In the middle of a new public park in San Francisco stands an old and unsightly utility vent. The challenge was to devise an artful enclosure while adhering to strict spatial constraints set by BART, the regional transportation authority that owns the vent. Since I knew that we were working under a tight budget and timeline, I created a parametric 3D modeling tool that allowed our team to generate complex prototypes at the speed of thought.
Team
Nico Wright
Client
Strada Investment Group / Kennerly Architecture
Role
Project Designer / Parametric Designer
Skills / Tools
Rhino, Grasshopper, Sketchup, 3D Printing, Illustrator
CONTEXT
In the heart of our project's open space stands a structure that provides access to and vents the BART tunnel passing under our site. My team's challenge was to come up with an artful approach to screening the structure, so in essence, an attractive utility fence. As per BART’s regulations, our design had to adhere to a 3’ clear zone around the utility structure, while at the same time avoiding manhole covers on either side. The design also had to funnel air pushed by passing trains from the structure’s side vents in an upward direction. Working under these tight constraints challenged us to come up with a design that balanced form and function.

BART vent structure's location within the larger project context (Credit: CMG)

Diagram of physical constraints surrounding the vent structure (Credit: CMG)

EARLY ITERATIONS
Right before I joined the project, our team created a series of designs were meant to push the boundaries of what could be done on the site. We were trying to create an iconic structure, something that would serve as a landmark in the city, which led to some bold sculptural ideas. After a few sessions of brainstorming, the client decided that they wanted to pursue a purely geometrical form with some transparency in the landscape. 
Early conceptual iterations for the BART vent structure (Credit: CMG)
ADVANCING THE LANTERN SCHEME
We named the client’s choice “The Lantern” since we intended for this geometric, semi-transparent structure to light up at night. This is the point where I was first brought onto the project to advance the concept into a concrete design. The first steps in advancing the lantern involved refining the geometries, iterating on our initial concept by adjusting the structure’s faces, angles, and proportions. I then dove into drawing the structure to scale, tweaking our initial design to make sure that it fit our tight constraints. In this stage I found that studying the structure in section and elevation from all sides was crucial to advancing our design.
Elevations showing the BART vent structure in relation to landscape and architectural elements.
FINALIZING THE GEOMETRY
After the detailed 2 dimensional studies were completed, I was confident that the structure would meet all the spatial requirements set by BART. Those studies yielded exact dimensions which I carried over into Rhino to build a simple 3D massing model. In order to better understand our final shape, I used a 3D printer to bring our massing model to life. The physical model allowed our team to identify problem areas in our design that we didn't catch while looking at a screen. Namely, one of the sides was too visually static, prompting a slight redesign of that face. 
3D printed model of "The Lantern" scheme, alongside another iteration of the enclosure that we were exploring. 
MODELING THE SKIN SYSTEM
Once the structure's geometries were finalized, I dove into detailed design explorations of the structure's skin. The intent behind the structure’s skin system was to create a visual gradient that increased in transparency as your eye moves up the structure. A higher opacity towards the bottom fit our desired aesthetic while also serving a dual purpose by deterring climbing and funneling air upwards. I devised a system that layered four different spacings of welded wire mesh, starting with a 1”x 1” grid and moving up to 4”x 4”. As you move up the structure, you lose a layer, increasing the skin’s transparency. 
My project manager liked this idea, but wanted to be able to test this design in 3D in order to accurately show the nuances of how light and shadow interacted with the mesh. I was also asked to model a few iterations of the same design using different sizes and configurations of welded wire mesh. At first this seemed like an almost impossible task in the short time frame I was given, since a mesh with this level of detail can take countless hours to model. Since we had so little time and budget, I proposed that we automate the design process by using the parametric design software Grasshopper in tandem with Rhino 3D.
The first step in developing the parametric script was researching which parameters we wanted to modify in the model. I found a manufacturer's website that carried a wide range of wire meshes, and discovered that the meshes have three main product specifications: gauge, spacing, and weave pattern. 
Material specifications for wire mesh pulled from the manufacturer's website (Credit: McNichols Wire Mesh)
With these parameters in mind I began to develop the script. My basic thinking process behind writing the script was to break down how I would model the skin system manually, and translate it into code for Rhino 3D using the Grasshopper plug-in. To help me visualize the code's structure I created a simple flow diagram that broke down the process from start to finish. 
After creating this high-level roadmap I broke my coding effort into three manageable chunks, which were developed one at a time. The code took a few iterations to get perfectly right, and ultimately was a hybrid of custom code and code that was pulled from the internet and past projects. Below are screenshots of what the final product looked like in the Grasshopper terminal, alongside it's corresponding place in the overall flowchart.
The resulting script allowed me to iterate at the speed of thought, and I was able to bring a dozen unique design options to the table at our next design meeting. After sharing the results of my visual studies, the team ultimately landed on the option shown below, which started off with a 4x4 grid on top and incrementally added layers that decreased in spacing towards the bottom, with the outermost grid maintaining 4x4 spacing the entire way.
RESULTS
After the internal review we presented our preferred design to the client. Our solution was admired for meeting all the spatial requirements set by BART while balancing monumental character with cost-effective materials. The legacy of my work on this project is that the script I developed can easily be modified by anyone in the office as the project goes through design development on its way to construction documentation. If at any time we decide to use a different type of welded wire mesh, the new specifications can be plugged into the script to help visualize what the final product will look like. In addition, this script can easily be adapted to any future 3D modeling projects that come into the office that use welded wire mesh in their design. 
Back to Top