mathematically-think algorithmically

3

CURVESLOFT ING IN GRASSHOPPER

WEEK ONE

I created a series of 4 curves in rhino and lofted them to produce the form (left). Following this, I added these curves into rhino, which allows for the manipulaiton of the lofted form multiple tmes.

Lofted grasshopper forms, manipulated using control points.

Loft: Nurb surface that is created through a set of curves

4 WEEK TWO

WEEK TWOUNDERSTANDING GEOMETRY, TRANSFORMATIONS

AND INTERSECTIONS

CURVE MENU

Creating a set of points then con-verting them to lines and curves in grasshopper

Examples of using curves in grasshopper as ele-ments in ideas for our parametric model; drives a more complex geometry

Interpolate curves

Polyline

5WEEK TWO

OR IENT COMPONENT2D REPRESENTAT ION OF GEOMETRY

XY plane - 2D projection of surface. The image left de-scribes a 2D representation of geometry which is good for working with 2D panels. Much fater to construct in grasshopper.

It shows potential for use in the project proposal as a means of preparing for laser cutting.

Grasshopper component

Left image: orient component turned into hundreds of pieces, I only wanted about 20 to 30 to work with, so I employed a slider attached to the contour/grid and changed it from .250 to 5. The right image despicts this change in panels drasticall reduced. The image on the left slowed my computer down when I tried to work with it.

6

WEEK TWOUNDERSTANDING GEOMETRY, TRANSFORMATIONS

AND INTERSECTIONS

WEEK TWO

Rolled out components of original curved surface, much quicker than

manually adjusting in Rhino like we did in Virtual Environments

Approximating 2D geometries using 2D elements - placing a rectangular surface on the points within a curved plane. Links to ‘centroid’

7WEEK TWO

Design iterations and playing with point reference to 2D shape. The rectangular component has been copied onto the curve here, and the centroid of the component is on the curve. X-Y planes have minimal rotation, just chang-ing the position.Below: Baked component

8

WEEK TWOUNEXPECTED RESULTS

WEEK TWO

When trying to experiment with 2D shapes on a curve, I ended up with this form and texture of a grid. I was not intendingfor the form to take this shape.

During experimentation, I was trying to link ideas of 2D shapes to that of materiality, as computa-tional design allows us to take those notions into account.

9WEEK TWO

10

WEEK THREECONTROLLING THE ALGORITHM: LISTS, FLOW

CONTROL, MATCHING

GRID SHELL

WEEK THREE

Construction phases of shell. 1 - dividing curve, 2- arc loft then rebuild (missing component apparent in third image). All incorporating Grasshop-per utilities of curve, shift and BANG

1 2 3

11

GR ID SHELLGRASSHOPPER AND M ISTAKES

WEEK THREE

Final grasshopper layout, provided for an interesting pattern on the grid shell

Attempt at creating my own form, Grasshopper didn’t like it but this bold shaped ball form was the result

12

WEEK THREECONTROLLING THE ALGORITHM: LISTS, FLOW

CONTROL, MATCHING

WEEK THREE

CREATING POLYLINES FROM POINTS

Utilising voronoi components to create polylines for points. These patterns created were made by creating a sequence of numbers representing all of the indexes in voronoi list (cells)

The numbers are shuffled so cells are not stored in an order. The Jitter component shuffles numbers

13

SURFACE ITERAT IONSVORONOI COMPONENTS

WEEK THREE

Partition list componet, where the cells were offset by -0.05, therefore linking some of the cells to one another. This component created the most interesting surface pattern

14

WEEK FOURFIELD FUNDAMENTALS

WEEK FOUR

The points in the field either atract their sur-ounds or repel them, depending on their numeri-cal value (ie positive or negative).

15WEEK FOUR

FRACTAL TECTRAHEDRAIRREGULAR FORM

Triangulated form

Variations within the form, manu ‘unbalanced’ outcomes

16

FRACTAL TECTRAHEDRAOUTCOMES

output of one function as input of new function deconsstruct brep

Baked component, aesthetically pleasing

17

EXPRESS IONSIRREGULAR FORM

Expressions - mathematical expressions, with notions of - input parameters- associative definition by scaling points on lofted surface using attractor point, then manupulate radius of circle

WEEK FOUR

output of one function as input of new function deconsstruct brep

18


Small changes in circles, where a point attractor is used to make variations in the circle sequence

WEEK FOUR

19


WEEK FOUR

20

I TTERAT IONSVOLTADOM 10 ITERAT IONS

Perspective View Perspective ViewTop Top

WEEK FOUR

21

PAV IL IONLAG I PAV IL ION EXPLORAT ION

From the VoltaDom precedent, I created 10 iterations from the provided script for the VoltaDom project. Some of the forms resulted in a maze like pattern, whereas others showed potential to be utilised in the LAGI brief.

