Step 1 - Generative Design Framework
1. Material Selection for Façade Cladding (Cost vs. Sustainability)
- Design Variables:
- Material type (e.g., aluminum, GFRC, terracotta, wood)
- Cladding panel size
- Thickness or weight per unit area
- Evaluators:
- Embodied carbon per m² (kg CO₂e/m²)
- Material cost per m²
- Maintenance frequency or durability score
- Key Tradeoffs:
- Lower carbon materials may be more expensive or less durable.
- Thinner panels reduce cost and weight but may underperform in durability.
- Natural materials may perform well in sustainability but require more maintenance.
2. Column Grid Spacing (Structural)
- Design Variables:
- Bay size (spacing between columns in X and Y directions)
- Floor-to-floor height
- Structural system type (e.g., moment frame vs. braced frame)
- Evaluators:
- Structural material volume (steel or concrete usage)
- Lateral drift under seismic or wind loads
- Floor vibration performance (comfort)
- Floor plan flexibility for tenants
- Key Tradeoffs:
- Larger bays improve open floor layout but may increase beam depth and deflection (longer spans).
- Tighter grids reduce structural demands but increase material and labor cost.
- Floor-to-floor height impacts both structural efficiency and MEP integration.
3. Tower Orientation for Solar Performance
- Design Variables:
- Rotation angle of the entire tower (0° to 360°)
- Optionally: floor-by-floor twist angle (for twisted towers)
- Evaluators:
- Solar radiation on each façade (kWh/m²)
- Cooling load potential (especially on east/south façades)
- Daylight availability on interior zones
- Tradeoffs:
- Orienting narrow faces east-west reduces solar heat gain but limits views.
- Broad southern exposure maximizes daylight but increases cooling demand.
- Rotating the tower changes shadow impact on adjacent buildings or public space.
Step 2 - Generative Design Study
Design Decision: Façade Cladding Material Selection (Cost vs. Sustainability)
I chose to run a Generative Design Study on the material selection of the facade cladding to explore the tradeoffs between material cost and embodied carbon. Given that the tower is fully glazed or clad, the choice of cladding material significantly impacts both construction cost and the carbon footprint of the building. This decision is meaningful early in the design process and simple to model in Dynamo, hence it was chosen for this study.
Materials may vary via a material index (i.e. 0 = Aluminum, 1 = Terracota, 2 = Timber, 3 = Glass), while the cladding area may vary as a continuous variable depending on the geometry of the parametric tower model. Additionally, thickness of the cladding may be used as an adjustable input to weigh between the durability and the material cost or embodied carbon. A higher durability typically comes from a greater material thickness, yet this will also increase embodied carbon and material costs, which should idealize be minimized. A higher durability is valuable as it will reduce maintenance costs in the long-term.
Other evaluation metrics that can be observed based on material thickness include thermal mass and structural load, which both increase in thickness. While increasing thermal mass is beneficial for improving energy performance by buffering temperature swings, increasing structural load requires more structural support and overall increases building costs. For simplicity, these metrics will not be taken into account for this study.
Generative Design Framework structure
Below is an outline of the node logic used to create the study graph for the Generative Design Study. It starts with definitions of varied inputs (pink) and constant inputs (purple) that are used. Varied inputs include the top and base radii, material type (via an index), and material thickness. The outputs observed will be the total material cost, total embodied carbon, and the durability of the facade based on the material and material thickness.
The constant inputs are the remaining variables needed to define the tower geometry (i.e. story height, rotation). Then the tower is defined using three polygon profiles and lofted into a solid. 6
While the custom node “BuildingForm.SelectWallSurfaces.dyn” could have been used to extract the wall surfaces of the tower, I decided to embed the node logic instead. It takes a list of the faces in the tower and identifies which are ground, roof, and wall surfaces based on a normal vector.
Once the wall surfaces (facade) are extracted, the total surface area is calculated. The purple group contains node logic that extracts the necessary information for material type, cost per m^2, carbon footprint in kg Co2e/m^2, and a durability score out of 10, based on the index called. Below is a summary of rough values used to characterize each material.
The total surface area is multiplied by the unit cost and unit carbon footprint to generate the total values for the facade. Note that since durability not only depends on the material type but also the material thickness, thus the following formula is used: durabilityScore = baseDurability × log10(thicknessMM + 1)
Using log models the idea that increased thickness improves durability, but there are diminishing returns — doubling thickness doesn't double durability. For example, going from 10 mm to 20 mm has noticeable improvement, but going from 90 mm to 100 mm has a much smaller impact.
Once the study graph was saved, a study was defined as follows using the “Optimize” method. Total cost of facade and total embodied carbon are directly proportionally to each other, thus both are minimized while durability was maximized. The final outputs generated the top 20 performing models.
Step 3 - Generative Design Study Results
Below are the input and output values for the 20 optimized outputs from the Generative Design study. Although there were 4 materials provided (index 0-3), it was observed that Glass (index = 3), never made it to the optimized list. This makes sense since it has the most undesirable combination of characteristics that we wish to optimize, such that it is relatively costly, has a high carbon footprint, and the lowest durability score. Additionally, I’ve identified the cases where certain output values are at their highest or lowest. As hypothesized, the best-case iterations for lowest cost and lowest carbon footprint coincide and it corresponds to the worst-case for durability score (lowest). This also applies vice versa, where the highest durability (best-case) also coincides with the same iteration where it has the highest cost and embodied carbon (worst-case).
Based on the parallel graph shown below, it is observed that the total cost and total carbon footprint are in fairly distinct ranges depending on the material used, whereas durability is much more varied as it depends on material and thickness, which is a more continuous variable. While geometry varies, the top and base radii don’t seem to have a major impact on carbon or cost. This suggests that the material choice and thickness are much more influential.
For the scatterplot, designs farther to the right have higher durability, but they also tend to have a higher cost (move up the Y-axis). Bubble size and color both indicate carbon footprint, with larger/darker dots representing higher emissions.
The bolded line and bolded dot in the two graphs below highlight a potential compromise between the output values, where cost and embodied carbon are relatively low on the scale, while durability is in mid-range. This corresponds to timber at a high thickness, which in practical industry, is in fact a sturdy sustainable alternative as seen in mass timber structures.
This information directly impacts the design decision by highlighting the tradeoffs between cost, carbon footprint, and durability. It shows that achieving higher durability results in higher facade costs and greater embodied carbon, whereas lower-cost, lower-carbon options tend to yield lower durability scores. Understanding these relationships allows the designer to prioritize according to project goals—if durability is essential due to environmental exposure or lifespan requirements, a higher-cost, higher-carbon material may be justified. If the focus is on reducing environmental impact or staying within budget, a moderately durable but more sustainable option may be preferable. With this in mind, I would use the graph to filter and identify designs that offer the best possible balance where no objective can be improved without sacrificing another, and select from among those based on project-specific priorities.
- An image of your Dynamo Study Graph (showing all your nodes and the connecting logic)
