PART B: CRITERIA DESIGN
TABLE OF CONTENTS
B.1 RESEARCH FIELD - GENETICS B.2 CASE STUDY 1.0 - L-SYSTEMS B.3 CASE STUDY 2.0 - BLOOM REVERSE ENGINEERED B.4 TECHNIQUE DEVELOPMENT B.5 TECHNIQUE PROTOTYPES B.6 TECHNIQUE PROPOSAL B.7 LEARNING OBJECTIVES & OUTCOMES - REFLECTION B.8 APPENDIX - ALGORITHMIC SKETCHES REFERENCES
B.1 Research Field
GENETICS Genetics in design describes the process of generative design, which produces a large number of prototypical forms that are then evaluated on their performance in a simulated environment. A number of similar forms are eventually narrowed down due to their beneficial and survival enhaving traits, which are passed on to new generations. L-Systems can be used to model, or computationally ‘grow’, prototypes that mimick natural forms. It’s ease of creating numerous prototypes by simply changing paramter values makes culling unsuccessful iterations fast and effective. Through this method of culling, successful traits can be carried on to later generations.
B.2 Case Study 1.0 L-SYSTEMS
2 POINTS, 2 CLUSTERS, XZ PLANE
3 POINTS, HOOPSNAKE 6 LOOPS
ADD 3 MORE CLUSTERS ADJUST X,Y,Z VALUES ADJUST X,Y,Z VALUES ADJUST ADJUST X,Y,Z VALUES
ADJUST X,Y,Z VALUES ADJUST X,Y,Z VALUES
ADJUST X,Y,Z VALUES 7 LOOPS
3 POINTS, HOOPSNAKE ADJUST X,Y,Z VALUES, 5 LOOPS ADJUST X,Y,Z VALUES 6 LOOPS
3 POINTS HOOPSNAKE 5 LOOPS
ADJUST X,Y,Z VALUES 6 LOOPS
ADJUS 6 LOO
ADJUST X,Y,Z VALUES ADJUS 6 LOOPS
ADJUST X,Y,Z VALUES ADJUST X,Y,Z VALUES 5 LOOPS
ADJUST X,Y,Z VALU 6 LOOPS
T X,Y,Z VALUES ADJUST X,Y,Z VALUES CHANGE TO XY PLANE ADD 2 MORE CLUSTERS ADJUST X,Y,Z VALUES ADJUST X,Y,Z VALUES
ST X,Y,Z VALUES ADJUST X,Y,Z VALUES ADJUST X,Y,Z VALUES OPS
ST X,Y,Z VALUES ADJUST X,Y,Z VALUES ADJUST X,Y,Z VALUES
UES
ADJUST X,Y,Z VALUES
ADJUST X,Y,Z VALUES 4 LOOPS
REVERSE VECTOR DIRECTION ADJUST X,Y,Z VALUES
ADJUST X,Y,Z VALUES
ADJUST X,Y,Z VALUES
ADJUST X,Y,Z VALUES
ADJUST X,Y,Z VALUES 5 LOOPS
ADJUST X,Y,Z VALUES
ADJUST X,Y,Z VALUES
I believe this iteration has a very distant spine which I can relate to natural organisms in a way. The spine acts as a structural apparatus to both hold and add movement to an organism. It also has a sense of hollowness towards to bottom, almost like an empty frame, or skeleton.
I like the effect the increasing compactness of the top backbone gives. It makes the structure seem dynamic. The top section is in contrast to the bottom, which is a lot softer and doesn’t hold any particular recurring shape. In both the top and bottom sections however, they fan out to become more open than the centre of the object.
This one is very frame-like, I can imagine it acting as a structural frame that forms the primary structure of a building. I like how it remains abstract; the lack of a base or orientation makes it quite playful in nature. The varying line weights and areas of openness have made it a compelling iteration.
I was drawn to the symmetrical effect of this iteration; it appears like a structure that has been reflected through a body of water. It seems very clean and organised as all the lines are almost parallel. It is however very dense and comes across as quite a heavy structure with rare areas of relief.