The baked surface above was one of the more suc-cessful iteraitons of the 10, where a pavilion like structure was created. the circular shapes provide a sheltering from the external environment, and create a

WEEK FOUR

22

WEEK FIVEGRAPH CONTROLLERS

Base Y value from 0 to 1The following patterns are all variations in graph mapper, where a series of infinite patterns could be generated.

WEEK FIVE

23

Graph Mapper script, allowing for many variations withing the patterning

WEEK FIVE

24

IMAGE SAMPL ING

Variable y=0.266, Surface divide component, then re-paramaterise

WEEK FIVE

25

IMAGE SAMPL INGSOMETHING HERE???

Low U and V count Two images imposed on one another Balanced U and V count of com-bined images

Green differentiates one image from another

WEEK FIVE

26

IMAGE SAMPL INGTAN INPUT

Tan input expression to offset circles upwards

Iteration 1

WEEK FIVE

27

UNEXPECTED RESULT

This was the second iteration, where a series of cones forms all projected to a designated point due to the tan input, producing a very strange form

Iteration 3

WEEK FIVE

28

EVALUAT ING F IELDSUSING PO INT CHAR GERS

Using positive point chargers, and curve division components, the points in space put lines through the fields.

The process of pushing points through the field and interpolating the design points, where a ‘field line’ component was utilised.

WEEK FIVE

29

Merge Field component

Chargers pushing the points away from field, line component utilised as opposed to circle

WEEK FIVE

30

EVALUAT ING F IELDSLINE CHAR GE

Line charge component within the field, chargers pointing away from it. Provides variation within the overall form, where the baked from below looks split at the charge of the inserted line.

WEEK FIVE

31

Using line charge to create this field, both baked components, not merged with ‘Merge Fields’

WEEK FIVE

32

REVERSE ENG INEER INGATMOSPHER IC TESSELLAT ION

Formulate the frame/skeleton on grasshopper Populate a triangle

surface with the ‘barnacles’. 3, four edged elements on each triangle.

Triangulate the surface

PHASE ONE PHASE TWO

Combining the two creates the Atmospheric Tessellation installation

ASSUMED DES IGN PHASES

WEEK FIVE

33

ATMOSPHERIC TESSELLATION

REVERSE-ENGINEERSTEPS 1-4

1. Tri-grid, 2D grid with triangular cells, size of 6, element X 10, element Y 7

3. Point Component fol-lowing output of polygon centre, X and Y direction

4. Voronoi component, plugged into region inter-section

2. Polygon Centre - area centroid of polygon shape focused on

WEEK FIVE

34



5. Scale component - objects were then scaled, creating a series of 3 four edges shapes within the triangles, just like the precedent project.

6. Loft - Following graft tree and region intersec-tion, two lofitng processes occur. This is the underside.

WEEK FIVE

35



7. Loft - Upper loft, attaching to the underside through a mesh component. The ‘simple mesh’ and ‘mesh join’ components need to be utilised.

8. Loft - Combining the two loft meshes together as one.

WEEK FIVE

36

WEEK SIXARANDA LASCH CONTINUOUS PATTERNING

1. Previous fractal tetrahedra component 2. Evaluate curve component, finds and con-nects the mid point of every edge to znother random edge. Creating a continuous pattern

3. Bezier span, combined with ‘jitter’ compo-nent

4. Shift Paths, shifting the indices in all data paths, joining bezier span and unroll

WEEK SIX

37

ARANDA LASCHUNROLL ING BREPS

Python scriptable component. Allowing the abil-ity to work with the form in a 2D mode

WEEK SIX

38

TREE MENUSHIFT PATH COMPONENT

FIRST PHASE SECOND PHASE

Set three surfaces

Surface Divide U-20 V-10

Average the points

Surface Divide -sphere

Divide Surface U-10 V-10

Polyline component{0;0}{1:1}

Polyline component{0;0}{1;1}{0;1}

Polyline component{0;0} set boolen true{1;1}{0;1}

WEEK SIX

39

Polyline component{0;0}{2;0}{0;2}

Polyline component{0;0}{2;0}{2;2}{0,2}

WEEK SIX

40

TREE STAT IST ICS AND V ISUAL ISAT ION

Tree StatisticsL=11 C=11Next phase is to visualise how objects are stored in a tree

