STUDIO AIR 2018, SEMESTER 2, Isabelle Jooste ANGUS JEFFREY 836314
Contents Introduction - 4 A.1 Design Futuring - 6 A.2 Design Computation - 8 A.3 Composition/Generation - 10 A.4 Conclusion - 13 A.5 Learning Outcomes -13 A.6 Appendix - Algorithmic Sketches - 14
My name is Angus Jeffrey and I am studying a double major of architecture and construction within the Bachelor of Environments, currently in my third year. Up to this point, my knowledge of digital architecture has largely revolved around ways in which it can speed up the design documentation phase of a project with software such as AutoCAD, Revit or Archicad. I have occasionally considered how a computer may be used to design structures inconceivable to the human mind, but have never felt the urge to improve my understanding of this area. I look forward to the opportunity presented by Studio Air to explore this area of design, though I am unsure if what I learn will be of great use to me in the future. In the past, I have used AutoCAD in conjunction with Photoshop to create two dimensional architectural drawings and Revit to create rendered three dimensional models. I also have some experience using Grasshopper, as I did Studio Air in semester 2 last year, though decided to withdraw for the sake of workload management. Generally, I would say that my technical skills are lower than those of my peers a because of the construction major, due to less time spent design courses where skills are often developed.
A.1 Design Futuring
Architecture has always been for its users, seeking to meet their needs and desires, taking on new form and nature as required. Should the occupants stay in the one spot, the architecture would remain where it was constructed. Should the occupants be nomads, any structures would need to be disassembled and re-erected. For much of architectural history, these two doctrines were almost entirely separate, with anything attempting to bridge the divide falling more into the category of vehicle than structure. Enter the Japanese Metabolists of the mid-20th century. In 1972, the Nakagin Capsule Tower had been completed thirty days after works began. Composed of capsules which may be inserted, removed and joined together to form larger spaces, the project represented the movement as a whole, being able to shape and ‘grow’ to the demands of the occupants. The capsules themselves were designed to be mass produced, with the eventual goal of there being several structures ready to receive them spread across the city. It is here perhaps that the first effort to bridge the divide between permanent and nomadic living with architecture is seen; one or several capsules may be removed from one tower and sent to another during the day for the occupants to ‘relocate’ in the evening. Though the Nakagin Capsule Tower did not realise its fanciful and radical aspiration, it significantly advanced the technologies and methodologies of prefabrication in architecture, which was a booming industry at the time and continues to be so today. Today, few of the original capsules are permanently occupied, with most either available for short term rent or in a state of neglect.
TOP: NAKAGIN TOWER BOTTOM: CAPSULE INTERIOR 6
TOP TO BOTTOM: THE ILLINOIS SKY CITY 1000 SIMZU MEGA CITY PYRAMID NOAH (NEW ORLEANS ARCOLOGY HABITAT
If the Nakagin Capsule Tower could be described as a city which can move around between buildings, the opposite would be a city that exists in a single building. This notion was first introduced by visionary American architect Frank Lloyd Wright in 1957 with his 1.6km high skyscraper named ‘The Illinois’. Wright’s goal was to create a structure capable of housing half of New York City (as of 1957)1, so that the land around the structure could be returned to nature. Though the reasoning behind The Illinois was to defeat the congestion of a horizontal city and to bring back the natural environment, the land recovered could be used for if suitable agriculture, easing the shortage of arable land. Alternatively, entire structures could be dedicated to hydroponic farming. The Illinois was never built however, as it exceeded the technologies of the time and would even prove impractical to construct today. It did however spark interest in the idea of super dense mixed-use structures, such as Sky City 1000 (Takenaka Corporation), Shimzu Mega-City Pyramid (Dante Bini & David Dimitric) or “arcological” structures as posited by Paolo Soleri2 , who spent a year and a half in the Taliesin Fellowship 3. Though above projects have not come to fruition, there now exists a polemic field of architecture focusing on such structures, which may provide a future solution either on Earth or as interplanetary colonies.