BLOOM PROJECT ANALYSIS
The Bloom Plethora Project designed by Alisa Androsek and Jose Sanchez is unique in the way that it connects architecture and gaming to create a highly interactive experience. It was commissioned by the City of London for the 2012 Olympic Games and is still being exhibited at many locations around the world, including RMIT in Melbourne. The success of the project is largely due to its interactive nature. Although it may find moments of stability or failure, it is never finished. Due to this ongoing effect, Bloom can be related to the concept of genetics raised by John Frazer (1995) in ‘An Evolutionary Architecture’. Frazer states that the concept of biological growth can be applied to architectural form. The Bloom Project is designed through a natural process of growth, decline, and points of stability. This allows for a large number prototypical forms to be produced, what Frazer refers to as ‘pseudo-organisms’ and they often have unexpected results. The natural process of genetics can link to the process of growth that is observed through generation as opposed to composition, as is noticed in the opposing methods of modelling a tree by Aristid Lindenmayer and Thomas Gainsborough. Where Gainsborough described the tree once it was grown (composition), Lindenmayer grew the tree (generation). This relates closely to genetics as both look at natural processes over a long period of time, growth and evolution. Generation allows for numerous different versions using the algorithm to be achieved, so the designer can experiment. From there the pattern can be reasoned with through problem solving in computation to arrive at an optimum design. In the case of the Bloom Project, the concept of an optimum design is ever-changing.
Rhino and Grasshopper were used in the design of Bloom. Grasshopper allows for the use of genetic-like algorithms, which produce many prototypes that can then be evaluated in their simulated environment. This permits the transfer of preferred traits to new generations. The architects would have used computational thinking, which allowed them to bend computation to their own needs of creating something that can be assembled in many different ways. They would have been able to understand the aspects that worked well with computation and the traits that may be limiting. Algorithms such as cluster can change one group, or ‘generation’ and this will update all future generations. This mimics the way in which beneficial and survival enhancing traits are passed on to new generations in nature. It can be designed, altered, and dismantled by the public, which brings it to life as a piece.
Manual Recursion 1. SOFT AND BLOB-LIKE This component is very soft in nature with no hard or sharp edges. Even when manually recursed it blocked together to form a very dense array of components.
2. CURVED This aggregation has quite curved components that change direction, thus tricking the eye into which way the formation may go.
3. SHARP AND JAGGERED This component has very sharp edges that make it appear cold and menacing. The way they were oriented however has created an open structure that has areas that cluster and areas that rope the clusters together.
4. CURVED AND WARPED This component was curvey with very rounded edges, which when combined together gave a pattern that appeared twisted and deformed.
1. SOFT AND BLOB-LIKE
RULE SET: A = AC B = BC C = ABC CONDITIONAL RULES: If A intersects B, keep A If B intersects C, keep B If C intersects A, keep C
2. CURVED RULE SET: A = BC B = AB C = ABC CONDITIONAL RULES: If A intersects B, keep A If B intersects C, keep B If C intersects A, keep C
3. SHARP AND JAGGERED
RULE SET: A = AC B = AC C = ABC CONDITIONAL RULES: If A intersects B, keep A If B intersects C, keep B If C intersects A, keep C
4. CURVED AND WARPED
RULE SET: A = BC B = AC C = ABC CONDITIONAL RULES: If A intersects B, keep A If B intersects C, keep B If C intersects A, keep C
B.3 Case Study 2.0
BLOOM REVERSE ENGINEERED
1. I learnt how to do automatic recursion with Grasshopper using some of the components from manual recursion exercise.
2. I formed a Bloom component piece by looking at photos of Bloom Project.
3. I tested my Bloom component in the automatic recursion algorithm.
4. Going back to basics, I lowered the number of recursions and started to re-orient the base components.
5. I did many trials with different orientations of the base components.
6. With 2 recursions, I found a pattern with good spread and many orientations although a definite directionality.
7. I increased the recursions to 4 and was still happy with the outcome.
8. I increased the recursions to 5.
9. Example of aggregation reacting to environmental factors - done manually.
Between using automatic and manual recursion I found that although automatic was a lot faster and easier to change, I had more control when manual. I feel like I was able to create more interesting patterns using manual recursion although automatic was a lot more rule driven. In manual recursion, each component had to be oriented by hand, which was extremely time consuming and prone to straying from the rules. However I believe it gave me greater control over the aggregations made.
Reference the axiom polyline from Rhino into Grasshopper in the Curve component; it must be an L-shaped polyline.
Reference the b from Rhino into G Curve componen L-shaped polyline
To be able to view the bloom compo create a new Orient component. Th End, Plane, and Mesh components. O will also appear on the axiom and ba the directions of these curves.
New growth rules can be set using a ‘branch = growth rule’ panel.
The number of re adjusted by adjus number slider. Ad ton can be pres component step-
To bake, the orig ponent (not the step 4) must be se
branch polylines Grasshopper in the nt; they must be es.
Reference the bloom mesh created in Rhino into the ‘Mesh’ Grasshopper component.
onent on the axiom and base curves, his must be connected to the Plane Once preview is on, the components ase curves, making it easier to orient
ecursions can be sting the ‘Repeat’ dditionally the Butssed to grow the -by-step.
ginal Orient comone created in elected.