Simplify and Graft- alternatively (and much easier/quicker) to just right click and graft

Text statistics component and text tag

Baked text tag

WEEK SIX

41

Baked text tag

BASE FORMSREVERSE ENG INEERED PROJECT

SCRIPT 1CONSTANT QUAD SUBDIVIDE

SCRIPT 2HEXAGON CELLS

SCRIPT 1SUBDIVIDE TRIANGLES

BASE FORM

BASE FORM

BASE FORM

WEEK SIX

42

MATR IX50 ITERAT IONS

WEEK SIX

SPECIES 1

PETR

USIO

NS

SCALE FACTOR 1 (TOP): 0.450Z FACTOR: 5SCALE FACTOR 2 (BOTTOM):0.371

SCALE FACTOR 1 (TOP): 1.225Z FACTOR: 4SCALE FACTOR 2 (BOTTOM): 0.130

TRIANGULATED SUB PANELSCAE FACTOR 1 (TOP): 1.560Z FACTOR: 1SCALE FACTOR 2 (BOOTTOM): 0.1635PATCH DISABLED

TRIANGULATED SUB PANELSCAE FACTOR 1 (TOP): 1.560Z FACTOR: 1SCALE FACTOR 2 (BOOTTOM): 0.162PATCH DISABLED

TRIANGULATED SUB PANELSCAE FACTOR 1 (TOP): 1.560Z FACTOR: 1SCALE FACTOR 2 (BOOTTOM): 1PATCH DISABLED

TRIANGULATED SUB PANELSCAE FACTOR 1 (TOP): 1Z FACTOR: 1SCALE FACTOR 2 (BOOTTOM): 1PATCH ENABLED

TRI PANELCONSTANT QUADSCALE FACTOR 1 (TOP): 2.3Z FACTOR: 1SCALE FACTOR 2 (BOTTOM) 0.41

U DIVISION: 1V DIVISION: 3SCALE FACTOR 1 (TOP): 0.488SCALE FACTOR 2 (BOTTOM): 0.846Z FACTOR : 2PATCH DISABLEDTRIANGULAR PANELS

U DIVISION: 1V DIVISION: 3SCALE FACTOR 1 (TOP): 0.6SCALE FACTOR 2 (BOTTOM): 0.488Z FACTOR : 2X FACTOR: 6PATCH ENABLEDTRIANGULAR PANELS

U DIVISION: 1V DIVISION: 3SCALE FACTOR 1 (TOP): 0.795Z FACTOR : 2SCALE FACTOR 2 (BOTTOM): 0.914

TRIANGULATED SUB PANELSCAE FACTOR 1 (TOP): 1.560Z FACTOR: 1SCALE FACTOR 2 (BOOTTOM): 1PATCH ENABLED

43

MATR IX50 ITERAT IONS

WEEK SIX

SPECIES 2

SOLI

D FO

RMS

SPECIES 3

HEXA

GONS

TRI PANEL: CONSTANT QUADSUBDIVIDE: 2SCALE FACTOR 1 9TOP0: 0.853Z FACTOR: 6SCALE FACTOR 2 (BOTTOM) 0.964PATCH DISABLED

U DIVISION: 4V DIVISION: 7SCALE FACTOR 1 (TOP): 0.795Z FACTOR: 4SCALE FACTOR 2 (BOTTOM): 0.914

SCALE FACTOR 1 (TOP): 1.548Z FACTOR: 2SCALE FACTOR 2 (BOTTOM): 0.769PATCH DISABLED

U DIVISION: 5V DIVISION: 5SUBDIVIDE: 1SCALE FACTOR 1 (TOP): 1SCALE FACTOR 2 (BOTTOM): 0.9Z FACTOR: 2X FACTOR: 2Y FACTOR: 2