References: 1) Darlur Namsill, ‘Frank Lloyd Wright interview’ <13/1/14> https:// www.youtube.com/watch?v=DeKzIZAKG3E [accessed 2/8/18] 2) Sean O’Byrne, ‘Frank Lloyd Wright’s Taliesin West <n.d.> https://www. desertusa.com/desert-arizona/taliesen-west.html [accessed 2/8/18] 3) Paolo Soleri, The Bridge Between Matter & Spirit is Becoming Spirit: The Arcology of Paolo Soleri (Anchor Books, 1973), p46 NB: All knowledge of Nakagin Tower from MoMo to Pomo
A.2 Design Computation
When designing a structure in a traditional manner, an architect will consider what shapes and forms are used to define the structure. From this idea of what the structure should look like, it would be decided upon as to how it would be built with an arduous detailing stage. Calculations were also required to determine if the structure is capable of remaining upright and carrying loads. Fortunately, most architects opted to design using relatively simple Euclidean geometries, which made the above process trivial compared to those who used more complex forms for more elegant solutions. With the advent of computational design, designing, detailing for construction and structural calculation for more complex forms has become far easier and is bale to be done in a much shorter time span. Take for instance Bollinger-Grohmannâ€™s gridshell for the Sound Bites City exhibition in 2013 named The Torus 4. Using Karamba, the gridshell was modelled in a 3D space to perform structural analysis tests to determine what loads were being applied to which members. In addition, by working through a virtual model from the beginning of the project, the issue of fabrication was always being considered, and given the complex grids and curvature of The Torus, precision in detail for the placement of screws and correct overlapping of grid lines was critical for the structure to perform and appear as intended. It is clear then that the use of computational design allowed the designers to develop this complex array of individually simple geometries into an impressive and feasible design outcome, with a project timeframe to match. Without the use of a computer to inform the design, the entire process would have undoubtedly taken far more than the six weeks required with the use of a computer. As using a computer for drafting allows for results to be yielded in considerably less time than drafting by hand, so too does using a computer to inform the design yield results faster than merely using a computer to produce drawings.
SOUND BITES CITY GRID SHELL
CONNECTION OF GRID LINES
LOAD ANALYSIS, RED = HIGHER LOADS
TOP TO BOTTOM: CONCEPT DEVELOPMENT DIAGRAMS ARTIST’S IMPRESSION OF TOWER UPON COMPLETION ARTISTS’ IMPRESSION OF TOWER SKY LOBBY
“Form follows function”, the creed that has delivered some of the finest works of modernist architecture during the 20th century. This doctrine is still followed by many prevalent designers, though to a much higher or arguably ‘purer’ standard. Once again, SMG is an industry leader in this field of research, as seen in their execution of the National Bank of Kuwait Headquarters. This project highlights how computational tools such as Bentley System’s Generative Components can be used to define the building’s performance using parameters outlining structure, environmental considerations and operations and functionality. A parametric model was used during the early stages of design to test several variations responding to the established parameters, and later to fully realise the chosen design solution5. The model also presents the builder with all the information required to construct the project such that it performs as designed, facilitating the realisation of the project as a whole. With the accuracy of documentation being a major concern of construction contractors and the root cause of many defects during a project, using a single model to consolidate design ideas prevents discrepancies from occurring during the documentation phase of a project. In addition, using this integrated computational design method, SMG were able to produce results which could only fit the designated parameters, which eliminates the time required to test ideas or run simulations if the project were designed in a traditional sense, even with the aid of computer programs. Additionally, by creating their own system to measure the performance of the building, SMG would have been able to rate its performance as they saw fit. Some energy efficiency or environmental sustainability rating systems only focus on the building after it was built or with some other misleading and unannounced catch 6 , whereas if a custom rating system is used, the building can be studied in a much more honest sense, resulting in a finer quality of outcomes. References: 4) Bollinger+Grohmann, ‘Gridshell’ <n.d.> https://www. bollinger-grohmann.com/en.projects.gridshell.html?f=9778E4045801-0B76-5EE6-EFE87B64483F [accessed 8/8/18] 5) Dusanka Popovska, ‘Integrated Computational Design: National Bank of Kuwait Headquarters’, Architectural Design, Volume 83, Issue 2 (March/April 2013), 34-35