Set the orientations of the branch polylines in relation to the axiom.
In reverse engineering the Bloom Project, I have been able to create a form that strongly resembles it. This is due to the initial component that I modelled, which is almost identical to the original. This was achieved through using the Picture Frame command on Rhino and then tracing it, filleting edges, extruding, trimming, and capping it. The other main similarity is the colour. The use of pink may distract a viewer from seeing the actual array of components, but simply associate that pink with the Bloom Project. Differences arise in the actual orientation of the components. The Bloom Project, although its ability to be re-arranged infinitely has more directionality and organised curving nature. In comparison, the forms that I created using automatic recursion were slightly more random in appearance. I feel like I was able to creating more Bloom-like forms using manual recursion. But as I progress with using Grasshopper I would like to push the boundaries and create forms that are unique. It may be interesting to use different components also that separate my arrangements from previous works. I want to create an arrangement that has good spread without clumping together. Of course there should be areas of varying density and openness. I want a good sense of directionality although without it being completely linear. The form should seem random with a sense of organisation.
B.4 Technique Development
In this section, the Bloom Project has been used as a starting point from which four new components are formed and arranged with more branching options. This creates more complex forms that strive towards the enhancements mentioned in B3: spread, directionality, and randomness. They are designed with digital fabrication in mind.
Component 1 Ruleset 1 This aggregation has an organic, coral-like appearance. It mimicks such natural form in the way that it has areas of randomness that are however clustered into regular forms. I like the way it spreads out in multiple directions from a very central starting point. The components are planar and quite simple, which would make fabrication of the pieces fast and easy. It has area that are quite dense, which could affect assembly.
RULE SET: A = ABC B = AC C = BD D = AB E = BCD F = CD
Component 1 Ruleset 2 In comparison with the first ruleset using this component, this aggregation is a lot more random and bush-like. I would have prefered it to spread out a bit more in areas to relieve some of the mass. I believe the first rule set was more effective with this component as due to its blob-like nature, space is important in the aggregation.
RULE SET: A = CE B = AC C = BD D = AB E = AE F = CD
Component 2 Ruleset 1 The component used here is jaggered and pointy. I expected it to create a menacing array of pieces but was pleasantly surprised with the delicate aggregation that was produced. Although there are pieces facing almost all directions, there is a sense of directionality created seen through thinner areas trailing behind a denser mass. These pieces would be more difficult to fabricate than the previous component. This is due to the surface having many different heights, edges, and cut-outs. Due to the delicacy, I also worry that it would not support its self weight when many pieces are assembled.
RULE SET: A = CE B=A C = BD D=B E = AE F = ACD
Component 2 Ruleset 2 This aggregation is delicate and fragile. It is extremely open, which is a nice change from the dense aggregations I had been creating. It is light weight and clear. Although I like the way this aggregation presents digitally, I would be concerned for its ability to hold its weight when fabricated due to the extremely thin components and connections. If a suitable material was found it could be successful.
RULE SET: A = CE B=A C = BC D=B E=C F = ACD
Component 3 Ruleset 1 This component is an interesting one because on one side it is jaggered and hard and the other side is soft and rounded. When arranged in a composition however, this jaggered edge produces almost a blurred effect, which is not aesthetically appealing. Despite the unsuccess of the component, the aggregation does have nice spacing and directionality. I could imagine a similar aggregation that composes of a different component.
RULE SET: A = CE B=A C = BC D=B E = AB F = ACD
Component 3 Ruleset 2 This aggregation is interesting in that it is quite tightly woven in its body, but then has some random arms that stick out in certain places. It makes it seem very random. It also opens up the aggregation and provides areas of the aggregation that are denser than others. With a different component, this could be successful. However I feel as though the structure does lack balance, with some tweaking of the directions of the base curves a better solution could be found.
RULE SET: A = CE B = DE C = BCD D=B E = AB F = ACD
Component 4 Ruleset 1 This aggregation does have less base curves than the others and that is due to the fact that I stumbled across this pattern that I really liked and was unable to reproduce it to its same extent with more base curves. I like the way it has a rope-like element that runs through the middle from which many smaller lines or clusters form off. I would make the component slightly thinner just to make it more delicate. This could be fabricated easily and I believe it would work as a structure.
RULE SET: A = ABC B = AC C = DB D = AB E = BCD
Component 4 Ruleset 2 This aggregation has a good spread but it is lacking in direction, there are too many elements roping out of the main structure, making it seem very disorganised and too close to being symmetrical. In reality this would not be able to stand as the elements are too thin and they would not be able to support the denser centre. The previous ruleset with this piece made better use of the component.