U DIVISION: 1V DIVISION: 5SUBDIVIDE: 1SCALE FACTOR 1 (TOP): 2SCALE FACTOR 2 (BOTTOM): 0.908Z FACTOR: 2

HEXAGONSCALE FACTOR 1 (TOP): 0.417Z FACTOR: 6SCALE FACTOR 2 (BOTTOM): 1.316PATH ENABLED



HEXAGONSCALE FACTOR 1 (TOP): 1.48Z FACTOR: 3SCALE FACTOR 2 (BOTTOM): 0.8PATCH DISABLED

HEXAGONSCALE FACTOR 1 (TOP): 1Z FACTOR: 1SCALE FACTOR 2 (BOTTOM): 1PATCH ENABLED

TRI PANELCONSTANT QUADSCALE FACTOR 1 (TOP): 0.189Z FACTOR: 6SCALE FACTOR 2 (BOTTOM): 0.964PATCH DISABLED

U DIVISION: 10V DIVISION: 15SCALE FACTOR 1 (TOP): 7Z AFCTOR: 3SCALE FACTOR 2 (BOTTOM): 1.3

44 WEEK SIX

SPECIES 4

POD

POTE

NTIA

L

U DIVISION: 8V DIVISION: 10SCALE FACTOR 1 (TOP): 0.6Z FACTOR: 5SCALE FACTOR 2 (BOTTOM): 0.9PARAMETER (T): 0.75PATCH DISABLED


U DIVISION: 13V DIVISION: 15SCALE FACTOR 1 (TOP): 0.6Z AFCTOR: 2SCALE FACTOR 2 (BOTTOM): 0.9PARAMETER (T): 0.1, 0.3PATCH DISABLED

U DIVISION: 13V DIVISION: 15SCALE FACTOR 1 (TOP): 0.6Z AFCTOR: 4SCALE FACTOR 2 (BOTTOM): 0.869PARAMETER (T): 0.1PATCH ENABLED

U DIVISION: 6V DIVISION: 8SCALE FACTOR 1 (TOP): 0.6Z AFCTOR: 6SCALE FACTOR 2 (BOTTOM): 0.9PARAMETER (T): 0.75PATCH ENABLED

SCALE FACTOR 1 (TOP): 0.928Z FACTOR: 1SCALE FACTOR 2 (BOTTOM): 0.769PATCH ENABLED

U DIVISION: 8V DIVISION: 20SCALE FACTOR 1 (TOP): 1.0Z FACTOR: 1SCALE FACTOR 2 (BOTTOM): 0.7PARAMETER (T): 0.8

REVSRF 3: REVERSE UVU DIVISION: 6V DIVISION: 2SCALE FACTOR 1 (TOP): 0.8Z FACTOR: 7SCALE FACTOR 2 (BOTTOM): 0.9PATCH ENABLED, SUBDIVIDED QUADSKEWED QUADS T: 0

U DIVISION: 6V DIVISION: 8SCALE FACTOR 1 (TOP): 0.6Z FACTOR: 6SCALE FACTOR 2 (BOTTOM): 0.9CAP HOLES, CULL FACESBOOLEEN (FTFFF)

SCALE FACTOR 1 (TOP): 1.555Z FACTOR: 2SCALE FACTOR 2 (BOTTOM): 0.807PATCH ENABLED


U DIVISION: 4V DIVISION: 7SCALE FACTOR 1 (TOP): 0.795Z FACTOR: 4SCALE FACTOR 2 (BOTTOM): 0.914

45

SPECIES 5

EXTR

USIV

E PO

D PO

TENT

IAL

U DIVISION: 5V DIVISION: 8SCALE FACTOR 1 (TOP): 0.488SCALE FACTOR 2 (BOTTOM): 0.846Z FACTOR: 2PATCH ENABLEDRANDOM QUAD PANEL S:5

U DIVISION: 3V DIVISION: 5SCALE FACTOR 1 (TOP): 0.3SCALE FACTOR 2 (BOTTOM): 0.9Z FACTOR: 5PATCH DISABLEDRANDOM QUAD PANEL S:1, SUBDIVIDE QUAD

TRI PANELCONSTANT QUADSUBDIVIDE: 1SCALE FACTOR 1 (TOP): 1Z FACTOR: 1SCALE FACTOR 2 (BOTTOM): 1

SCALE FACTOR 1 (TOP): 1.555Z FACTOR: 2SCALE FACTOR 2 (BOTTOM): 0.807PATCH DISABLED

TRI PANELCONSTANT QUADSCALE FACTOR 1 (TOP): 1Z FACTOR: 2SCALE FACTOR 2 (BOTTOM): 0.8PATCH ENABLED

TRI PANELCONSTANT QUADSCALE FACTOR 1 (TOP): 0.432Z FACTOR: 8SCALE FACTOR 2 (BOTTOM): 2PATCH DISABLED

U DIVISION: 5V DIVISION: 10SCALE FACTOR 1 (TOP): 1.3Z FACTOR: 5SCALE FACTOR 2 (BOTTOM): 0.3PARAMETER (T): 0.9, 0.7PATCH ENABLED


