Previously, I created a tower using grasshopper that responded to the level and position of occupants within each floor to create a system of variable extrusions that opened the facade to the exterior when people gathered near them. This solution focused on the adaptation of a single space to changes in the distribution and total number of occupants. This was sought as both a way to express the use of the building to outside observers and as a way to maximize views for inhabitants as they approach the windows from the inside. Since the focus of the building for inhabitants is the management of light and views within the structure, the general goal of my optimization is to ensure an effective but evenly muted daylight within the structure under normal conditions, but to allow views to open as an occupant or group of occupants approaches the exterior wall.
While I enjoyed the concept of this experiment, further refinement is necessary to make this system operate at peak efficiency. Since this project was conceived as an office building, we know that it is an internally dominated system with regards to heating and cooling demands. Since the two factors that most contribute to generally high cooling loads in a high rise building are occupancy level and lighting I wanted to optimize the design to allow for the maximum amount of daylight autonomy within the building during typical hours of occupancy: Monday through Friday 9-5. The less electric lighting necessary within the building, the lower the energy use will be in terms of both lighting and cooling loads.
To achieve this goal I plan to maximize daylight within the building by mass optimization that allows for full daylight autonomy, then to reduce glare through a variable perforated paneling skin. The first step in this process was to go back to basic considerations of building massing with the goal of creating a two step optimization process. In the first stage I have created a basic mass that is variable between 6 and 12 floors. It is constrained to between 3 and 10 sides, between 14 and 40 meters in diameter, and with a floor to floor height of 3 to 6 meters. Rotation of the orientation is allowed to ensure ideal solar exposure to the floor plates.
This mass has been attached to an environmental analysis script using the Honeybee plugin and Radiance which analyzes total radiation exposure on the floorplates of the building. The results of this analysis are then read and each point is assigned a daylight autonomy value which measures the percentage of time during active occupancy that daylight over a certain threshold is attained (300 lux in this case). Before my final optimization I will be researching ideal lux threshold levels in office buildings to find a more realistic value for this number.
Using the Galapagos optimization method variable massing parameters are optimized for maximization of daylight autonomy within the building. Initial attempts at optimization with a previous script highlighted several issues with my parameters that required tweaking before my final result was obtained. One example is that because Galapagos averages multiple fitness values the optimized results naturally trended towards fewer floors. In my initial script I found that higher level results all had only a single floor so I had to set the number of floors to a consistent number – my height being 12 stories.
I also found that the optimization consistently reduced the number of sides to three, but that high daylight autonomy was achievable with more sides as well. Therefore, the number of sides of the building is mainly a design decision and in this case I would prefer a pentagon shaped mass. This decision allowed me to reduce the rotation constraint to a range of 72 degrees.
Another issue with my original script that is yet unresolved is that my building mass doesn’t always split into floors in the way that it is supposed to. I checked multiple inputs for incongruities but have yet to find anything that explains this issue.
The full script and a screenshot are included below for anyone wishing to see in more detail how this system works. An update will be coming soon that shows the results of this optimization routine. For my next steps I will be reintroducing the panelized system to this optimized mass and finding the optimal shading pattern for each panel on the mass to reduce glare and provide the desired amount of even daylighting within the interior.