PART
1 | Case Study and Parametric Modeling Practice
Case Study: The Water Cube
In the context of ARCH
655 - Parametric Modeling in Design, the selected case study is the façade of the
National Swimming Center (also called Water Cube) in China, designed by CSCEC +
PTW + CCDI and ARUP.
 |
| https://www.arup.com/en-us/projects/national-aquatics-center-water-cube/ |
https://www.westchinatour.com/beijing/attraction/water-cube.html
The National Aquatics Center, also known as the 'Water Cube', is one of the most dramatic and exciting sporting venues constructed for the 2008 Beijing Olympic Games. Arup’s designers and structural engineers came up with a design for the walls inspired by the natural formation of soap bubbles, a unique geometry that was both highly repetitive and buildable, yet which appears pleasingly random and organic. With Ethyl tetrafluoroethylene (ETFE) used for the cladding, Arup designed the building to be sustainable, well-lit and seismically resistant (ARUP).
This endeavor investigates the integration of design intents within a parametric modelling framework, utilizing Rhino/Grasshopper for the initial modeling phase.
Parametric Modeling Practice:
A systematic investigation of the Water Cube's exterior facade skin and structure was conducted using Rhino/Grasshopper software. The objective was to dissect and reimagine the facade structure through parametric design methods, laying the groundwork for iterative development and exploration of design intent. Due to laptop configuration limitations, this project only reconstructed and optimized the roof of the Water Cube.
 |
| https://www.arup.com/en-us/projects/national-aquatics-center-water-cube/ |
This process first generates random points within the surface using the Populate2D battery. Subsequently, based on these points, a Voronoi diagram is employed to create a cellular-like pattern, simulating the facade structure of the Water Cube.
Mesh Refinement and Simulation
Utilize the TriReMesh tool to subdivide the generated faces, creating a mesh that lays the groundwork for the next step using Kangaroo. The Kangaroo plugin in Grasshopper can be employed to simulate the air-supported film of the Water Cube's skin, generating a more realistic roof structure.
Detailed Optimization
After generating the simulated skin using Kangaroo's Solver,
parameter details were optimized by adjusting the length of the mesh
subdivision to maximize the fidelity of the Water Cube's roof structure. Due to
the limitations of the laptop configuration, this result still has room for
further refinement in the future.
PART 2 | Reinterpretation with Generative AI
After generating the original roof of the Water Cube using Rhino/Grasshopper, utilize ChatGPT's built-in Dall-E image generation feature to create an optimized exterior facade design for the Water Cube. The specific process was as follows: The top image was uploaded to ChatGPT, accompanied by the prompt: “This is the Water Cube. I want to enhance the design of its exterior facade. Could you generate an image for me? Keep all other parts unchanged. I want the cellular texture of the facade to transform into a flowing, water-like sensation. Use this membrane material to fill the structure, making each unit more elongated and fluid, like flowing water”. ChatGPT generated the image shown on the bottom.
 |
| Origin image. https://www.westchinatour.com/beijing/attraction/water-cube.html |
 |
| Image generated by ChatGPT |
Upon its completion, the Water Cube pioneered innovative architectural design through its extensive use of ETFE membrane structures, while also setting new standards in energy efficiency and environmental sustainability. Consequently, during the redesign of its exterior facade, the membrane structure skin was retained. Simultaneously, a functional analysis of the building identified water-inspired design elements, ensuring the building's skin form and functional requirements are now more closely aligned.
This method leverages artificial intelligence to expand the feasibility of architectural design schemes, delivering diverse design outcomes. It enables early-stage control and adjustment of design proposals, playing a crucial role in establishing Grasshopper models.
PART 3 | Parametric Form – Mass & Skin
Based on the
optimization scheme generated by DALL·E, the skin of the Water Cube was
reconstructed using Rhino/Grasshopper. The challenge in this process lay in
generating random yet directionally consistent curves on a plane to serve as
both the plane's dividing lines and the foundation for its structural
framework.
Establishing Base Points
Filtering the Data Tree
The purpose of this step is to filter
the points that need to be randomly moved. As shown in the figure below, the
points are divided into 11 branches, each containing 8 points. Points located
at the surface edges form the fixed framework and cannot be moved. Therefore,
all points except those in the first and last branches must be filtered. The
result, shown in the figure below, is that the Tree Branch battery obtains all
points from the 2nd to the 10th branches.
Move
Points Randomly and Generate Curves
Utilizing
a series of List batteries + Random battery, control points to move randomly
along the x-axis. Link these points to generate random yet directionally
consistent curves that subdivide the plane, yielding 10 faces as the foundation
for the film of the surface. Note: Before proceeding, ensure all points are
flattened so each point can move independently. If not flattened,
points on each branch will move as a single unit.
Mesh Refinement and Detailing
Utilize the curves generated in the previous step to segment the original surface. Subdivide the resulting multiple surfaces using the TriReMesh battery, then generate the skin simulating the air film with the Kangaroo plugin. Control the length input via the Number Slider at the Length input port of the TriReMesh battery. Due to laptop limitations, finer processing was not feasible. Apply Pipe processing to the structural framework to obtain the final model.
Comments
Post a Comment