RULE SET: A=A B = ACD C=D D = AB E = BCD F = AD G = BD
B.5 Technique Prototypes
An interesting method of fabrication that I have not yet used is 3D printing a mold and using this to cast many elements. 3D printing is an additive manufacturing process that makes three-dimensional objects. To create a mold, a solid block must be created around the 3D model on Rhino, and then split this block in two so the components can be easily cast inside it later. Once the mold is printed, there are various materials that can be casted such as silicone or plaster. Silicone could work well in this project as I believe it will hold better than plaster, which may break under the loads. If slotted in correctly together, the pieces should be able to hold themselves, just like the components in Bloom. Silicone can be translucent and mixed with a dye to give colour, which will stand out in Dulux Gallery.
This component could work well with laser cutting since it is a flat surface. I initially thought perspex could work well for this component but the jagged edges and excessive loads that will be on the entire structure led me to think otherwise. Perspex snaps easily and without warning, which could jeopardise the full composition. For this reason I believe MDF (medium density fibreboard) would be the most suitable. There is a Rhino template for laser cutting on the FabLab website. The components to be cut must be made 2D and then fitted onto the template. Once the job is complete, you must remove the pieces from the board and then can start assembling. When fitted together, the strength of many pieces will allow to structure to stand on its own. MDF does not look attractive in its post-laser cut state due to colour and visible burn marks so it should be painted.
Similarly to Component 2, this component could be done using the laser cutter due to its flat nature. As it is slightly more compact than the previous component, perspex could be a suitable material. I believe this may add a nice effect of translucency to the overall model and play with the transmission and distortion of light through the components. When laser cutting perspex, it is melted rather than burned (like MDF), which does not leave any marks along the edges. Once the laser cutting is complete, each piece must be carefully removed from the sheet and then can be fabricated by slotting the pieces together at the ends. As there is not as much space for the slotting together of elements, some kind of glue may also be needed in order to prevent collapse.
Unlike the other components, this one is not flat but rather slopes to a point at the base. For this reason I would choose to use 3D printing to create this element. 3D printing can only print in the colour of the material being used, most likely white, which eliminates the opportunity to make a coloured structure. However white could be nice in the way that it would blend more with the Dulux Gallery instead of completely stand out. The elements are quite small and delicate, the 3D printer will be able to create many of them with ease and efficiency.
B.6 Technique Proposal
The images in this section demonstrate how I have applied my aggregation to the Dulux Gallery. The site is very geometric and restrained in its linear shaping, contrasting to living, organic nature of my aggregation. Additionally, the use of colour in the aggregation makes it stand out from an otherwise dull and plain setting. When entering my aggregation into the site I avoided simply placing it there; I wanted it to interact. It protrudes over and is shaped by the walls which contain it, playfully avoiding contact with the surfaces of the gallery. I have left areas of the gallery empty to imply that there is room for more growth and to encourage further interaction with the structure. I could imagine people of all ranges interacting and rearranging the structure, as is done in Bloom. I could also see it as being a haven for small creatures such as birds or insects to explore and take refuge in. This is due to both to the way it mimicks a more natural formation and the many nooks and crevices. Despite areas of the aggregation being quite compact and dense, there are also areas of light and openness that provides a nice contrast.
B.7 Learning Objectives & Outcomes
REFLECTION
Part B has been extremely challenging, having reached the end I feel a strong sense of accomplishment. From not having any experience with algorithmic techniques a few weeks ago, I’m quite proud of the work I have now produced. From using Rhino in the past I thought I would be more comfortable with using manual recursion as it was all commands I had used before. However this proved to get confusing very quickly as the form rapidly grew in size and components. When doing a similar exercise automatically with Grasshopper it made me realise why using algorithmic techniques is so important. Automatic recursion made it easier to be able to alter my technique many times and produce many generations of the same component to find the ones that work best. I could test how these aggregations worked when placed in a simulated environment, the Dulux Gallery.
B.8 Appendix
ALGORITHMIC SKETCHES
References
Kolarevic, Branko, Architecture in the Digital Age: Design and Manufacturing (Spon Press: New York, 2003). MSD (2017), ‘FabLab’, accessed 15 September 2017, https://msd.unimelb. edu.au/fablab Plethora Project (2017), ‘Bloom’, accessed 29 August 2017, https://www. plethora-project.com/bloom/ Woodbury, Robert F. ‘How Designers Use Parameters’, in Theories of the Digital in Architecture, ed. by Rivka Oxman and Robert Oxman (London; New York: Routledge, 2014).