REVSRF 3: REVERSE UVU DIVISION: 2V DIVISION: 1SCALE FACTOR 1 (TOP): 0.3Z FACTOR: 7SCALE FACTOR 2 (BOTTOM): 0.9PATCH ENABLED, TRIANGULAR PANELS


U DIVISION: 3V DIVISION: 3SCALE FACTOR 1 (TOP): 0.3SCALE FACTOR 2 (BOTTOM): 0.9Z FACTOR: 7PATCH ENABLEDSUBDIVIDE QUAD, SKEWED QUAD T=0

U DIVISION: 1V DIVISION: 2SUBDIVIDE: 3SCALE FACTOR 1 (TOP): 0.427SCALE FACTOR 2 (BOTTOM): 0.583Z FACTOR: 2

46

SUCCESSFUL ITERAT IONSSELECT ION CR ITER IA

WEEK SIX

SELECTION CRITERIA - Ease of fabrication/assembly- Aesthetically pleasing and interesting- Differs from original base pattern- Creation of suitable pod structure to house

algae

In terms of feasibility and generation of the pod structure, our chosen successes in the iterative pro-cess need to be able to be applied to the concept of a ‘pod’. This pod pattern and form must be aesthetically pleasing and intriguing for the viewers of the pavilion, as the form generated will respond directly to the energy source of algae biofuel.

47

Second Definition - Hexagonal Pod Potential in tessellating hexagons, clean form and spatial diversity

Third Definition - Blockwork Interesting surface pattern stray-ing from original definition, limited pod use

Second Definition - Pertrusions Spacing between pod structures for possible pipes, interesting assembly

Third Definition - Valuted Interesting patterning that varies on top and bottom of design, pod structures evident

48

NON TEACHINGCLUSTERS - TRAVELLING SALESMAN

Populate a 2D surface, thinking about how to cluster a definition to a single point

Cull Index - after list item, use cull index to search through points to find the closest point

Create a line between the two points, show-ing what we did ‘visually’

Travelling Salesman - script for Cluster

NON TEACHING

49

GRAD IENT DECENTRECURS IVE PATTERNS

1. Set one surface

2. Surface divide and graft, keeps track of points

3. New surface point and number comp. + Unit Z (-1 downwards)

4. Surface closed point. Then cluster

5. Continue Cluster for 5 times, then attach to NURBS curve

NON TEACHING

50

GRAD IENT DECENTRECURS IVE PATTERNS

6. U+V count of 25

NON TEACHING

51

7. U=64 V=10

7. U=15 V=90

NON TEACHING

52

FORM EXPLORAT IONCURVES

Applying the three different scripts to differnt sets of base curves to see the results. In general terms, the curves seemed weel supported by the tessellating patterns.

NON TEACHING

53

PODS ON FORMEXPLORAT IONS

Constant Quad Subdivide

LUNCHBOX BAKED FORM

Hexagon Cells

FORM VARIATION

Subdivide Triangle

NON TEACHING

54

LARGER FORM GENERATION

55

GENERATED FORM

This is a render of what we propose our final design to look twards, as it curves adjusting to the levels of sunlight throughout the day in a Southeen manner. the curved surface provides a visually interesting object to

look at, where the pods can be seen in all of their presence.

In terms of parametric design, Part B has encouraged us to push our boundaries and creativity to the absolute limit with the aid of Grasshopper, and many of the designs that both myself and my peers have come up with suggest that we have taken a lot away from this subject already in terms of computational design and thinking.

NON TEACHING

56

PART CDETAILED DESIGN

Following the interim presentation, we explored form variation at a greater level and concluded on a form

that incorporated the pods, spacing for pipes and a steel frame to hold the structure together. The form

showed highlights the first major iteration we came up with, but this form was not optimized to our site just yet. We needed to develop our form in response to the sun and angle at which it hits the site.