6) Franklyn Carter ‘Critics say LEED Program Doesn’t Fulfill Promises <8/9/10> https://www.npr.org/templates/story/ story.php?storyId=129727547 [accessed 3/8/18]
Technology is a strong instigator for change in any industry or field, which then goes on to change the methodologies and culture of that industry. Architecture is no different, and has seen considerable change since the increase in focus on generative design. To properly understand the impacts this new technique has had on the industry, one must consider how algorithmic thinking differs from traditional methods of designing and how it has changed the use of parametric models as an effective working environment for projects. Once this has been established, the cultures which have grown around these methods should be studied to gauge the impact upon the industry on a human level, rather than a technological one. Human beings are one of the few creatures on this planet to use tools, and are the only ones to fashion new tools to perform a specific task; while an otter may carry a rock to break open clam shells, humans have developed a tool specifically for this task, aptly named, a clam opener. With new tools come new ways of thinking and going about specific tasks, which has been demonstrated in the field of generative design, where designers are now developing their own software to use, rather than relying on packages from software distributors7. By creating and using their own tools, designers are able to produce works which tackle a given problem as they see fit. For instance, Programmer Joel Simon experimented with genetic algorithms to optimize the existing layout of a primary school to minimize walking time 8. The two principles underlying his work are graph contraction and ant colony pathing, which when used in conjunction produced a layout resembling a cluster of cells. Simon admits that his generations were not programmed to be practical, though it is perfectly achievable to produce results which are through generative design. Bollinger-Brohmann, the same firm that produced the gridshell in section A.2 once again used Karamba to create their own generative tool. It should be noted that while Karamba is a product and readily accessible, it does not provide a set of tools top use, but rather 10
a means through which to create tools to use which classifies the created tools it as a custom script. The algorithm was constructed to produce approximately 200,000 variations of truss layouts with random diagonal web members using an evolutionary process for a metro bridge at the Frankfurt Airport, Germany 9. The algorithm was also instructed to avoid complex member connections to reduce the cost of construction. Both this and Simonâ€™s floor plan show that through generative design, designers do not so much define the form as they would if designing compositionally, but rather define a set of rules for a computer to follow and determine viable design outcomes. With this in mind, the use of parametric models should now be examined.
ORIGINAL SCHOOL LAYOUT
LAYOUTS OPTIMISED FOR INTERNAL MOVEMENT
SKYLINK FROM GROUND
WORKING DIAGRAMS OF SKYLINK
As discussed in section A.2 with the use of the National Bank of Kuwait Headquarters as an example, parametric models offer the designers advantages in terms of evidence-based design outcomes and fully realised and realistic fabrication methods. The Torus, also from A.2, shows that a comprehensive parametric model greatly enables digital fabrication processes, which have seen a rise in popularity recently. This is because by using a parametric model, rapid digital fabrication of components can be easily achieved10. With an increased use in popularity of digital fabrication as a means to produce custom components for projects, a change in how we build similar to how we design as studied in the above paragraph can be observed. Beforehand, architects would use universally standardised elements such as the 90x45mm timber stud, 230x110x76mm brick or 410UB54, they may now create their own components, as Greg Lynn did in 2008 with his Blobwall project. Such components would be difficult, or impossible if complex enough, to design and realise without a parametric model which only allowed them to be defined in a virtual space based on algorithmic and generative rules. Between the new methods of thinking and new methods of constructing which arise therefrom, it would appear that the practice for architecture is poised for a series of discipline redefining breakthroughs in the near future. One must consider however, the drawbacks or dangers of carelessly pursuing this new and exciting field.
