- Automotive Complex, ONL [Oosterhuis_Lénárd] & Hyperbody, 2005
- Automotive Complex
- Abu Dhabi
- Project architect
- Kas Oosterhuis
- Design team ONL [Oosterhuis_Lénárd]
- Kas Oosterhuis, Ilona Lénárd, Gijs Joosen, Cas Aalbers, Sander Boer, Barbara Janssen, Henrike Michler, Eirini Logara, Han Feng, Brenda Vonk Noordegraaf
- Project team Hyperbody
- Thomas Jaskiewicz, Chris Kievid, Dieter Vandoren
Automotive Complex, behavioural modelling applied in urban-scale development
Tomasz Jaskiewicz and Chris Kievid
There’s no better way to validate theories than by applying them in practice. Hyperbody design application prototypes, created to allow non-linear design of complex architectural and urban behavioural models, were tested by being applied to a master planning project of the Automotive Complex by ONL [Oosterhuis_Lénárd].
The Automotive Complex project assignment was not trivial. It was to create a master plan for the site located on the outskirts of Abu Dhabi, covering more than 6 km². This huge site came with a huge program of demands. 800,000 m² of covered showrooms, car shops, car-themed leisure and experience, restaurants, hotel, conference centre and even two gas stations. All accompanied by over 500,000 m² of covered parking spaces and many more outdoor facilities such as a Formula 1 and 4x4 racetracks. All of this had to be organized into one project, while retaining a certain amount of flexibility as to the exact location and floor spaces of planned functions. What’s more, this project not only had to work as a well-organized complex system; it also had to become the landmark of a global destination hub for all car enthusiasts.
The scale of the designed master plan is comparable to a small town. However, the complexity of the planned development in many ways surpasses the complexity of a town system. To be able to handle this, a new kind of a parametric design strategy was used. The program of demands was broken down into individual program cells, ranging from 1,000 to 13,000 m², ordered in a Fibonacci sequence. These cells were subsequently given additional parameters; function type, floor space area, number of floors, average floor height, additional shape defining parameters and connections to other program elements. All this information was put into a spreadsheet and exported to the custom-made application based on Hyperbody’s Swarm Toolkit developments and protoSPACE design tools .
The subsequent steps taken in the design process can be described as behavioural modelling . In a behavioural modelling process, each of the unique parts of a project, in this case the program cells become autonomously acting entities.
The behaviour given to the individual cells of the Automotive Complex master plan was simple in its principles. Each cell had to avoid overlapping with other cells. If that were to happen, the cell would be forced to move away from the collision point. Connections between particular cells were formed in a top-down manner. If two cells were connected, they would try to stay closer to each other than the specified maximum and further away than the specified minimum distance. Handles attached to each of them could control the bulkiness of cell capsule shapes. Additionally, spatial attractors: points, lines and curves, were introduced to the system in order to repel or attract all, or just a selection, of cells to specific areas in the master plan.
Out of the individual, simple, local actions and interactions of such entities, a complex, holistic system came into being. Even though all the rules and behaviours were very straightforward, the system as a whole exhibited a very high degree of adaptability, allowing the emergence of unpredictable qualities. To use a relevant analogy; the behavioural model can operate similarly to how a city evolves over decades or centuries, but with the main difference that the same and many more new and unprecedented qualities can now emerge within minutes instead of years or centuries of the system’s operation, while at the same time allowing for explicit top-down control of their development throughout the process.
While the distribution of the functional program coming from its decentralized behaviour was entirely emergent and unpredictable, attractors and explicit relations between selected cells were used to ‘tame’ the whole system, add deterministic qualities and impose constraints on it, in order to allow its operation within chosen development scenarios and an overall design vision. Introduction of global parameters such as the total floor space or volume of all cells together made it possible to change these values locally while maintaining strict control of the overall project vision and its feasibility.
The ability to achieve holistic design quality and exceptionally efficient, yet flexible spatial organization of the project were both high-priority aspects of the master plan. The project is meant to become a global landmark location, visible from most airplanes approaching to land at the Abu Dhabi international airport. In order to give it the desired exceptional appearance, both design and spatial organization systems were inspired by the ripples that appear on the desert surrounding the project site as an effect of wind interacting with individual sand particles and the forces keeping these particles together.
In order to achieve a similar self-organizing behaviour, several kinds of geometric curve objects with unique behaviour called powerlines  were introduced to the system. These powerlines enforced a behaviour of the project components analogous to that, which in nature causes the appearance of wind ripples on sand. Different powerlines follow the spatial arrangement of program cells and form a pattern of communication flow as well as creating the shape of the building skin wrapping the entire project.
Inside the complex, the flow of cars, people, energy and information follows these powerlines. All aspects of movement are smooth and continuous for both cars and people. The result of this is a 3D woven pattern of circulation for cars and people crystallized in the functional plans of the project. The circulation begins on the southwest entrance of the complex and continues throughout the entire 1.5 kilometre long building. The user never has to stop; he or she naturally follows the communication routes while being informed both physically and virtually about the directions to take in order to reach particular destinations.
On the outside, the building skin uses a different type of powerline to achieve its unique articulation embedded in the skin covering the entire building. Subtle ripples form a new landscape emerging from the desert, with the most articulated parts covering the hotel and conference centre and voids from which thrives the greenery of newly created oases.
In the same way that the scale model of the project was created using a file-to-factory process, the digital data of the project can be sent directly to contractors and their machines to accurately and efficiently build the entire complex in full scale and detail.
The qualities coming from the swarm logic embedded in the design process of the Automotive Complex master plan are in many ways absolutely revolutionary. As Kas Oosterhuis envisioned; ‘Building and architecture have from now on two aspects to them: on the one hand one they create a physical environment, while on the other hand, they design the behaviour, the rules of the game, the states of mind of the buildings and the environments directly connected to the physical places’ . The Automotive Complex project proves that this idea is valid not only for buildings, but also for an entire urban-scale development. The designed master plan has a unique and consistent design and organization, yet it can be dynamically changed and adjusted throughout the design process, and potentially also after completion of subsequent phases of its construction. All this without compromising any of its functional, engineering and design qualities.
1. Oosterhuis (et al.) ‘Protospace Software’, in Second International Conference World of Construction Project Management (2007).
2. Reynolds, Craig, ‘Flocks, Herds, and Schools: A Distributed Behavioural Model’, in: Computer Graphics 21(4), SIGGRAPH '87 Conference Proceedings (1987).
3. Lénárd, Ilona, ‘Powerlines’, in iA#2 (Rotterdam: episode publishers, 2009).
4. Oosterhuis, Kas, ‘Swarm Architecture’, in: Hyperbodies: Towards an E-Motive Architecture (Basel: Birkhäuser, 2003).