PART C

57PART C

58 PART C

59PART C

60

FORM EXPLORAT IONGRID SHELL

This is the grid shell form utilized for the final of our design. In order to rationalize this form, we conducted a series of analytical tests, such as:- Radiation Analysis- Shadow Studies- Pod Studies

AERIAL VIEW

EASTERN VIEW

SOUTHERN VIEW

PART C

61

PROBLEMPRINT F ILE

In order to 3D print there had no be no revealed edges on the form. Therefore, on the open polysurface, the command ‘show edges’ was used, then ‘naked edges’ was selected to then ‘join edges’ to one another. This occurred on our polysurface.

UNJOINED EDGES

PART C

62

SOLUT IONPRINT F ILE

JOINED EDGES

REVEAL EDGES

NAKED EDGES

JOIN EDGES

PART C

63

JOINED EDGES

OPEN SURFACE - NEEDS TO BE CONVERTED TO A MESH IN ORDER TO PRINT IN STL FORMAT

CLOSED MESH - ABILITY TO BE USED AS AN STL FILE FOR THE 3D PRINTER

PART C

64

PR INT F ILESIMPL IF ICAT ION

Using Meshlab, specializing with 3D objects and fil-ters, I was able to simplify the mesh created for the

print file. While our mesh was quite simple and small in size, reaching just over 10cm in length, the more simple the object, the faster it will print.

Using the filters tab, then remeshing, simplification and reconstruction, quadric edge collapse deformation, and specifying the percentage of reduction at 0.5, (by half) a simpler version of the mesh is created.

This mesh was too simplified, and did not give a smooth surface, such as the one that would be generated from the connecting pods. This mesh was reduced by a percentage of 0.5 four times.

This mesh was a better percentage of simplifica-tion than the first, as there are no hard edged peaks such as the one above has. The mesh was simplified throguh Meshab at a reduction of 0.5 twice.

PART C

65

PODSTEMPLATE FOR LASER CUTTER

PART C

66

FIRST FORM - INTERIM PRESENTATION SECOND FORM - GENERATION

LADYBUGRADIAT ION ANALYS IS

PART C

67

THIRD FORM - FINAL FORM

HIGH

LOW

LEVELS OF RADIATION

PART C

68

FINAL FORM - NO PODS

FINAL FORM - POD INTEGRATION

SOLAR RADIATION IN LADYBUG DEFINITION PART C

69

LADYBUGRADIAT ION ANALYS IS

HIGH

LOW

LEVELS OF RADIATIONThis is the final form optimised with ladybug for solar radiation. The pavilion relies on the sun to promite alge growth, and the pods need the maximum amount of sun-light that they can access.

PART C

70

LADYBUGSHADOW STUD IES

LADYBUG GRASSHOPPER DEFINITION

PART C

71

WESTERN FACADE

NORTHERN FACADE

SOUTHERN FACADE

EASTERN FACADE

The lighter the shadowing, the less exposed it is to shade. These diagrams depict the final form over a shadow study of the annual amount of sunlight projected over the site (in a southerly direction).

PART C

72

LADYBUGPOD SHADOW STUD IES

These are the projected sun paths of Copen-hagen over 5 different pod shapes. We chose to cotinue with hexagons as they produced the least amount of shadowing onto sur-rounding pods.

HEXAGONS

CIRCLES

SQUARES

TRIANGLES

TRI GRID

PART C

73

1CM THICK

2CM THICK

3CM THICK

4CM THICK

7CM THICK

8CM THICK

KEY

Ideal range

High in compression

KARAMBASTRUCTURAL ANALYS IS

Karamba was used to analyse the strcture, and yielding an 8cm steel system as effective to work across the paviion. By this stage, our framing traced the pod shapes in a hexago-nal manner, differing from the form gener-ated after the interim presentation.

PART C

74

STEELSTRUCTURAL DEF IN IT ION

PART C

75

FORMSURFACE DEF IN IT ION

PART C

76

PODSPATTERN DEF IN IT ION - HEXAGONS

FINAL POD SURFACE PATTERN DEFINITIONUSING LUNCHBOX TO GENERATE HEXAGONAL POD SHAPES AND HEIGHTS

PART C

77

WESTERN FACADE

SOUTHERN FACADE

EASTERN FACADE

NORTHERN FACADE

PART C

Algorithmic Sketchbook

Documents

d panels

d shapes

d geometries

d elements

d projection of surface

d representation of

rectangular component

lofted grasshopper forms