DISASSEMBLED BLOBLWALL References: 7) Peters, Brady. ‘Computation Works: The Building of Algorithmic Thought’, Architectural Design, 83, Issue 2 (March/April 2013), pp. 08-15 8) Simon, Joel, ‘Evolving Floor Plans’ <n.d.> http://www. joelsimon.net/evo_floorplans.html [accessed 8/8/18] 9) Bollinger+Grohmann, ‘Skylink Squaire Metro/Airport’ <n.d.> https:// www.bollinger-grohmann.com/en.projects.skylink-squaire-metro-airport. html?f=9778E404-5801-0B76-5EE6-EFE87B64483F [accessed 8/8/18] 10) David Celanto, ‘Innovate or Perish: New Technologies and Archietcure’s Future <n.d.> http://www.harvarddesignmagazine.org/issues/26/innovateor-perish-new-technologies-and-architectures-future [accessed 9/8/18]
A.3 Composition/Generation cont’d
As the former head of Foster + Partner’s SMG, Hugh Whitehead is indisputably well informed when it comes to the nature of the industry in terms of scripting cultures. In a special edition of the journal Architectural Design, he posits states that many within the field of scripting design operate as “lone gun[s]” and posits that this hinders the practice from becoming wholly integrated into architectural practice as a whole11. This perceived schism would create an ‘us/them’ mentality within creative teams, which can only lead to miscommunication or interpersonal friction. It may also harbour or aggrandise the egos of some designers, further causing tension in a highly cooperative field. Whitehead believes that to present this from happening, the real design goal must be kept in plain sight all throughout the process. Another possible problem with the dissemination of scripting culture into the architectural practice at large is that if the tools used to create a project are readily shared, some designs derived from that code may misuse or appropriate it to produce ineffective outcomes. A counterpoint to this is that by doing so, architecture becomes more democratised, perhaps even to the extent where non-designers may be able to create architecture, returning the role of architect to the human race as a whole. Such considerations are pure speculation, though what they mean for the future of architecture is sure to generate discussion as endless as the designs that the generative may yield. References: 11) Peters, Brady. ‘Computation Works: The Building of Algorithmic Thought’, Architectural Design, 83, Issue 2 (March/April 2013), pp. 08-15
A.4 Conclusion Having discussed and considered all of the above information, it seems that a generative approach will be best suited for my later design projects to create and fabricate facilities for my client. Given that my client is an animal, more specifically a boobook owl, the design must be something that it will find familiar and comforting without any persuasion whatsoever. For this reason, I believe that a generative approach in an attempt o recreate natural forms will deliver the optimal result. Of course, this is only a preliminary assumption, and discussion with my partner is required.
A.5 Learning Outcomes
Since this is my second attempt at studio air, I have already completed this task, though it was useful for me to research new cases to further broaden my understanding of the topic. I feel now a renewed vigour to continue learning what computational design can produce and how it can enhance a design, and I feel that this will make this second attempt more enjoyable. I look forward to taking what I have learnt and applying it to later assignments with my partner.
A.6 Appendix - Algorithmic Sketches
Basket Shells Together, these piped breps create an enclosure similar to what my partner and I had discussed to use as a space for the boobook owl to nest. The hollow space within is not too dissimilar to a hollowed tree trunk, satisfying the familiarity quality parameter. The surface of the larger iterations off no grip for possums or cats, which are the main predators of the boobook owl, which satisfies the security quality parameter. It should be noted that the â€˜skinniestâ€™ iteration would probably be the easiest to construct as some kinf of gridshell, though it would leave the owls more open in its current form, so a wrapping element such as canvas would be required to fulfill the secritury criterion. An single opening will need to be created in the surface and the ends will need to be capped before it could service the needs of tthe client. In relation to what was discovered in part
A.6 Appendix - Algorithmic Sketches
By experimenting with projection, I was able to map geometries of 1, 2 and 3 dimensions onto a reference surface to create the arrays above. This may well prove useful during the fabrication stage of the project in order to create defensive yet welcoming surface textures for the owl.
These tesselated spheres proved to be a simple way of creating a frame for a structure using a series of random points and Voronoi cells. As with the skinniest of the basket shells, they do not protect the owl from predators sufficiently for them to quality nesting hollows, however the openness they provide could allow them to serve as observation posts when suspended from branches.